Second Edition
William b. White
David c. culver
ENCYCLOPEDIA
OF CAVES
SECOND EDITION
ENCYCLOPEDIA
OF CAVES
SECOND EDITION
Editors
William B. White
The Pennsylvania State University
David C. Culver
American University
ELSEVIER
AMSTERDAM • BOSTON • HEIDELBERG • LONDON NEW YORK • OXFORD
PARIS • SAN DIEGO • SAN FRANCISCO • SYDNEY • TOKYO
Academic Press is an imprint of Elsevier
Academic Press is an imprint of Elsevier
225 Wyman Street, Waltham, MA 02451, USA
The Boulevard, Langford Lane, Kidlington, Oxford, OX51GB, UK
Radarweg 29, PO Box 211, 1000 AE Amsterdam, The Netherlands
525 B Street, Suite 1900, San Diego, CA 92101-4495, USA
First edition 2005
Second edition 2012
Copyright © 2012 Elsevier Inc. All rights reserved
No part of this publication may be reproduced, stored in a retrieval system or transmitted in any form or by any means
electronic, mechanical, photocopying, recording or otherwise without the prior written permission of the publisher.
Permissions may be sought directly from Elsevier's Science & Technology Rights Department in Oxford,
UK: phone (+44) (0) 1865 843830; fax (+44) (0) 1865 853333; email: permissions@elsevier.com.
Alternatively you can submit your request online by visiting the Elsevier web site at http://elsevier.com/locate/permissions,
and selecting Obtaining permission to use Elsevier material.
Notice
No responsibility is assumed by the publisher for any injury and/or damage to persons or
property as a matter of products liability, negligence or otherwise, or from any use or operation of any methods, products,
instructions or ideas contained in the material herein.
Library of Congress Cataloging-in-Publication Data
Encyclopedia of caves/editors David C. Culver, William B. White. - 2nd ed.
p. cm.
Includes bibliographical references and index.
ISBN 978-0-12-383832-2 (alk. paper)
1. Speleology-Encyclopedias. 2. Caves-Encyclopedias. I. Culver, David C., 1944- II.
White, William B. (William Blaine), 1934-GB601.E534 2012
551.44'703-dc23
2011039751
British Library Cataloguing in Publication Data
A catalogue record for this book is available from the British Library
ISBN: 978-0-12-383832-2
For information on all Academic Press publications
visit our website at www.elsevierdirect.com
Printed and bound in China
12 13 14 109 8 765432
Working together to grow
libraries in developing countries
www.elsevier.com | www.bookaid.org | www.sabre.org
ELSEVIER
BOOK AID
International
Sabre Foundation
Contents
Contents by Subject Area ix
List of Contributors xi
Guide to the Encyclopedia xiii
Preface xv
A
Adaptation to Low Food 1
Kathiin Hiipfmp
Adaptive Shifts 9
Francis C. Howarth tout Hotmehn Hoch
Anchihaline (Anchialinc) Caves and Fauna 17
Boris Sfc-f
Ancient Cavers in Eastern North America 25
Pally /t> IVoiso'i
Asetius aquatints: A Model System for Historical
Biogeography 30
Riuh Vermnik
Astyanax mexicanus: A Model Organism for Evolution
and Adaptation 36
Wtllinm K [effe ry
B
Bats 45
VtoHia* H Kiuk. SusijN VV. Murray and \atlmt IV. Fuller
Beetles 54
Oana Tetxlora Moltlovan
Behavioral Adaptations 62
lakob Panrfall
Breakdown 68
Elizabeth L. White
Burnsville Cove, Virginia 74
Gregg $. CUmmer
c
Camps 85
Gregg S. Clemmer
Castleguard Cave, Canada 89
Davit Ford
Cave Dwellers in the Middle East 94
Paul Goldberg and Ofcr Bar-Yostf
Cave Ecosystems 99
Kevin S. Simon
Cave, Definition of 103
William B. Mute and David C. Culver
Cavcfish of China 107
Li Met and Ya lnn Zhao
Chemoautotrophy 125
Annette Summers Engel
Clastic Sediments in Caves 134
Grt',v>i>n/ 5. Springer
Closed Depressions in Karst Areas 140
Ugo Satini
Coastal Caves 155
[oh n F. Mylmie
Contamination of Cave Waters by Heavy Metals 161
Dorothy j Vesper
Contamination of Cave Waters by Nonaqueous
Phase Liquids 166
Caroline M. Loop
Cosmogonic Isotope Dating of Cave Sediments 172
Parry! £. Granger and Derek Fabel
Crustacea 177
Horton II Hobbs III
D
Dinaric Karst: Geography and Geology 195
Madia Z.upan Hajna
Diversity Patterns in Australia 203
William F. Humphreys
Diversity Patterns in Europe 219
/.ants Deharwng, famne Giber! and David C. Culver
Diversity Patterns in the Dinaric Karst 228
Boris Sket
Diversity Patterns in the Tropics 238
Louis Deharoeng and Anne Bedos
Diversity Patterns In the United States 251
Hinton II 'llnhh III
Documentation and Databases 264
Peter Matthews
VI
CONTENTS
E
Ecological Classification of Subterranean
Organisms 275
Eleonora Trajano
Entranceless Caves, Discovery of 277
Mil'll tv. Davis
Entrances 280
William B. While
Epikarst 284
Michel Bakalaipicz
Epikarst Communities 288
Diivid C. Cutivr, Anton Unmeet/ ami Taiijit Pipon
Evolution of Lineages 295
Eleonora Trajano and .Marina Ciilsuffi
Exploration of Caves—General 304
Williatn P. While
Exploration of Caves—Underwater Exploration
Techniques 310
/iff ffeinerlh
Exploration of Caves—Vertical Caving
Techniques 314
MurJr Minton and Yvonne Proms
F
Folklore, Myth, and Legend, Caves in 321
Paul lav Steward
Food Sources 323
Thomas L. Paulson
Friars Hole System, West Virginia 334
Stephen K.H. Worthington ami Douglas M. MedvtUe
G
Gammarus minus: A Model System for the Study of
Adaptation to the Cave Environment 341
Daniel W. Fong
Geophysics of Locating Karst and Caves 348
Barlwa Anne am Ende
Glacier Caves 353
fason P. Culhy and Andrew C Fountain
Guano Communities 357
Pedro Gnasphii
Gypsum Caves 384
Alexander Khmcluruk
Gypsum Flowers and Related Speloothems 374
William II. White
H
Helictites and Related Speleothems 379
Donald G. Davis
Hydrogeology of Karst Aquifers 383
William B White '
Hydrothermal Caves 391
Yuri Dublyamkp
I
Ice in Caves 399
Atirel Perfoiu and Bogdan P. Onac
Invasion, Active versus Passive 404
Dan L. Danicttrpol and Kat/moml Pouch
J
Jewel Cave, South Dakota 411
Mike Wiles
K
Karren, Cave 419
Joyce Lundlvrg
Karren, Surface 425
Joyce Lundberg
Karst 430
William K. limes and William H. White
Kazumura Cave, Hawaii 438
Kevin Allreti
Krubera (Voronja) Cave 443
Alexander Klimctmik
L
Lampenflora 451
Jancz Midee
Lechuguilla Cave, New Mexico, U.S.A. 45b
Patricia Kamlvsis
Life History Evolution 46.5
David C. Cnber
M
Mammoth Cave System, Kentucky 469
Roger IV. Brucker
Mapping Subterranean Biodiversity 474
Man/ C. Christman and Ma/a Zagmn/ster
Marine Regressions 482
Claude Boutin and Nicole Coineau
CONTENTS
vii
Mava Caves 48b
Aadrto Stone and /arm’s £. Urndy
Microbes 490
Annette Sumnurs Engel
Minerals 499
Bogdan P. Onac
Modeling of Karst Aquifers 508
Caw# Kaufnunn. Douchko Romanov mid Wolfgang Dreybrodt
Mollusks 512
PiK'Jif C. Culver
Morphological Adaptations 517
Kenneth OmMntsm
Multilevel Caves and Landscape Evolution 528
Darlene M. Anthony
Mulu Caves, Malaysia 531
Inny Waltham and Joel Despain
Myriapods 538
C Cuhvr and William A . Shear
N
Natural Selection 543
Peter Twntclj
Neutral Mutations 549
Horst Wilkens
Niphargus : A Model System for Evolution and
Ecology 555
Cene FBer
Nitrate Contamination in Karst Groundwater 564
Brian C. Katz
Nullarbor Caves, Australia 568
Julia M. lames. Amtulisa K. Cantos and Craig XI. flames
P
Paleoclimate Records from Speleothems 577
Victor /. Polyak and Rhawn F. Dcmiiston
Paleomagnctic Records in Cave Sediments 585
Ira D. Sastmsbf
Paleontology of Caves 590
Blame IV. Schubert ami Jim 1. Mead
Passage Growth and I>evelopment 598
Arthur K. Palmer
Passages 603
Gouge Veni
Population Structure 608
Valerio Sbonbni, Giuliana Allegrucei and Donatella Cesaroni
Postojna—Planina Cave System, Slovenia 618
Shmka Selvla
Protecting Caves and Cave Life 624
William R. Elliott
Q
Quartzite Caves of South America
Augusta S. Auler
635
R
Recreational Caving 641
John M. Wilson
Rescues 648
John C. 1 lempel
Responses to Low Oxygen 651
Frdtdric Hermitl and Florian Malard
Root Communities in Lava Tubes 658
Fred D. Stone. Francis G. Howerth, llannelote Floch and
Manfred Asche
s
Salamanders 665
Spela Conch, Matthew L. Niemdler and Dante B. Fentdio
Saltpetre Mining 676
Diii’irf A. Hubbard Jr.
Scallops 679
Phillii> I Murphy
Shallow Subterranean Habitats 683
Tanfa Pijian and David C. Cuhvr
Show Caves 690
Arrigo A. Cigna
Siebenhengste Cave System, Switzerland 698
Pierre- Y'ivs leannin and Philipp lldusehmmn
Sinking Streams and Losing Streams 707
fosofih A. Ray
Sistema Huautla, Mexico 712
C. William Steele and lames H Smith, Jr.
Soil Piping and Sinkhole Failures 718
Bany Beck
Solution Caves in Regions of High Relief 723
Philipp Hduselmatin
Solution Caves in Regions of Moderate Relief 733
Arthur N. Palmer
Species Interactions 743
David C. Cuhvr
Spelcogenesis, Hypogcnic 748
Alexander Khmehouk
VIM
CONTENTS
Speleogenesis, Telogenetic 765
Fraud Calnwfek
Speleothem Deposition 769
Wolfgang Drej/brodl
Speleothems: General Overview 777
William If, White
Spiders and Related Groups 786
james R. Raided
Springs 797
William B. l\7r/ft*
Stalactites and Stalagmites 805
Silvia Frisia a nd /on D. Woodhead
Sulfuric Acid Caves 810
Arthur V. Pahurr and Catol A. Hill
T
Tiankeng 821
XMftwn 7.1m. Weihai Chen and Yuanluti 7han$
u
Ukraine Giant Gypsum Caves 827
Alexander Klimchouk
Underwater Caves of the Yucatan Peninsula 833
lames G. Cake, IV
Uranium Scries Dating of Speleothems 838
Christoph S/uill and Runny Boeh
V
Vertebrate Visitors—Birds and Mammals 845
Mikola Tvrlkovie
Vicariance and Dispcrsalist Biogeography 849
/idw R Hobinger
Vjetrenica Cave, Bosnia and Herzegovina 858
In' Lucie
Volcanic Rock Caves 863
Slepfutu Kenifie
w
Wakulla Spring Underwater Cave System,
Florida 875
Barlvira Anne am Lnde
Water Chemistry in Caves 881
land S. Herman
Water Tracing in Karst Aquifers 887
William K. /ones
Wetlands in Cave and Karst Regions 897
Tanja Pipan and David C. Culver
White-Nose Syndrome: A Fungal Disease of North
American Hibernating Bats 904
Marianne S. Moore and Thomas II. Kurtz
Worms 910
FJshiehi Dumnu ka
Index 917
Contents By Subject Area
TYPES OF CAVES
Anchihaline caves and fauna
Cave, definition of
Coastal caves
Entranceless caves, discovery of
Glacier caves
Gypsum caves
Hydrothermal caves
Multilevel caves and landscape evolution
Quartzite caves of South America
Show caves
Solution caves in regions of high relief
Solution caves in regions of moderate relief
Sulfuric acid caves
Volcanic rock caves
CAVE FEATURES
Breakdown
Clastic sediments in caves
Entrances
Ice in caves
Karren, cave
Paleomagnetic records in cave sediments
Passage growth and development
Passages
Scallops
Nitrate contamination in karst groundwater
Passages growth and development
Speleogenesis, hypogenetic
Speleogenesis, telogenetic
Sinking streams and losing streams
Water chemistry in caves
Water tracing in karst aquifers
SPELEOTHEMS AND OTHER
CAVE FORMATIONS
Gypsum flowers and related speleothems
Helictites and related speleothems
Karren, cave
Minerals
Paleoclimate records from speleothems
Speleothem deposition
Speleothems — general overview
Stalactites and stalagmites
CAVE AGES AND PALEOCLIMATE
Cosmogenic isotope dating of cave sediments
Multilevel caves and landscape evolution
Marine regression
Paleomagnetic record in cave sediments
Uranium series dating of speleothems
SURFACE KARST FEATURES EXCEPTIONAL CAVES
Closed depressions in karst areas
Dinaric karst — geography and geology
Karren, surface
Karst, general overview
Sinking streams and losing streams
Soil piping and sinkhole features
Springs
Tiankeng
Wetlands in cave and karst regions
Burnsville Cove, Virginia
Castleguard Cave, Canada
Friars Hole Cave system, West Virginia
Jewel Cave, South Dakota
Kazumura Cave, Hawaii
Krubera (Voronja) Cave,
Lechuguilla Cave, New Mexico, USA
Mammoth Cave system, Kentucky
Mulu Caves, Malaysia
Nullarbor Caves, Australia
Postojna-Planina Cave system, Slovenia
Quartzite caves of South America
Siehenhengste Cave system, Switzerland
Sistema Huautla, Mexico
Ukraine giant gypsum caves
Underwater caves of the Yucatan Peninsula
Vjetrenica Cave, Bosnia and Hercegovina
Wakulla Spring underwater cave system, Florida
HYDROLOGY AND HYDROGEOLOGY
Contamination of cave waters by heavy metals
Contamination of cave waters by nonaqueous phase liquids
Epikarst
Hydrogeology of karst aquifers
Modeling karst aquifers
IX
X
CONTENTS BY SUBJECT AREA
GROUNDWATER CONTAMINATION
AND LAND-USE HAZARDS IN
CAVE REGIONS
Contamination of cave waters by heavy metals
Contamination of cave waters by nonaqueous phase liquids
Nitrate contamination in karst groundwater
Soil piping and sinkhole failures
HISTORICAL USE OF CAVES
Ancient cavers in eastern North America
Cave dwellers in the Middle East
Folklore, myth, and legend, caves in
Maya caves
Saltpetre mining
CONTEMPORARY USE OF CAVES
Lampenflora
Mapping subterranean biodiversity
Protecting caves and cave life
Recreational caving
Show caves
EXPLORATION OF CAVES
Camps
Documentation and databases
Entranceless caves, discovery of
Exploration of caves — general
Exploration of caves — underwater exploration techniques
Exploration of caves — vertical exploration techniques
Recreational caving
Rescues
BIOLOGY OF PARTICULAR
ORGANISMS IN CAVES
Asellus aquaticus — a model system for historical biogeography
Astyanax mexicanns — a model system for evolution and adaptation
Bats
Beetles
Cavefish in China
Crustacea
Gammarus minus — a model system for the study of adaptation to
the cave environment
Microbes
Molluscs
Myriapods
Niphargus — a model system for evolution and ecology
Paleontology of caves
Salamanders
Spiders and related groups
Vertebrate visitors — birds and mammals
White Nose Syndrome — a fungal disease of North American
hibernating bats
Worms
COMMUNITIES AND HABITATS
Cave ecosystems
Ecological classification of cave organisms
Epikarst communities
Guano communities
Lampenflora
Root communities in lava tubes
Shallow subterranean habitats
ECOLOGY
Chemoautotrophy
Food sources
Niphargus — a model system for evolution and ecology
Population structure
Responses to low oxygen
Species interactions
BIOGEOGRAPHY AND DISPERSAL
Asellus aquaticus — a model system for historical biogeography
Invasion, active versus passive
Marine regression
Vicariance and dispersalist biogeography
DIVERSITY
Diversity patterns in Australia
Diversity patterns in Dinaric karst
Diversity patterns in Europe
Diversity patterns in the tropics
Diversity patterns in the United States
Mapping subterranean biodiversity
EVOLUTION AND ADAPTATION
Adaptation to low food
Adaptive shifts
Astyanax mexicanus — a model system for evolution and adaptation
Behavioral adaptations
Evolution of lineages
Gammarus minus — a model system for the study of adaptation to
the cave environment
Life history evolution
Morphological adaptation
Natural selection in caves
Neutral mutation theory
List of Contributors
Giuliana Allegrucci Tor Vergata University, Italy
Kevin Allred Hawaii Speleological Survey
Barbara Anne am Ende Deep Caves Consulting
Darlene M. Anthony Roane State Community College
Manfred Asche Museum fur Naturkunde, Germany
Augusto S. Auler Instituto do Carste, Brazil
Michel Bakalowicz Universite Montpellier 2, France
Craig M. Barnes University of Sydney, Australia
Ofer Bar-Yosef Harvard University
Barry Beck P.E. LaMoreaux & Associates, Inc.
Anne Bedos Museum National d'Histoire Naturelle, France
Claude Boutin Universite Paul Sabatier, France
James E. Brady California State University, Los Angeles
Anton Brancelj National Institute of Biology, Slovenia
Roger W. Brucker Cave Research Foundation
Donatella Cesaroni Tor Vergata University, Roma, Italy
Weihai Chen Chinese Academy of Geological Sciences,
China
Kenneth Christiansen Grinnell College
Mary C. Christman University of Florida
Gregg S. Clemmer Butler Cave Conservation Society, Inc.
Marina Cobolli University of Rome "La Sapienza", Italy
Nicole Coineau Observatoire Oceanologique de Banyuls,
France
James G. Coke, IV Texas
Annalisa K. Contos University of Sydney, Australia
David C. Culver American University
Dan L. Danielopol Austrian Academy of Sciences,
Austria
Nevin W. Davis Butler Cave Conservation Society, Inc.
Donald G. Davis National Speleological Society
Louis Deharveng Museum National d'Histoire Naturelle,
France
Rhawn F. Denniston Cornell College
Joel Despain National Park Service
Wolfgang Dreybrodt University of Bremen, Germany
Yvonne Droms U.S. Deep Caving Team
Yuri Dublyansky Innsbruck University, Austria
Elzbieta Dumnicka Polish Academy of Sciences, Poland
William R. Elliott Missouri Department of Conservation
Derek Fabel University of Glasgow, Scotland
Dante B. Fenolio Atlanta Botanical Garden
Cene Fiser University of Ljubljana, Slovenia
Daniel W. Fong American University
Derek Ford McMaster University, Canada
Andrew G. Fountain Portland State University
Silvia Frisia The University of Newcastle, Australia
Nathan W. Fuller Boston University
Franci Gabrovsek Karst Research Institute at ZRC SAZU,
Slovenia
Janine Gibert Universite' Claude Bernard Lyon I, France
(deceased)
Pedro Gnaspini University of Sao Paulo, Brazil
Paul Goldberg Boston University and Harvard University
Spela Goricki University of Maryland
Darryl E. Granger Purdue University
Jason D. Gulley University of Texas
Philipp Hauselmann Swiss Institute for Speleology and
Karst Studies (SISKA), Switzerland
Jill Heinerth Heinerth Productions, Inc.
John C. Hempel EEI Geophysical
Janet S. Herman University of Virginia
Frederic Hervant Universite Claude Bernard Lyon 1,
France
Carol A. Hill University of New Mexico
Horton H. Hobbs III Wittenberg University
Hannelore Hoch Universitat zu Berlin, Germany
John R. Holsinger Old Dominion University, Norfolk,
Virginia
Francis G. Howarth Bernice P. Bishop Museum
David A. Hubbard Jr. Virginia Speleological Survey
William F. Humphreys Western Australian Museum
Kathrin Hiippop Institute for Vogelforschung, Germany
Julia M. James University of Sydney, Australia
Paul Jay Steward Cave Research Foundation
Pierre-Yves Jeannin Swiss Institute for Speleology and
Karst Studies (SISKA), Switzerland
William R. Jeffery University of Maryland
Patty Jo Watson Washington University
William K. Jones Karst Waters Institute
Patricia Kambesis Cave Research Foundation
Brian G. Katz U.S. Geological Survey
xi
Xll
LIST OF CONTRIBUTORS
Georg Kaufmann Free University of Berlin, Germany
Stephan Kempe University of Technology Darmstadt,
Germany
Alexander Klimchouk Ukrainian Institute of Speleology
and Karstology, Ukraine
Thomas H. Kunz Boston University
Caroline M. Loop Groundwater Management Associates
Ivo Lucic Speleological Association Vjetrenica, Bosnia and
Herzegovina
Joyce Lundberg Carleton University, Ottawa, Canada
Li Ma Department of Biology, University of Maryland
Florian Malard Universite Claude Bernard Lyon 1
Jim I. Mead East Tennessee State University
Douglas M. Medville West Virginia Speleological Survey
Mark Minton U.S. Deep Caving Team
Marianne S. Moore Boston University
Janez Mulec Karst Research Institute, at ZRC SAZU,
Slovenia
Phillip J. Murphy University of Leeds, UK
Susan W. Murray Boston University
John E. Mylroie Mississippi State University
Matthew L. Niemiller University of Tennessee
Bogdan P. Onac University of South Florida, Tampa,
U.S.A., and "Emil Racovita" Institute of Speleology,
Romania
Arthur N. Palmer State University of New York Oneonta
Jakob Parzefall University of Hamburg, Germany
Aurel Per§oiu University of Suceava, Romania
Tanja Pipan Karst Research Institute at ZRC SAZU,
Slovenia
Victor J. Polyak University of New Mexico
Thomas L. Poulson Jupiter, Florida
Joseph A. Ray Crawford Hydrology Laboratory, Western
Kentucky University
James R. Reddell The University of Texas at Austin
Douchko Romanov Free University of Berlin, Germany
Raymond Rouch Centre National de la Recherche
Scientifique, France (Retired)
Ira D. Sasowsky University of Akron
Ugo Sauro University of Padova, Italy
Valerio Sbordoni Tor Vergata University, Italy
Blaine W. Schubert East Tennessee State University
Stanka Sebela Karst Research Institute at ZRC SAZU
William A. Shear Hampden-Sydney College
Kevin S. Simon The University of Auckland,
New Zealand
Boris Sket Univerza v Ljubljani, Slovenia
James H. Smith, Jr. Environmental Protection Agency
Gregory S. Springer Ohio University
C.William Steele Boy Scouts of America
Andrea Stone University of Wisconsin-Milwaukee
Fred D. Stone University of Hawai'i at Hilo
Annette Summers Engel University of Tennessee
Oana Teodora Moldovan Emil Racovitza Institute of
Speleology, Romania
Eleonora Trajano Universidade de Sao Paulo, Brazil
Peter Trontelj University of Ljubljana, Slovenia
George Veni National Cave and Karst Research Institute
Rudi Verovnik University of Ljubljana, Slovenia
Dorothy J. Vesper West Virginia University
Tony Waltham British Cave Research Association, U.K.
Elizabeth L. White Hydrologic Investigations
William B. White The Pennsylvania State University
Mike Wiles Jewel Cave National Monument
Horst Wilkens University of Hamburg, Germany
John M. Wilson Marks Products, Inc.
Jon D. Woodhead The University of Melbourne, Australia
Stephen R.H. Worthington Worthington Groundwater,
Canada
Maja Zagmajster University of Ljubljana, Slovenia
Yuanhai Zhang Chinese Academy of Geological Sciences,
China
Ya-hui Zhao Institute of Zoology, Chinese Academy of
Sciences, China
Xuewen Zhu Chinese Academy of Geological Sciences,
Guilin, China
Nadja Zupan Hajna Karst Research Instituteat ZRC SAZU
Guide to the Encyclopedia
The Encyclopedia of Caves is a complete source of
information on the subject of caves and life in caves,
contained within a single volume. Each article in the
Encyclopedia provides an overview of the selected topic
to inform a broad spectrum of readers, from biologists
and geologists conducting research in related areas, to
students and the interested general public.
In order that you, the reader, will derive the maxi¬
mum benefit from the Encyclopedia of Caves, we have pro¬
vided this Guide. It explains how the book is organized
and how the information within its pages can be located.
SUBJECT AREAS
The Encyclopedia of Caves presents 128 separate arti¬
cles on the entire range of speleological study. Articles
in the Encyclopedia fall within 17 general subject areas,
as follows:
• Types of Caves
• Cave Features
• Surface Karst Features
• Hydrology and Hydrogeology
• Speleothems and Other Cave Deposits
• Cave Ages and Paleoclimate
• Exceptional Caves
• Biology of Particular Organisms in Caves
• Communities and habitats
• Ecology
• Cave Invasion
• Biogeography and Dispersal
• Evolution and Adaptation in Caves
• Evolution and Adaptation in Caves
• Diversity
• Contemporary Use of Caves
• Historical Use of Caves
• Ground Water Contamination and Land Use
Hazards in Cave Regions
ORGANIZATION
The Encyclopedia of Caves is organized to provide the
maximum ease of use for its readers. All of the articles
are arranged in a single alphabetical sequence by title.
An alphabetical Table of Contents for the articles can
be found beginning on page v of this introductory
section.
So they can be more easily identified, article titles
begin with the key word or phrase indicating the topic,
with any descriptive terms following this. For exam¬
ple, "Invasion, Active versus Passive" is the title
assigned to this article, rather than "Active versus
Passive Invasion," because the specific term Invasion is
the key word.
You can use this alphabetical Table of Contents by
itself to locate a topic, or you can first identify the topic
in the Contents by Subject Area on page x and then go
to the alphabetical Table to find the page location.
ARTICLE FORMAT
Each article in the Encyclopedia begins with introduc¬
tory text that defines the topic being discussed and
indicates its significance. For example, the article
"Behavioral Adaptations" begins as follows:
Animals living in darkness have to compete for
food, mates, and space for undisturbed reproduction
just as their epigean conspecifics do in the epigean
habitats, but there is one striking difference: In light,
animals can use visual signals. Thus, important aspects
of their behavior driven by visual signals cannot apply
in darkness. The question arises, then, of how cave
dwellers compensate for this disadvantage in complete
darkness. This article uses several examples to com¬
pare various behavior patterns among cave dwelling
populations with epigean ancestors.
Major headings highlight important subtopics that
are discussed in the article. For example, the article
"Beetles" includes these topics: Adaptations, Coloniza¬
tion and Geographical Distribution, Systematics of
Cave Beetles, Ecology, Importance and Protection.
CROSS-REFERENCES
Cross-references appear within the Encyclopedia as
indications of related topics at the end of a particular
xm
XIV
GUIDE TO THE ENCYCLOPEDIA
article. As an example, a cross-reference at the end of
an article can be found in the entry "Camps." This arti¬
cle concludes with the statement:
See Also the Following Articles
Recreational Caving • Exploration of Light Sources
This reference indicates that these related articles all
provide some additional information about Camps.
BIBLIOGRAPHY
The Bibliography section appears as the last element
of the article. This section lists recent secondary
sources that will aid the reader in locating more
detailed or technical information on the topic at hand.
Review articles and research papers that are important
to a more detailed understanding of the topic are
also listed here. The Bibliography entries in this
Encyclopedia are for the benefit of the reader to
provide references for further reading or additional
research on the given topic. Thus they typically consist
of a limited number of entries. They are not intended
to represent a complete listing of all the materials con¬
sulted by the author or authors in preparing the article.
The Bibliography is in effect an extension of the article
itself, and it represents the author's choice as to the
best sources available for additional information.
INDEX
The Subject Index for the Encyclopedia of Caves con¬
tains more than 4600 entries. Within the entry for a
given topic, references to general coverage of the topic
appear first, such as a complete article on the subject.
References to more specific aspects of the topic then
appear below this in an indented list.
Preface
Throughout history, caves have always been of at
least some interest to almost everyone. During the past
few centuries, caves have been of passionate interest to
at least a few people. The number of those with a pas¬
sionate interest has been continuously growing. The
core of the cave enthusiasts are, of course, the cave
explorers. However, scientists of various sorts, mainly
geologists and biologists, have also found caves useful
and fascinating subjects for scientific study.
There have always been cave explorers. Some, such
as E.A. Martel in France in the late 1800s, achieved
amazing feats of exploration of deep alpine caves. In
the United States, the number of individuals seriously
interested in the exploration of caves has grown con¬
tinuously since the 1940s. Cave exploration takes many
forms. Some cavers are interested in caving simply as
a recreational experience, not intrinsically different
from hiking, rock climbing, or mountain biking. But
many pursue genuine exploration. Their objective is
the discovery of new cave passages never before seen
by humans. As the more obvious entrances and the
more accessible caves have been explored, cave explo¬
ration in the true sense of the word exploration, has
become more elaborate and more difficult. To meet the
challenge of larger, more obscure and more difficult
caves, cavers have responded with the invention of
new techniques, new equipment, and the training
required to use them. To meet the challenge of long
and difficult caves, cavers have been willing to accept
the discipline of project and expedition caving and to
accept the arduous tasks of surveying caves as they
are explored. The result has been the accumulation of
a tremendous wealth of information about caves that
has been invaluable to those studying caves from a sci¬
entific point of view.
In the early years of the twentieth century, a few
geologists became interested in the processes that
allowed caves to form. Biologists were interested in
the unique habitats and the specialized organisms that
evolved there. In both sciences and in both Europe and
the United States, the interest was in the caves them¬
selves. The study of caves was focused inward and
some proposed the study of caves to be a separate sci¬
ence called speleology. In Europe, largely due to the
influence of the Romanian biologist Emil Racovita,
other subterranean habitats were included in the study
of cave life. In the latter decades of the twentieth cen¬
tury, there was a gradual change in perspective, and
the study of caves came to be seen as important for its
illumination of other realms of science.
In the past few decades, the geological study of
caves has undergone a tremendous expansion in point
of view. The caves themselves are no longer seen as
simply geological oddities that need to be explained.
Caves are repositories and are part of something
larger. As repositories, the clastic sediments in caves
and the speleothems in caves have been found to be
records of past climatic and hydrologic conditions.
Cave passages themselves are recognized as fragments
of conduit systems that are or were an intrinsic part of
the groundwater system. Active caves give direct
insight into the hydrology and dry caves are records
that tell something of how drainage systems have
evolved. Techniques for the dating of cave deposits
have locked down events much more accurately in the
caves than on the land surface above. Caves then
become an important marker for interpreting the evo¬
lution of the surface topography. Even the original,
rather prosaic, problem of explaining the origin and
development of caves has required delving into the
chemistry of groundwater interactions with carbonate
rocks and on the fluid mechanics of groundwater flow.
Cave biology has likewise evolved from an exercise
in taxonomy—discovering, describing, and classifying
organisms from caves—to the use of caves as natural
laboratories for ecology and evolutionary studies. The
central question that has occupied the attention of biol¬
ogists at least since the time of Lamarck is how did
animals come to lose their eyes and pigment. The
question gets answered each generation using the sci¬
entific tools available and its most contemporary form
is a question of the fate of eye genes themselves. Cave
animals have also served as models for the study of
adaptation because of their ability to survive in the
harsh environments of caves. There are also interesting
biological questions about the evolutionary history of
cave animals that are being unraveled using a variety
of contemporary techniques. Finally, there is increasing
xv
XVI
PREFACE
concern about the conservation of cave animals.
Nearly all have very restricted ranges and many are
found in only a single cave. The past two decades
have seen a phenomenal growth in the understanding
of how to manage cave and karst areas to protect the
species that depend on them.
One should not suppose that caves are of interest
only to geologists and biologists. Caves are repositories
of archaeological and paleontological resources.
Ancient art has been preserved in caves. Caves appear
in folk tales, legends, mythology, and in the religions
of many peoples. Caves appear frequently in literature,
either as an interesting setting for the story or as a met¬
aphor. The latter has a history extending at least back
to Plato.
In planning the content of the first edition of the
Encyclopedia of Caves, the editors were faced with this
great variety of "clients" with their highly diverse
interests in caves. Several decisions were made. One
was that we would address the interests of as many
"clients" as possible given the limitations of space.
Thus, the Encyclopedia, in addition to the expected
articles on biology and geology, also contains articles
on exploration techniques, archaeology, and folklore.
A second decision was to allow authors a reasonable
page space so they could discuss their assigned subject
in some depth. As a result of this decision, the
Encyclopedia contains a smaller number of articles
and thus a smaller number of subjects than might be
expected. The object was to provide a good cross-sec¬
tion of contemporary knowledge of caves rather than
attempt an entry for every possible subject.
The level of presentation was intended to be at the
college level. In this way, the articles would have
sufficient technical depth to be useful to specialists but
would still be accessible to the general reader. Some of
the subjects are intrinsically more technical than
others, but we have attempted to keep to a minimum
the specialist jargon and in particular the obnoxious
acronyms that turn many technical subjects into a
secret code known only to insiders.
The editors maintained that same criteria and guide¬
lines in constructing the second edition. Authors of
articles in the first edition were invited to revise and
update their articles and most of them did. Many of
the entries from the first edition have been completely
revised and expanded; others received only minor
updates. Many new articles have been added. The first
edition contained 107 articles; the second edition con¬
tains 128. In addition, a few old entries were dropped
and new articles with new authors were added.
The selection of authors was made by the editors.
We attempted to select contributors who we knew
were expert in the subject being requested of them. For
many subjects there was certainly a choice of potential
experts and our selection was to some degree arbi¬
trary. We sincerely hope that no one is offended that
some other person was selected rather than them.
We take this opportunity to thank the authors for
their hard work. The Encyclopedia is a collective effort
of many people in many disciplines. We are particu¬
larly appreciative of everyone's efforts to communicate
with cave enthusiasts outside of their particular
discipline.
William B. While
David C. Cidver
May, 2011
ADAPTATION TO LOW FOOD
Kathrin Hiippop
Helgoland
INTRODUCTION
Subterranean environments are characterized not only
by continuous darkness but also by a reduced variability
in the number of specific abiotic conditions such as mois¬
ture, temperature, and water chemistry, as well as by iso¬
lation and restriction in space. Additionally, hypogean
systems are relatively energy-limited compared to photo-
synthetically based epigean systems. As a response, many
cave animals share numerous adaptations to the food
scarcity of their environment. They not only show mor¬
phological and behavioral adaptations but also have
evolved several special physiological characters, most
notably energy economy, a reduction in energy consump¬
tion, which has a high selective advantage in cave animals
and has been observed in numerous species in a variety
of phyla (Poulson, 1963; Culver, 1982; Hiippop, 2000).
Factors concerning adaptation to food scarcity in caves
are illustrated in the causal network shown in Figure 1 .
On one hand, the high environmental stability in
caves, including darkness and sometimes predator
scarcity, allows the evolution of characters; on the other
hand, it requires character changes. In fact, most charac¬
teristics of adaptation to food scarcity can only be real¬
ized in ecologically stable and, above all, predator-poor
caves (Fig. 1). Food scarcity acting as a selective force
in caves requires adaptations. Possible adaptations of
cave animals to survival in caves low in food are an
improved food-finding ability, an improved starvation
resistance, a reduced energy demand, and life history
characters changed toward more k-selected features
(Hiippop, 2000). Further, a higher food utilization effi¬
ciency, dietary shift, and feeding generalism may be
realized. Many of these characteristics have evolved
coincidentally, depending on the kind of food scarcity.
As a consequence, most real cave animals show no or
only minor signs of malnutrition despite the low food
availability in their environment.
TYPES OF FOOD SCARCITY
Not only the intensity but also the quality of the food
scarcity and the duration of this selective force determine
the degree of adaptation. Food scarcity in caves can have
three facets: general food scarcity, periodic food supply,
and patchy food scarcity. General food scarcity holds for
nearly all caves and occurs especially in caves with no or
low but continuous food input. Additionally, many caves
are not stable throughout the year. Periodic food supply
characterizes caves that are flooded periodically (nor¬
mally several times during the rainy season) or caves
with periodic food input by animals visiting the cave reg¬
ularly. Seasonally flooded caves are subject to severe
changes regarding food input, water quality, oxygen con¬
tent, temperature, and competitors or predators. During
the wet season, food supply can be very high and even
abundant for some weeks or months. After exhaustion of
these food reserves, animals in such caves suffer food
scarcity like animals in generally food-poor caves. Some
cave animals have to cope with patchy food scarcity.
This means that food is not necessarily limited but is
locally concentrated and difficult to find and exploit.
Under such conditions, cave organisms can be observed
aggregated at patchy food resources.
WHAT TO FEED IN CAVES
Food Input
The basic food resource in most caves is organic mat¬
ter from external origin. Wind, percolating surface water,
flooding, and streams provide input of many kinds of
organic matter, such as detritus, microorganisms, feces,
and accidental or dead animals. Some caves are visited
Encyclopedia of Caves.
1
© 2012 Elsevier Inc. All rights reserved.
2
A
FIGURE 1 Network of characters observed in many cave animals, their causes, consequences, and interrelations. Dark gray: selection fac¬
tors; light gray: main characters influenced by food scarcity; close arrowheads require something, open arrowheads allow something. Redrawn
after Hiippop, 2000.
actively by epigean animals for shelter or reproduction.
Such caves are much richer in food than are more iso¬
lated ones, because the visitors provide an additional
food input in the form of their feces or their carcasses.
Bat guano can present an immense source of food for
guanobionts. Bacteria and above all microfungi decom¬
pose detritus and guano, thus building the basis for a
food pyramid in caves. Lava tubes can be rich in food
due to exudates from roots growing through the ceilings
into the caves (Poulson and Lavoie, 2000).
Chemoautotrophy
As the only primary producers in caves, a few species
of chemoautotrophic bacteria may support the survival
of cave animals, especially in caves that have no natural
entrance and where the absence of water infiltration
from the surface excludes the input of photosynthetic
food (Sarbu, 2000). However, these chemoautotrophic
systems are quantitatively important in only a few
exceptional caves, the best known example being the
Mo vile Cave in Romania.
Influence of Cave Type
The amount of food supply in caves depends on the
cave type, on surface connections, and on the geo¬
graphic location. Generally, the food supply in tropical
and subtropical caves is greater than in temperate ones
because the biomass in the tropical epigeum is greater
and its production is mostly uninterrupted (Poulson
and Lavoie, 2000). As a consequence, selection pres¬
sure can be expected to be weaker, the evolutionary
rate slower, and the appearance of troglobites not as
fast in such caves compared to caves with low energy
input such as temperate ones. In fact, troglobites are
far more abundant in temperate zones than in the tro¬
pics, and species richness in caves is often correlated
with the amount of available energy.
ENCYCLOPEDIA OF CAVES
ADAPTATION TO LOW FOOD
3
FOOD-FINDING
A variety of morphological and physiological adapta¬
tions and changes in foraging behavior are the basis for
more efficient foraging and increased food-finding abil¬
ity of cave animals. Such alterations are only advanta¬
geous under food scarcity and in darkness. In the case
of high food—prey density or in light conditions, cave
species are inferior to competing epigean relatives.
Appendages and Sensory Equipment
The most obvious morphological alterations in cave
animals are longer legs, antennae, fins and barbels,
or enlarged or flattened heads. If these body parts
bear sensory organs, their enlarged surface can be
correlated to an increased number of chemosensitive or
mechanosensitive organs. As a consequence, an
increased sensitivity to chemical and mechanical stimu¬
lants and changes in foraging behavior are possible.
Cave animals can detect the food faster and at a greater
distance from their bodies than can epigean ones and,
as a side effect, spend less energy for food searching.
All of these characters have been observed in a broad
variety of taxa, from amphipods to crayfish, isopods,
spiders, beetles, fish, salamanders, and more. Cavefish
have been studied most intensively in this respect,
mainly those of the famous North American cavefish
family the Amblyopsidae, the Mexican characid fish
Astyanax fasciatus, and cave salamanders.
In the Amblyopsidae, a positive trend in several of
the specified troglomorphic features progresses from
an epigean species over four gradually more cave-
adapted species (Poulson, 1963). Adaptive alterations
to the cave conditions are correlated with enlarged
associated brain parts, whereas smaller optic lobes
reflect the reduction of eyes as a consequence of dark¬
ness and uselessness. For cave salamanders, the most
likely function of the elongated limbs is to raise the
body and particularly the head above the cave floor to
increase efficiency of the lateral-line system. They also
permit the salamanders to search a larger area per unit
of energy expended and thus increase feeding effi¬
ciency. In interstitial species, the evolution of append¬
age length is different than in cave species. Due to the
small size of the interstitial gaps, they tend to have
shorter appendages and a more worm-like appearance
(Coineau, 2000).
Behavior
Changes in foraging behavior can also increase the
food-finding ability. In the darkness of caves, a food¬
searching behavior concentrated on only the two-dimen¬
sional bottom or other surface areas can be much more
economic in time and cost than a food search in a three-
dimensional space, as exhibited by most surface animals
in light and what they also try to do in darkness. Several
cave animals have abandoned the shoaling or grouping
behavior and adopted a continuous moving mode as a
consequence of darkness and food scarcity in the cave
habitat. They compensate for the optically orientated
and spatially limited food-searching mode of epigean
relatives by covering a greater area using chemo- and
mechanosensors. The amblyopsid cavefish have devel¬
oped a different swimming behavior, referred to as
glide-and-rest swimming. This behavior, also enabled by
the larger fins, not only conserves energy but also results
in a reduction of interference noise for neuromast recep¬
tors, thus improving prey detection.
Other Factors
Most cave animals cope with food scarcity by taking
a wide range of food or exhibiting a different food pref¬
erence compared to surface relatives. Sometimes they
show a dietary shift if one food source becomes scarce.
A higher food utilization efficiency in cave animals as
adaptation to the food scarcity is still not proven.
Finally, the improvement of one feature sometimes may
have more than one positive effect on the cave animals.
The elaboration of the antennae in amphipods not only
enhances food-finding ability, and thus survivorship,
but also improves the mate-finding ability in popula¬
tions with often low densities. Elongated bodies pre¬
sumably facilitate movement through an interstitial
medium.
Back to the Network
An improved, that is, a more efficient, food-finding
ability is adaptive predominantly in patchily food-
limited cave habitats and may be realized through
changes in foraging behavior and improved ability for
sensory orientation. The latter includes improvements
not only in taste and smell senses but also in spatial ori¬
entation, as is required in the darkness of caves. A mul¬
tiplied and improved sensory equipment can be
causally connected with longer appendages and this can
be a result of neoteny. An improved food-finding ability
itself can reduce the general energy demand of cave ani¬
mals or improve the ability of fat accumulation.
GENERAL ENERGY DEMAND
A reduced general energy demand is highly adaptive
in the food scarcity of caves. It reflects a resistance not
only to starvation during periodic food limitation or to
general food scarcity but also to food patchiness, low
ENCYCLOPEDIA OF CAVES
4
A
oxygen content, or other abiotic factors in the cave envi¬
ronment. The energy demand of an animal is usually
quantified by its metabolic rate. Meaningful information
on the metabolic rate is given by the measurement of oxy¬
gen consumption of the entire organism or parts of it. In
addition, indirect parameters such as respiratory fre¬
quency, resistance to anoxia, ability to survive starvation
periods, body composition, growth rate, gill area, or turn¬
over rate of adenosine triphosphate (ATP) have been
used to compare metabolic rates, and in most investiga¬
tions the metabolic rates of the hypogean species were
found to be more or less lower than those of their epigean
relatives.
Aquatic Cave Animals
Not only in caves but also in interstitial habitats,
aquatic animals above all practice striking energy
economy. Several hypogean amphipod, isopod, deca¬
pod, and fish species have been shown to live with
metabolic rates much lower than those of surface rela¬
tives. The most detailed analysis of cave adaptation in
fish (Poulson, 1963) demonstrates a decreasing trend
in the metabolic rate from the epigean species in the
Amblyopsidae over the troglophilic to gradually more
cave-adapted ones. High fat reserves together with
low metabolic rates explain the long survival time of
the most troglobitic amblyopsid species when starved.
However, high fat contents may lead to misinterpreta¬
tions of metabolic rates. Because fat tissue is known to
have a relatively low maintenance metabolism com¬
pared to other tissues or organs, lean body mass or
bodies with comparable fat contents should be pre¬
ferred as a metabolic reference to avoid misinterpreta¬
tions of the metabolic rate. In a variety of the Mexican
cavefish A. fasciatus a very high fat content (Table 1)
suggested a reduced metabolic rate compared to the
epigean relative (Hiippop, 2000; see Table 1). The
recalculation of the metabolic rate, taking fat content
into account, resulted in nearly identical values in both
varieties of the fish species. Although obviously
adapted to a periodically low energy environment (see
"Periodic Starvation" below), as can be seen from the
high fat content, the hypogean A. fasciatus were not yet
able to reduce their energy turnover rate in adaptation
to a general food scarcity.
Terrestrial Cave Animals
Only a few investigations on metabolic rates of ter¬
restrial cave animals exist. Although food scarcity gen¬
erally is even greater in the terrestrial than in the
aquatic cave environment, only a few cave arthropod
species were found to show a tendency toward energy
economy.
Activity
Every activity increases the energy consumption of
animals. The standard metabolic rate (i.e., the lowest
oxygen consumption rate that can be measured during
a test) excludes motion activity and is a measure of the
physiological adaptation of cave animals to food scar¬
city. However, the routine metabolic rate (i.e., the
mean metabolic rate over 24 hours, which includes
spontaneous activity) is a more appropriate index of
actual energy expenditures in nature; it actually may
have the highest rank among the parameters determin¬
ing adaptation to food scarcity in cave animals. The
routine metabolic rate may be reduced in cave animals
due to minimized motion activity, to changed motion
patterns (temporal as well as morphological), to
reduced or no-longer-practiced aggressive and terri¬
tory behavior, or to reduced fright reactions. Actually,
in most cave animals activity is reduced. Although an
increase in food-finding ability in cave animals often
seems to go along with an increase in food-searching
activity, changed motion patterns result in a reduction
of energy expenditure, sometimes to a fantastic exten¬
sion. For example, in the most cave-adapted species of
the amblyopsid fish in North America, over 90% of the
total energy savings by adaptations are based on the
reduced activity (Poulson, 1985).
Excitement and Aggression
Metabolic rates are definitely elevated by an animal's
reaction to disturbance (excitement) and by aggressive
behavior. The standard metabolic rate is elevated by
interior activity or by excitement without expression in
motion activity. The routine metabolic rate increases
due to external activity, including motion activity
resulting from excitement or aggression. In the Mexican
characid fish Astyanax fasciatus the reduction of
TABLE 1 Fat Contents and Condition Factors of the Epigean
and One Hypogean Variety (from La Cueva de El Pachon) of the
Mexican Characid Fish Astyanax fasciatus in Relation to the
Number of Experimental Starvation Days
Days of Starvation
0
109
174
Fat content (% wet body mass)
epigean
9
2
hypogean
37
28
27
Fat content (% dry body mass)
epigean
27
8
hypogean
71
63
62
Condition factor (100 g cm~ 3 )
epigean
2.0
1.6
hypogean
2.9
2.5
1.9
Source: Adapted from Hu'ppop (2000).
ENCYCLOPEDIA OF CAVES
ADAPTATION TO LOW FOOD
5
aggressive behavior as a consequence of the loss of
vision in the darkness of caves was proven. An
increased resistance to disturbance has been shown to
be important for energy economy in the amblyopsid
fish (Poulson, 1963). The generally low standard and
routine metabolic rates of cave amblyopsids and their
resistance to disturbance are interpreted not only as
adaptations to the reduced food supply, by a factor of
about 100 compared to the surface, but also as a by¬
product of relatively stable cave conditions and a gen¬
eral lack of predators in the amblyopsid cave
environment.
The Conflicts of Body Size
The energy reserves of larger animals last longer
and are more resistant to food shortage than are small
animals because the metabolic rate of animals is not
directly proportional to body mass but is related to
mass by the following equation
metabolic rate = am 1 '
where a = intercept, m = wet body mass, and b = mass
exponent/slope smaller than 1 (Withers, 1992).
Consequently, subterranean animals have to resolve
the conflict between two advantages: (1) to be larger
with a lower energy demand per unit mass but a high¬
er one per individual; or (2) to be small, thus requiring
more energy per unit mass but less per animal, and/or
being able to live in crevices.
A special case of subterranean habitat is the intersti¬
tial. In addition to food scarcity it is constrained by the
grain size of the substrate. Interstitial animals are limited
in size and shape due to the small size of the interstitial
gaps and often have shortened appendages, excluding
the posterior appendages, which tend to be elongated
(Coineau, 2000). However, although they have a com¬
paratively higher mass-specific routine metabolic rate
than their surface relatives, their routine metabolic rate
per individual is smaller. Thus, many interstitial forms
may have reduced their body size not only to fit better
into the small crevices but also to cope better individu¬
ally with food scarcity in their special habitat. That the
motion activity, necessary for foraging within the small
crevices, and thus the routine metabolic rate can in turn
be increased in interstitial animals (Danielopol et al.,
1994) has a counterproductive effect.
Ectothermy and Neoteny
Troglobites are exclusively ectotherms. The generally
very low metabolic rates of ectotherms (only 10 to 20% or
even less that of similar sized endotherms) are the basis
for their success in zones characterized by limited
resource supplies, such as shortages in food, oxygen, or
water. Ectotherms can utilize energy for reproduction
that endotherms are forced to use for thermoregulation.
Finally, ectotherms are able to exploit a world of small
body sizes unavailable to endotherms. Body sizes less
than 2 grams are not feasible for endotherms because the
curve relating metabolism to body mass becomes asymp¬
totic to the metabolism axis at body masses lower than 2
grams (Withers, 1992).
Within the ectothermic vertebrates, only fish and
amphibians evolved cave species. Because they are the
largest animals in cave communities, they usually repre¬
sent the highest trophic level in the cave food web and
can survive in large populations only in relatively food-
rich caves. All troglobitic salamanders are aquatic
throughout their life or at an early stage of their devel¬
opment. They often show the retention of larval charac¬
ters, known as neoteny, which enables them to survive
in a relatively less food-scarce aquatic cave habitat
compared to the terrestrial cave habitat (Culver, 1982).
Finally, suppression of the energetically expensive meta¬
morphosis in hypogean salamanders can be interpreted
as an adaptation to general food scarcity.
Hypoxic Conditions
Besides food scarcity, numerous cave or interstitial
species have to cope with temporary, permanent, or pat¬
chy hypoxic conditions. Also, this character of some
cave environments forces reduced metabolic rates and
has been proven in crustaceans and fish. High amounts
of fermentable fuels result in a more sustained supply
for anaerobic metabolism, and glycogen utilization rates
and lactate production rates are significantly lower in
hypogean crustaceans. In contrast to surface species,
several hypogean species have no sharp break in the
oxygen uptake lines under depleting oxygen concentra¬
tions. This absence of a discontinuity in the oxygen
uptake line is called oxyregulation and is considered to
be adaptive in environments characterized by variable
oxygen conditions (Danielopol et al., 1994).
Character Reduction
Many features become reduced during the evolution
of cave animals. This regressive evolution is generally
described as the reduction of "functionless" characters
in cave animals. It not only concerns structural but also
behavioral and physiological traits. Character reduction,
that is, saving energy from not building or not maintain¬
ing useless characters when living in strongly food-lim¬
ited cave environments, should be advantageous for
cave animals: they can transfer the saved energy to the
development or support of other characters or to
growth, reproduction, or survival during starvation
ENCYCLOPEDIA OF CAVES
6
A
periods. There exist a few hints among beetles and spi¬
ders of this possible strategy of cave animals to adapt to
a food-restricted cave environment. Nevertheless, more
often the reduction of characters in cave animals seems
to be the result of accumulated neutral mutations.
Back to the Network
A reduced energy demand is adaptive mainly in those
caves that are generally low in food. A lowered interior
activity, meaning a reduced standard metabolic rate, can
be the result of an increased fright resistance, a lowered
aggressive behavior, and changes in biochemical
mechanisms. Reduced motion activity (i.e., a lowered
routine metabolic rate) can be achieved by means of
reduced body movement to escape (a reduced number
of) predators or for aggression, changes in foraging
behavior and locomotion, and an improved sensory ori¬
entation, resulting in fewer movements for food search¬
ing. In the end, reduced metabolic rates result in a higher
availability of energy for growth and/or a greater resis¬
tance to starvation. Additionally, character reduction,
reduced growth rates, and smaller adult body size have
the ability to reduce energy demand in cave animals. The
reduced energy demand in cave animals can have two
effects. Under the aspect of metabolic span (see
"Longevity" below), a reduction per time is correlated
with a prolonged lifetime combined with iteroparity. On
the other hand, the reduction per individual life enables
higher survival rates of individuals or even the increase
of population size.
LIFE HISTORY CHARACTERS
The extremes of the spectrum of life history adapta¬
tions are characterized as r- and K-selection. Whereas
r-selection (r being the slope of the population growth
curve) means a trend toward high population growth
rate under temporarily good conditions in relatively
unpredictable and changing habitats, K-selection
(K being the carrying capacity of the habitat) can be real¬
ized only in more predictable and stable habitats, and the
appropriate fitness measure is the maximum lifetime
reproduction. K-selected species are characterized by low
or no population growth; they have reached a maximal
K. This situation is connected with fewer but larger and
more nutrient-rich eggs, increased time required for
hatching, prolonged larvae stage, generally decreased
growth rate, delayed and perhaps infrequent reproduc¬
tion, increased longevity, and parental care. Many cave
animals show a couple of these characters, demonstrat¬
ing a trend toward more K-selection in cave species. The
life history of cave animals has been the subject of some
reviews; the main investigations were done on a variety
of invertebrates, particularly crustaceans and arthropods,
and on fish and salamanders (Hiippop, 2000).
Egg Size
Bigger eggs with more energy-rich yolk release lar¬
vae that are bigger at the time of yolk depletion when
they have to start feeding on external food. These lar¬
vae have a bigger head with larger mouth, so they can
start external feeding on a wider spectrum of food par¬
ticles and may have a better chance of survival. For
example, in the Mexican characid fish Astyanax fascia-
tus the eggs of several hypogean varieties and the lar¬
vae of the most adapted variety have significantly
more yolk than those of the commensurate epigean
conspecific relative (Fig. 2 and Fig. 3; FIuppop, 2000).
The cave larvae live longer on their yolk reserve (4.5
versus 3 days), are longer when yolk is depleted (5 ver¬
sus 4 mm), and start storage of adipose tissue already
when fed ad libitum some days earlier than
epigean larvae do (6 versus 9 days).
Bigger larvae may also have a higher resistance to
starvation and a higher mobility for food searching
and for effective escape reactions. In extreme, the trend
toward fewer but bigger eggs in cave animals can
result in a single larva per reproductive season that
possibly never feeds, as in a cave beetle species.
Growth Rate
A reduced growth rate in cave animals is adaptive
to food scarcity because it means a reduced energy
demand per time. More animals can live on a defined
amount of food, or a defined group of animals can sur¬
vive longer on it. A reduction in energy demand per
unit time through a lower metabolic rate, together with a
reduction of absolute and relative costs of reproduction,
can make possible an increase of population density and
hence an increase in the number of females actually
breeding per year.
Longevity
There is evidence that the total metabolic turnover
in a lifetime not only of endotherms but also in
ectotherms is the product of the energy turnover rate
and the duration of life, called the metabolic span.
Generally, lower metabolic rates and slower growth
rates (that is, "slower living") are tied to increased lon¬
gevity. Because the reproductive success of an animal
might be defined by the ability to live long enough to
survive the gap between good years, an increased life¬
time in cave animals is advantageous in a generally
food-scarce environment or in caves where relatively
food-rich reproductive seasons occur irregularly. The
ENCYCLOPEDIA OF CAVES
ADAPTATION TO LOW FOOD
7
increased longevity of cave animals connected with a
delay in maturity and a trend from semelparity to iter-
oparity means that the population is less likely to dis¬
appear in years when food supply is too low to allow
females to produce offspring.
A Case Study
Extremely prolonged lifetimes of more than 150 years
are known among North American cave crayfish.
However, the amblyopsid cavefish, intensively investi¬
gated by Poulson (1963), are the best-known example of
how cave animals adapt their life history to food scar¬
city. Within this group of fish species it is obvious how
cave animals with increasingly slower energy turnover
rates have increasingly prolonged life cycles connected
with many increasingly K-selected features, such as
bigger and fewer eggs with prolonged developmental
Cave fish Surface fish
FIGURE 2 Egg size at an age of 12 hours and larvae size at an
age of 48 hours of a cave variety (from La Cueva de El Pachon) and
the epigean relative variety of the Mexican characid fish Astyanax fas-
ciatus. After Hiippop (2000).
time, prolonged branchial incubation time (=parental
care), bigger larvae at first external feeding, reduced
growth rate, delayed maturity, and multiplied chances
to reproduce with increasing cave adaptation (Table 2).
Population growth rate and population density decrease
with increasing phylogenetic age of the cave species,
and the population structure shifts toward adults
(Poulson, 1963).
Back to the Network
Life history changes toward more K-selected charac¬
ters are correlated with a prolonged lifetime (and conse¬
quently iteroparity) and/or with a shift toward more
adults in the population. Additionally, more ^-selection
may include bigger and fewer eggs and longer incuba¬
tion and brood care, giving the offspring a higher
chance of survival. A reduced growth rate and a smaller
adult size may save energy. A prolonged developmental
period may result in neoteny which in turn can have an
influence on food-finding ability through appendage
lengthening and changes in foraging behavior.
PERIODIC STARVATION^
Besides a general food scarcity, many cave animals
are faced with temporal periodicity of food; hence,
they need an improved ability to survive long periods
of starvation. Seasonality in caves, as already men¬
tioned, is based on periodic flooding or on animals vis¬
iting the cave periodically, such as bats. Normally, this
results in annual cycles.
CO
E
£
CD
E
o
>
o
0.28
0.23
0.18
0.13
0.08
Epigean
2 3 4 5
Hypogean
FIGURE 3 Yolk volume (mean, standard deviation, and range) at an age of 12 hours of five hypogean varieties of the Mexican characid
fish Astyanax fasciatus (from La Cueva Chica, El Sotano de Yerbaniz, La Cueva del Rio Subterraneo, El Sotano de Las Piedras, and La Cueva
de El Pachon) and the epigean relative. After Hiippop (2000).
ENCYCLOPEDIA OF CAVES
8
A
TABLE 2 Some Life History Characters of the Epigean
Amblyopsid Fish Chologaster cornuta Compared to the Troglobitic
Amblyopsis rosae
Ch. cornuta
A. rosae
Egg size (mm)
0.9 to 1.2
1.9 to
2.2
Developmental time
few weeks
5 to 6
months
Larval size (mm) when hatching
3
5
when leaving gill cavity
8
12
Adult size (mm)
23 to 55
36 to 48
Female body mass (g)
0.93
1.25
Adult maturity (year)
1
3
Maximal life span (years)
1
5 to 6
Number of mature ova per female
98
23
Average number of reproduction per lifetime
1
0.6
Maximum number of reproduction per lifetime
1
3
Source: Adapted from Poulson (1963, 1985) and Culver (1982).
Fat Accumulation
The main way to improve the survival capacity
in periodically food-scarce cave environments is the
accumulation of large amounts of adipose tissue during
food-rich seasons. High lipid contents have been
observed not only in cave animals but also in many
surface species subjected to seasonal changes in food
supply. The energy content per gram of lipids is
roughly twice that of proteins or carbohydrates; there¬
fore, fat accumulation is the best way to store energy.
This may be achieved through excessive feeding,
increased feeding efficiency, or improved metabolic
pathways favoring lipid deposition. Cave animals
build up fat reserves during the food-rich season and
store them in their abdominal cavity, in subdermal
layers, intra- or extracellularly in the hepatopancreas
(decapods) or in the muscles, within the orbital sockets
of the reduced eye, or in the more or less reduced
swim bladder (as some fish do). Hypogean amphipods,
decapods, remipedes, collembola, beetles, and several
fish species all over the world have been observed to
be able to survive starvation periods better than their
epigean relatives. Several of them accumulate fat
deposits, sometimes up to huge amounts. They are able
to survive starvation periods from several weeks to one
year (as proven with fish), probably even more. By pro¬
ducing eggs with more yolk, cave animals can enable
their young to better resist short starvation periods.
A hypogean variety of the Pyrenean salamander
Calotriton as per exhibited a hypermetabolism and
higher glycogen in liver and muscles (+25%) and tri¬
glyceride stores in muscles (+50%) in the fed state
than the epigean ones (Issartel et al., 2010). During
experimental fasting the energetic reserves always
remained higher in the hypogean individuals.
In the North American amblyopsid cavefish species
the increasing ability to resist starvation is correlated
with the different stages of various morphological
adaptations (Poulson, 1963).
Individuals of a hypogean variety of the Mexican
characid fish Astyanax fasciatus fed ad libitum in
the laboratory were able to accumulate fat up to 71%
of dry body mass compared to only 27% in the conspe-
cific epigean fish variety (Hiippop, 2000). Not until
after a starvation period of almost half a year did the
condition factor of individuals of the hypogean variety
of A. fasciatus fall below the condition factor of the
epigean fish in an unstarved condition (Table 1).
Energy Demand
A lowered energy demand during starvation can
mainly be attained by the reduction of the routine meta¬
bolic rate in minimizing the motion activity. But also the
reduction of the standard metabolic rate in lowering
internal activity is possible; even changes in biochemical
mechanisms, for example, changes in electron transport
system activity have been shown (Mezek et al., 2010).
In some hypogean crustaceans, the locomotory, ven¬
tilatory, and respiratory rates were drastically lowered
during long-term starvation, whereas surface species
showed lower decreases in these rates and responded
by a marked and transitory hyperactivity (Hervant
et al., 1997). In a state of temporary torpor the hypo¬
gean species were able to survive starvation periods
largely longer than 200 days.
A hypogean variety of the Pyrenean salamander
Calotriton asper exhibited a 20% decrease in oxygen
consumption under experimental fasting, whereas epi¬
gean individuals experienced no significant change
(Issartel et al., 2010).
Back to the Network
An augmented fat content is extremely useful in cave
animals confronted with periodic food scarcity in that it
increases their resistance to starvation. Cave animals
may be able to improve their capacity of fat accumula¬
tion by an improved food-finding ability, perhaps sup¬
ported by a higher food utilization efficiency. Lowered
metabolic rates reduce energy demand during starvation
and thus increase starvation resistance. An increased
starvation resistance possibly also has an influence on
the survival rate of the young.
ENCYCLOPEDIA OF CAVES
ADAPTIVE SHIFTS
9
See Also the Following Articles
Natural Selection
Food Sources
Bibliography
Coineau, N. (2000). Adaptations to interstitial groundwater life.
In H. Wilkens, D. C. Culver, & W. F. Humphreys (Eds.),
Subterranean ecosystems (pp. 189—210). Amsterdam: Elsevier.
Culver, D. C. (1982). Cave life: Evolution and ecology. Cambridge, MA:
Harvard University Press.
Danielopol, D. L., Creuze des Chatelliers, M., Mosslacher, F.,
Pospisil, P., & Popa, R. (1994). Adaptation of Crustacea to
interstitial habitats: A practical agenda for ecological studies.
In J. Gibert, D. L. Danielopol & J. A. Stanford (Eds.), Groundwater
ecology. New York: Academic Press.
Hervant, F., Mathieu, J., Barre, H., Simon, K., & Pinon, C. (1997).
Comparative study on the behavioral, ventilatory, and respiratory
responses of hypogean and epigean crustaceans to long-term star¬
vation and subsequent feeding. Comparative Biochemistry and
Physiology, A118( 4), 1277-1283.
Hiippop, K. (2000). How do cave animals cope with the food scarcity
in caves? In H. Wilkens, D. C. Culver, & W. F. Humphreys (Eds.),
Subterranean ecosystems (pp. 189—210). Amsterdam: Elsevier.
Issartel, J., Voituron, Y., Guillaume, O., Clobert, J., & Hervant, F. (2010).
Selection of physiological and metabolic adaptations to food depri¬
vation in the Pyrenean newt Calotriton asper during cave colonisa¬
tion. Comparative Biochemistn/ and Physiology, A155(l), 77—83.
Mezek, T., Simi, T., Arts, M. T., & Brancelj, A. (2010). Effect of fasting
on hypogean ( Nipliargus stygius) and epigean ( Gammarus fossarum)
amphipods: A laboratory study. Aquatic Ecology, 44(2), 397—408.
Poulson, T. L. (1963). Cave adaptation in amblyopsid fishes.
American Midland Naturalist, 70(2), 257—290.
Poulson, T. L. (1985). Evolutionary reduction by neutral mutations:
Plausibility arguments and data from amblyopsid fishes and
linyphiid spiders. National Speleological Society Bulletin, 47(2),
109-117.
Poulson, T. L., & Lavoie, K. H. (2000). The trophic basis of subsurface
ecosystems. In H. Wilkens, D. C. Culver & W. F. Humphreys
(Eds.), Subterranean ecosystems (pp. 231—249). Amsterdam:
Elsevier.
Sarbu, S. M. (2000). Movile Cave: A chemoautotrophically based
groundwater ecosystem. In H. Wilkens, D. C. Culver, & W. F.
Humphreys (Eds.), Subterranean ecosystems (pp. 319—343).
Amsterdam: Elsevier.
Withers, P. C. (1992). Comparative animal physiology. Fort Worth:
Saunders College Publishing.
ADAPTIVE SHIFTS
Francis Q. Howarth and Hannelore Hoch
Bishop Museum, Honolulu
Museum fur Naturkunde, Berlin
THEORY OF ADAPTIVE SHIFT
For over a century, the evolution of obligate cave
species (troglobites) was assumed to be restricted to
continental regions that had been influenced by glacial
events and that the ecological effects resulting from the
severe climatic fluctuations provided the isolation and
necessary elements to facilitate their evolution. Because
most troglobites appear to have no surviving ancestors,
it was assumed that they evolved allopatrically (i.e., as
a consequence of extinction of their closely related
epigean species); however, it is possible that many
temperate-zone troglobites evolved parapatrically
through adaptive shifts and that their current isolation
from surface relatives occurred after cave adaptation.
In the adaptive shift model, populations of epigean
species invade subterranean habitats to exploit novel
resources, and the drastic change in habitat is consid¬
ered to be the driving force for genetic divergence and
speciation. The presence of exploitable food resources
provides the evolutionary "incentive" for adaptation.
Evolutionary theory predicts that adaptation and new
species are more likely to arise from large, expanding
populations to exploit marginal resources at the edge
of their habitat and to make adaptive shifts to exploit
new environments than are small populations.
Adaptive shifts proceed in a three-step process.
First, a new habitat or resource becomes available for
exploitation. This opening of an ecological niche can
occur either by an organism expanding its range or
by geological or successional processes creating an
opportunity for a resident population. Second, there is
a shift in behavior to exploit the new habitat or food
resource. Third, if the behavioral shift successfully
establishes a new population, natural selection fosters
adaptation. In a classic comparative study of cave and
surface springtails, Christiansen (1965) demonstrated
that a behavioral shift to exploit a new resource or
survive a new environmental stress occurs first, and,
if it is successful, morphological, physiological, and
additional behavioral changes follow. Even if the popu¬
lations remain in contact, the selection pressures
imposed by the different environments and augmented
by environmental stress can force the two populations
to diverge.
Speciation by adaptive shift is parapatric in that adap¬
tive differentiation, accompanied by a reduction in gene
flow, proceeds across a steep environmental gradient
within a contiguous area. If the colonization is successful,
the founding population reproduces and expands into
underground habitats. Maintenance of the barrier to
gene flow could be facilitated by selection against hybrid¬
ization ( e.g ., decreased hybrid viability), assortative mat¬
ing, or spread of the incipient cave population away
from the narrow hybrid zone, thereby reducing the effect
of introgression from epigean individuals. It has been
hypothesized for cave species that selection for novel
mating behaviors may be the principal origin of isolation
(Hoch and Howarth, 1993). Many adaptive shifts
Encyclopedia of Caves.
© 2012 Elsevier Inc. All rights reserved.
10
A
undoubtedly fail; that is, the founding population dies
out. Conversely, interbreeding and selection following a
shift can produce a single more adaptable population
capable of exploiting both environments.
FACTORS UNDERLYING ADAPTIVE
SHIFTS INTO CAVE HABITATS
The trigger for an adaptive shift into caves is
the availability of suitable habitat and exploitable
resources. Only a small percentage of epigean groups
have representatives inhabiting caves, and several
factors appear to be involved in determining whether
or not an organism can take advantage of the opportu¬
nity to exploit caves. These factors often act in concert
but for clarity can be considered either intrinsic or
extrinsic. Intrinsic factors are characteristics inherent
in the organism that allow it to make the shift—for
example, the ability to live in damp, dark habitats.
Extrinsic factors are those imposed on the organism by
the environment.
Intrinsic Factors
Preadaptation
The role of preadaptation has been recognized for
more than a century as one of the principal factors
explaining which taxa have successfully adapted to
live in caves. Most preadaptations simply result
from the correspondence of an organism's preferred
environment with that found in caves. That is, organ¬
isms that have characteristics that allow them to live in
damp, dark, wet-rock microhabitats on the surface
have a better chance of surviving in caves than do
organisms that do not possess these traits. Similarly,
survival is enhanced if their normal food resource
naturally occurs in caves. Nymphs of epigean cixiid
planthoppers are admirably preadapted to caves as
they feed on plant roots and therefore already have
a suite of behavioral, morphological, and physiological
adaptations to survive and feed underground. In the
cave species, these preadapted nymphal characters
have been retained into adulthood so that the adults
are also able to live underground.
Genetic Repertoire
In order to adapt to caves, a population must have
the ability to change. For example, if an organism
is genetically hard-wired to require light, temperature
change, or other environmental cues to complete its
life cycle and reproduce, it is unlikely to colonize caves.
In cixiids, salamanders, and a few other groups, the
adaptive shift to living in caves was greatly facilitated
by neoteny or the retention of nymphal characters into
adulthood. Development of neoteny in salamanders
is believed to involve relatively small changes in their
genes that regulate development. This demonstrates
that a small genetic change can have a large effect on
phenotype.
Founder Events and Subsequent Population Flush
and Crash Cycles
Because only a few individuals from the parent
population make the initial shift, only a subset of the
genetic diversity found in the parent population can
be carried into the founding population. Certain alleles
may be lost, while certain rare ones may become more
abundant. Inbreeding, response to environmental
stress, and expression of previously rare alleles may
result in coevolved blocks of loci becoming destabi¬
lized, thus allowing new combinations of genes. If
the new colony expands rapidly into the new habitat
(e.g., due to abundant food or reduced competition
and predation), the relaxed selection pressure allows
the survival of new mutations and unusual recombi¬
nants that ordinarily would be lost due to low fitness.
Thus, the genetic variability may be quickly reestab¬
lished in the founding population; however, by this
time there may be significant genetic divergence from
the parent population. A subsequent population crash
(as when the new colony exceeds the carrying capac¬
ity) places extreme new selection pressures on the
population. These cycles of expansions and crashes
in founding populations can facilitate adaptation to a
new environment (Howard and Berlocher, 1998).
In parapatric divergence, occasional backcrosses with
the parent population can increase genetic diversity and
provide additional phenotypic variability on which nat¬
ural selection can act. Mixing of the two gene pools by
interbreeding becomes less likely as the new population
expands its range beyond that of the parent population
and adapts to the stresses of the new environment.
Response to Stress
Environmental stresses in a novel habitat may
facilitate adaptive shifts (Howarth, 1993). Organisms
living under environmental stresses often experience
higher mutation rates and display greater phenotypic
and genetic variation. In nature, selection and the
high energetic costs of stress usually reduce this varia¬
tion; however, where sufficient exploitable food
energy compensates for the extra costs of survival in a
novel habitat, some of the enhanced variation may
survive and in time result in a new population
adapted to cope with the stresses. The higher ener¬
getic costs required to cope with stress augment both
natural and relaxed selection pressures to favor the
loss of unused characters, such as eyes and bodily
ENCYCLOPEDIA OF CAVES
ADAPTIVE SHIFTS
11
pigment of troglobites. The lower metabolic and
fecundity rates possessed by troglobites may in large
part be an adaptation to cope with stresses found in
subterranean environments.
Mating Behavior and Hybridization
An important constraint among organisms coloniz¬
ing caves is the ability to locate mates and reproduce
underground. The normal cues used by a species may
not be present or the mate recognition signal system
may be confusing in the unusual environment. For
example, sex pheromones would not disperse in the
same way in caves as they do on the surface. Also, an
animal may have difficulty following the pheromone
plume in a three-dimensional dark maze. The rarity
of troglobites that use airborne sounds to locate and
choose a mate attests to the difficulty of using audible
sound underground and suggests that muteness in
crickets is a preadaptation for colonizing caves.
In addition, environmental stress can reduce sexual
selection and disrupt mating behaviors. Founder
events can exacerbate these effects and the resulting
release of sexual selection pressure may allow hybrid¬
ization with relatives. The maintenance of limited
gene flow between the diverging populations is
believed to be important in providing the additional
genetic variability to accelerate adaptation; however,
the evolution of more appropriate mating behaviors
may be the principal isolating mechanism separating a
newly established cave population from its surface
relatives.
Extrinsic Factors
Presence of Cavemicolous Habitats
Obviously, there must be an opportunity for a
surface population to colonize caves; that is, the sur¬
face population must live in contact with cavernous
landforms. Geological processes (such as lava flows)
can create new cave habitats, ecological succession
or the immigration of a new organism can introduce
suitable new food resources into caves, or a pre¬
adapted surface population can migrate into the cave
area. The Hawaiian Islands formed sequentially in
line, and the successful colonization of each island
after it emerged from the sea had to proceed in an
orderly fashion. Phytophagous insects had to wait
for their hosts to establish and so on. Terrestrial obli¬
gate cave species are not likely to have crossed the
wide water gaps between each island. Even though
suitable lava tubes and cave habitats were available
from the beginning, the evolution of cave species
had to wait for their surface ancestors to become
established and for appropriate food resources to
accumulate in caves. These circumstances provided
tremendous advantages to the first organisms that
could exploit a newly available resource and preempt
the resource from subsequent colonists.
Presence of Exploitable Food Resources
An exploitable food resource is the paramount
prerequisite that allows individuals of a population to
shift into a new habitat or lifestyle. In fact, in most
examples of adaptive shift, both the cave and surface
species feed on the same food resource. Sparse soil
and organic litter with areas of barren rock on the
surface are characteristic of cavernous landforms,
including those created by limestone, lava, and talus.
Where an interconnected system of cave and cave-like
subterranean voids occur, organic material does not
accumulate on the surface but is washed underground
by water, falls underground by gravity, or is carried
underground by living organisms (Howarth, 1983).
Many surface organisms will follow their food under¬
ground and attempt to exploit it. Even though their
food may be abundant in caves, most organisms
adapted to living in surface habitats may become lost
in deep cave passages because their normal cues are
reduced or absent or because they cannot cope with
the harsh environment. Thus, much of this sinking
material is outside the reach of most surface organ¬
isms; however, it becomes a rich reward for any organ¬
ism that can adapt to exploit it.
Environmental Stresses
The environment in deep caves and cave-like voids
is highly stressful for most surface animals, and
relatively few surface species can survive for long
underground. Even though food may be adequate, it is
much more difficult to find and exploit in the complex
three-dimensional dark maze, especially because
many environmental cues (e.g., light/dark cycles, tem¬
perature changes, and air currents) are absent.
Furthermore, the atmosphere is characteristically
above the equilibrium humidity of body fluids, mak¬
ing respiration and water balance difficult to sustain
(Howarth, 1980). The substrate is moist, and the voids
occasionally flood. Rock-lined pools are often pitfall
traps for small animals. Carbon dioxide and oxygen
concentrations can reach stressful levels. In many
cave regions, radiation from decaying heavy isotopes
may exceed hazardous levels. Because of the maze¬
like space, escape to a less stressful environment is
virtually impossible, and cave animals must either
adapt to cope with these and other stresses in situ or
die out.
ENCYCLOPEDIA OF CAVES
12
Ancestral Habitats
Troglobites originated from a variety of damp
surface habitats. Specifically, the ancestral habitats
include soil, leaf litter, mosses, and other damp micro¬
habitats in forests; damp, wet-rock habitats (such as
marine littoral, riparian, and cracks in barren lava
flows); and guano-inhabiting species living in caves
and animal burrows. These source habitats corroborate
the concept of preadaptation.
CASE STUDIES
Hawaii
Twelve arthropod lineages have independently
adapted to caves on two or more islands in Hawai'i, indi¬
cating that cave adaptation is a general phenomenon fos¬
tered by ecological and evolutionary factors. On the
young (less than one million years) Hawai'i Island, some
taxa have undergone (or are undergoing) adaptive shifts
into both caves and new surface habitats in a phenome¬
non called adaptive radiation. Many of these troglobites still
have extant close surface relatives living in nearby habi¬
tats. The better known among these are listed in Table 1
and discussed below. The number of examples will
increase as more taxa are studied. Cave species on the
older islands tend to be relicts today; however, they also
are believed to have originally evolved by adaptive shifts.
Cixiid Planthoppers (Hemiptera: Fulgoromorpha:
Cixiidae)
Cixiid planthoppers exemplify how adaptive shifts
might occur (Fig. 1). The nymphs of nearly all species
in the family live in or close to the soil and feed on
the xylem sap of roots. Cixiid nymphs of surface
species molt to adults, emerge above ground, and live,
feed, and mate on vegetation. Plants living on or near
cavernous landforms often send their roots through
the interconnected system of voids deep underground
following nutrients and percolating groundwater.
Thus, roots are often an important source of food
energy in caves, especially in the tropics and lava
tubes. Nymphs of surface species may migrate along
roots deep into caves from their normal shallow
habitat. These nymphs can exploit deep roots, as the
deep habitat is similar to their normal one except for
the perpetually saturated atmosphere, and for this
stress they already have mechanisms to excrete excess
water as xylem sap is very dilute. When they molt,
the adults may have little chance to find their way to
the surface to reproduce and would eventually die
from the stressful environment; however, if host plant
roots were abundant and large numbers of nymphs
continually made their way underground, an adult
might occasionally find a mate and reproduce fortu¬
itously. Most of these early chance matings, if success¬
ful at all, would produce normal surface adults that
would also become lost. Eventually, some offspring
might acquire traits from random mutation or genetic
recombination that are passed on and that allow
the offspring to survive and reproduce underground
more easily. If this incipient colony successfully estab¬
lishes an underground population, it could more fully
exploit the deeper root resources. In fact, a whole
new habitat and resource would be opened to the
expanding population as it could move throughout
TABLE f Parapatric Cave and Surface Species-Pairs Occurring on Hawai'i Island
Cave Species
Common Name
Surface Relative
Ancestral Habitat
Littorophiloscia species
Slater, Wood louse
Littorophiloscia hawaiiensis
Marine littoral
Oliarus makaiki
Cixiid planthopper
Oliarus koanoa
Mesic forest
Oliarus polyphemus*
Cixiid planthopper
Oliarus species
Rain forest
Oliarus lorettae
Cixiid planthopper
Oliarus species
Dry shrub land
Schrankia howarthi*
Moth
Schrankia howarthi*
Rain forest
Schrankia altivolans*
Dry to wet forests
Nesidiolestes ana*
Thread-legged bug
Nesidiolestes selium
Cryptic habitats in rain forest
Caconemobius varius*
Rock cricket
Caconemobius fori*
Barren lava flows
C. sandwichensis
Marine littoral
Anisolabis howarthi
Earwig
Anisolabis maritima
Marine littoral
Anisolabis hawaiiensis
Barren lava flows?
Lycosa howarthi
Wolf spider
Lycosa cf hawaiiensis*
Barren lava flows
*Represented by several distinct populah
ions or species. Polymorphic cave
populations may represent separate invasions or
could result from divergence of a single lineage after
cave adaptation.
ENCYCLOPEDIA OF CAVES
ADAPTIVE SHIFTS
13
the interconnected voids in the cave region and leave
its surface relatives behind. Adaptation to the new envi¬
ronment might be rapid in the invasive expanding
phase. Subsequently, the newly established subterranean
population will acquire the troglomorphic traits char¬
acteristic for obligate cavernicoles: reduction of eyes,
wings, and bodily pigment. The surface environment on
cavernous landforms provides little suitable habitat for
epigean planthoppers, so that the subterranean popula¬
tion is often larger than surface populations over
caves. This is especially true on volcanoes where erup¬
tions can create vast new habitats for cave populations.
In Hawaii, the principal pioneering plant on young
lava flows is the endemic tree Metrosideros polymorpha
(Myrtaceae), the roots of which are the hosts of Oliarus
polyphemus s.l. (sensu Into) nymphs. Even widely scattered
small trees on 100-year-old flows send abundant roots
deep into caves. O. polyphemus can migrate underground
to colonize new caves from neighboring older lava flows,
and in fact they have been found in caves less than
25 years old.
In Hawaii, the genus Oliarus has undergone exten¬
sive adaptive radiation in surface habitats after the
successful colonization of a single ancestral species,
and about 80 endemic species and subspecies have
been described from all major islands. Eighteen sur¬
face species are known from the island of Hawaii,
where at least three independent adaptive shifts
into subterranean habitats have occurred (Table 1).
The clearest case for adaptive shift among cave cixiids
involves O. makaiki and O. koanoa. O. makaiki has
reduced wings, eyes, and bodily pigment and is
known from a single 1500- to 3000-year-old lava tube
from Hualalai Volcano. Its male genitalia (which
usually provide an excellent means for distinguishing
species in the group) are virtually identical to the fully
winged and eyed epigean O. koanoa. O. koanoa is wide¬
spread in mesic forests on Hawaii. Its nymphs are
frequent in shallow caves and even occur near the
entrance in the same cave with O. makaiki. Adults
of O. koanoa may be accidental in caves, as accumula¬
tions of body parts from thousands of dead indi¬
viduals have been found in dead-end passages in
some caves (Hoch and Howarth, 1999).
The highly troglomorphic O. polyphemus s.l. (Fig. 2)
also occurs in the same cave with O. makaiki, as well
as in virtually every suitable cave containing its host
roots, Metrosideros polymorpha; however, each cave
system appears to harbor a unique population of this
troglobite. Each population differs only slightly in mor¬
phology, but the mating calls are highly distinctive and
probably sufficient to reduce hybridization if the popu¬
lations were to come in contact (Hoch and Howarth,
1993). The ancestor of O. polyphemus s.l. has not been
determined, although it certainly belongs in the
endemic Hawaiian group of species. It is also unknown
whether O. polyphemus s.l. represents a single coloniza¬
tion event in caves with subsequent divergence under¬
ground or whether some populations represent separate
invasions by the same or closely related surface species.
FIGURE 1 Diagrammatic view of an adaptive shift from surface to underground environments by a cixiid planthopper. AC: accidental;
AS: adaptive shift; (e): epigean; TP: troglophile; (tr): troglobite.
ENCYCLOPEDIA OF CAVES
14
A
Adult planthoppers use substrate-borne sounds to
recognize and locate their mates. Among surface spe¬
cies, females feed on the host plant while the males fly
from host to host and initiate a species-specific mating
song. If a receptive female hears an appropriate song,
she will answer with her own song. The male orients
to the call, and the two alternately sing and listen until
the male finds the female, who remains in situ. Studies
have shown that Hawaiian cave Oliarus essentially use
the same mate recognition system: substrate-borne
vibrations, with living roots being an excellent transmit¬
ting medium. In laboratory studies on the cave cixiid
Oliarus polyphemus s.L, the flightless females feed on the
host root and occasionally call. Males are also flightless
and wander from root to root, listening. When a male
hears a female, he answers, and if the song is appropri¬
ate the two sing and listen alternately until the male
locates the stationary female. Although the behavior
has not been confirmed in natural settings in caves, it is
believed that this switch in behavior occurred as an
adaptation to the subterranean environment. Singing is
energetically costly, and males expend extra energy
finding roots in the dark, three-dimensional maze. Also,
predators can use these songs to locate prey, and males
would be especially vulnerable if they had to sing on
every root they found; therefore, males would be more
likely to find a female if they did not have to initiate
calling. Females conserve energy by staying in place
and feeding, and they produce fine wax filaments that
deter predators (Hoch and Howarth 1993).
Terrestrial Isopods (Crustacea: Isopoda:
Philosciidae)
Species in the genus Littorophiloscia are marine
littoral and inhabit tropical shorelines worldwide. One
FIGURE 2 Adult female of Oliarus polyphemus. Photo by W.P. Mull.
Used with permission.
epigean species, L. hawaiiensis, is endemic to Hawai'i
and is so far known only from littoral habitats on the
islands of Hawai'i and Laysan. Surprisingly, an unde¬
scribed troglobitic species also lives on Hawai'i Island.
The cave form is remarkably similar in morphology
to L. hawaiiensis, differing only in the troglomorphic
characters (reduced eyes and pigment) displayed by
the cave population. The males even share the same
distinctive sexual characters. Results from a recent
molecular phylogenetic study showed that L. hawaiien¬
sis and the cave Littorophiloscia are distinct species that
diverged from each other most likely by an adaptive
shift (Rivera et al., 2002). The initial adaptive shift
must have been from salty marine littoral habitats to
freshwater terrestrial habitats, possibly along shore¬
lines of coastal springs in young lava flows.
Suitable freshwater shorelines are not well developed
on Hawai'i Island, but lava tubes would offer a vast
new habitat. The complex three-dimensional system of
anastomosing voids characteristic of young basaltic
lava flows provides an immeasurably large habitat for
cave animals. In contrast, L. hawaiiensis is restricted to
a narrow band along the coast where it lives under
rocks resting on soil. Thus, an adapting cave popula¬
tion would quickly move away from contact with its
ancestral population and greatly exceed it in size.
A possible alternative explanation is the classical
model of cave adaptation in which the coastal species
became extinct, stranding an incipient cave population;
however, this model still requires that the adaptive shift
to freshwater (and possibly to caves) had to occur
before extinction, and that circumstance is precluded in
the classical model. This scenario also requires a subse¬
quent colonization by the halophilic littoral species that
secondarily reestablished parapatry.
Crickets (Orthoptera: Gryllidae)
On the island of Hawai'i, the marine littoral rock
cricket Caconemobius sandwichensis has colonized both
unvegetated lava flows and caves, both of which are
barren-rock habitats like its ancestral home. Several
distinct populations of rock crickets inhabit each of
these inland environments, but it is unknown whether
these represent multiple adaptive shifts from the
marine littoral habitat or result from divergence after
colonization of the new habitat, or both. Even though
the marine littoral and lava flow cricket (C. fori) species
are distinct morphologically, a hybrid zone between
the two was found around a pool of brackish water in
a deep crack where a young lava flow entered the sea,
suggesting that the adaptive shift is ongoing. C. fori is
a nocturnal scavenger on very young (i.e., one month
to about a century old) lava flows on Kllauea, disap¬
pearing when plants colonize the flow. It hides in deep
ENCYCLOPEDIA OF CAVES
ADAPTIVE SHIFTS
15
cracks and caves during the day where it overlaps
with several distinct cave populations or species.
Additional recognizable populations of lava crickets,
resembling C. fori, occur allopatrically on Mauna Loa
and Hualalai volcanoes, also on Hawaii Island.
Up to three morphologically, behaviorally, and
physiologically distinct populations of cave crickets
can be found in larger caves, and individual caves
may harbor a unique population of one or more of
these forms. The total diversity of cave-adapted rock
crickets on Hawaii is astounding, especially given the
young age of the island. These multiple simultaneous
adaptive shifts into such different habitats from
the same ancestor further corroborate the contention
that cave adaptation via adaptive shifts is a general
phenomenon.
A possible scenario for the origin of troglobitic rock
crickets and wolf spiders on the island of Hawaii
might proceed like this. Marine littoral crickets could
have been among the earliest colonists of newly
emerging islands, because all that they required was
already present: a rocky shoreline and ocean-derived
flotsam. Barren lava was the original terrestrial habi¬
tat, and sea birds would have sought these new
islands for nesting, presenting an inviting habitat for
the littoral crickets to shift to the land where they
could feed on wind-borne debris. Subsequent accu¬
mulation of organic material underground would
have allowed the colonization of caves by either the
seacoast or lava cricket preempting some resources,
possibly before many plants and potential competitors
arrived on the island.
Wolf Spiders (Arachnida: Lycosidae)
The big-eyed wolf spiders (family Lycosidae) live
on barren landscapes in many parts of the world and
are good long-distance dispersers. They would also
have been among the earliest colonists on emerging
islands in Hawai'i. There are several barren-ground
populations of an unidentified species of Lycosa
(possibly L. hawaiiensis) on Hawai'i Island that differ in
behavior, morphology, and color pattern, but surpris¬
ingly they can hybridize. They occupy a range of habi¬
tats from hot, dry barren lava on coastal plains to the
freezing stone deserts above 4000 meters on Mauna
Kea and Mauna Loa. On young lava flows, they
prey on Caconemobius lava crickets. One member of
this group followed its prey underground, becoming
the anomalous small-eyed, big-eyed wolf spider
(Lycosa howarthi).
The lycosid wolf spiders are characterized by the
possession of four huge eyes (for spiders) and four
smaller eyes, and they are among the better-sighted spi¬
ders. The troglobitic species has six to eight vestigial
eyes, yet differs from the lava flow species only by the
characters associated with cave adaptation. As noted,
both the epigean spider and cricket were early colonists
on Hawai'i, and possibly the spider was able to pre¬
empt the cave habitat because it was among the first
predators to have the opportunity to exploit the young
caves on the island.
Moths (Lepidoptera: Noctuidae)
Recent molecular analysis of mitochondrial
(mtDNA) and nuclear (nDNA) of cave and surface
Schrankia has documented another example of ongoing
adaptive shifts into caves (Medeiros et ah, 2009). Based
largely on wing color pattern and external morphol¬
ogy, five endemic species had been described from
surface habitats in Hawai'i. Three species were
recorded from both Hawai'i and Maui islands. In addi¬
tion, two distinct forms were known from caves on
Hawai'i and Maui islands: a pale, small-winged, and
vestigial-eyed cave form and a drably patterned, larger
winged, sighted form found in the twilight and
entrances of caves. The two were believed to be sepa¬
rate species, but intermediate forms occurred in larger
caves, suggesting hybridization. The DNA analysis
revealed a different and more interesting story. All
five surface species were shown to belong to a single
highly polymorphic species Schrankia altivolans, and all
cave populations belong to another polymorphic spe¬
cies S. howarthi. The two species are closely related but
can also be separated by the form of the male genitalia
as well as wing pattern. The DNA sequences confirm
that hybridization occurs between the cave and twilight
zone morphs. However, they also show that S. howarthi
is segregating into distinct geographic populations. For
example, the Maui S. howarthi form a distinct group
within the Hawai'i Island group, indicating that
S. howarthi arose on Hawai'i and subsequently dispersed
to Maui where it independently became cave-adapted.
The west Hawai'i and montane Mauna Loa populations
each also form distinct groups, whereas populations
from southeast Hawai'i are more diverse and include
members related to the other groups. The latter happen¬
stance might be explained by its youth as the caves sam¬
pled are all less than 700 years old, and nearly the entire
surface of Kllauea is probably less than 10,000 years old.
Surprisingly, the montane S. howarthi group is closer to
S. altivolans than it is to the remaining S. howarthi. This
might result from lineage sorting or past hybridization,
but it also suggests independent adaptation to caves.
Schrankia larvae feed on succulent new growth of
roots and occasionally on rotting organic matter, and
thus are preadapted to live in caves. Like Oliarus, cave
Schrankia have adapted to exploit plant roots in the
vast system of voids in young lava flows. However,
Schrankia use close-range pheromones to locate mates,
and unlike Oliarus, cave-adapted Schrankia are not
ENCYCLOPEDIA OF CAVES
16
A
fully isolated reproductively from their surface rela¬
tives. Volant males would be able to find mates by fol¬
lowing the female scent plume at least in larger cave
passages. Pheromone communication probably would
be less effective within the smaller voids. By necessity,
all specimens used in the DNA study came from acces¬
sible cave passages. Moths living in the more isolated,
inaccessible smaller voids might include a greater pro¬
portion of cave-adapted individuals than the popula¬
tions sampled.
Other Islands
Adaptive shifts are best known from islands in large
part because of their youth and isolation. Thus, many
diverging species pairs are still extant, and their evolu¬
tionary history has not been obscured by geological
or ecological events. Examples of troglobite evolution
by adaptive shift are known from the Galapagos,
Canary, and Greater Antilles islands. In the Galapagos,
at least ten species of troglobitic terrestrial invertebrates
still live parapatrically with their putative epigean
ancestors (Peck and Finston, 1993). At least seven extant
parapatric pairs of troglobitic and epigean species
are known from the Canary Islands (Oromi et ah, 1991).
Some of these, such as the Dysdem spiders (Arnedo
et ah, 2007), Trechus ground beetles (Contreras-Diaz
et ah, 2007), and cockroaches, represent multiple inva¬
sions into caves, in parallel with Hawai'i.
Continents
Clear examples of adaptive shifts in continental caves
are relatively rare, possibly because the great age of
these systems has obscured the evolutionary history of
the taxa involved or because the phenomenon has not
generally been considered; nevertheless, there are a few
cases. Perhaps the best documented is the detailed
study (Culver et ah, 1995) done on the spring- and cave-
inhabiting amphipod Gammarus minus, which occurs in
freshwater basins in temperate eastern North America.
That study demonstrated that several cave populations
were derived from independent invasions from springs
and that some hybridization had occurred; however, the
constraints imposed by the dynamics of water flow and
other factors may have isolated some populations from
each other. In spite of considerable genetic and morpho¬
logical diversity, these populations are all considered one
species. In parallel with some Hawaiian forms, the defini¬
tion of what constitutes a species is sometimes not clear.
Adaptive shifts have been proposed to explain the origins
of cave species in North Queensland, Australia, but the
few phylogenetic analyses completed to date have not
supported the contention.
CONCLUSIONS
In cavernous regions, organic energy is continually
being transported into an extensive system of subterra¬
nean voids, as evidenced by the characteristic presence
of areas with a barren exposed rocky surface. Epigean
species can exploit food resources on and near the sur¬
face, but increased environmental stress levels and
absent or inappropriate cues prohibit their access to this
resource in the deeper voids. Animals following their
normal food deeper may eventually become lost and be
unable to return to the surface. Surface species continue
to be frequent accidentals in caves and provide a large
proportion of the food to the cave-adapted predators
and scavengers. Over time, a few of these accidentals
may eventually survive and establish a new population.
A major initial factor facing a new cave population may
be the ability to locate a mate and reproduce under¬
ground. Once a population is able to live and reproduce
underground, a large new habitat would be opened up
to it, and the population could expand rapidly.
Evolution of troglobites by adaptive shift may be a com¬
mon phenomenon; however, many examples of adaptive
shift remain unrecognized, especially because the phe¬
nomenon is not generally considered. Other scenarios
(such as allopatric evolution, which appears to be sup¬
ported for many temperate troglobites) are also conceiv¬
able. The adaptive shift model and the classical theory of
troglobite evolution may not be mutually exclusive.
THE FUTURE
Adaptive shifts offer extraordinary opportunities for
research. Comparative studies of closely related pairs of
species adapted to such different environments as caves
and surface habitats should provide better understanding
not only of the processes of cave adaptation but also of
selection and species formation in general. Molecular sys-
tematics will clarify the evolutionary history of taxonomic
groups as well as elucidate the genetic basis of adaptive
shifts. Phylogenetic analyses can elucidate the evolution¬
ary transformation of specific morphological, behavioral,
physiological, and ecological traits that occurred during
the process of adaptation to novel habitats. The remark¬
able divergence from their epigean relatives displayed by
troglobites makes such studies especially interesting.
Comparative physiological experiments can potentially
dissect out for study the specific mechanisms involved in
coping with a variety of environmental stresses. For
example, terrestrial cave species live in an air-filled
aquatic environment, while the close relatives of some
live in desert-like environments. These species pairs are
ideal models for studies on the physiological mechanisms
involved in regulating water balance.
ENCYCLOPEDIA OF CAVES
ANCHIHALINE (ANCHIALINE) CAVES AND FAUNA
17
See Also the Following Article
Root Communities in Lava Tubes
Bibliography
Arnedo, M. A., Oromi, P., Murria, C., Macias-Hernandez, N. E., &
Ribera, C. (2007). The dark side of an island radiation: Systematics
and evolution of troglobitic spiders of the genus Dysdera (Araneae,
Dysderidae) in the Canary Islands. Invertebrate Systematics, 21,
623-660.
Christiansen, K. (1965). Behavior and form in the evolution of cave
Collembola. Evolution, 19, 529—537.
Contreras-Diaz, H. G., Moya, O., Oromf, P., & Juan, C. (2007).
Evolution and diversification of the forest and hypogean ground-
beetle genus Trechus in the Canary Islands. Molecular Phylogenetics
and Evolution, 42, 687—699.
Culver, D. C., Kane, T. C., & Fong, D. W. (1995). Adaptation and
natural selection in caves: The evolution of Gammarus minus.
Cambridge, MA: Harvard University Press.
Hoch, H., & Howarth, F. G. (1993). Evolutionary dynamics of
behavioral divergence among populations of the Hawaiian cave¬
dwelling planthopper Oliarus polyphemus (Homoptera: Fulgoroidea:
Cixiidae). Pacific Science, 47, 303—318.
Hoch, H., & Howarth, F. G. (1999). Multiple cave invasions by spe¬
cies of the planthopper genus Oliarus in Hawaii (Homoptera:
Fulgoroidea: Cixiidae). Zoological Journal of the Linnean Society,
127, 453-475.
Howard, D. J., & Berlocher, S. H. (Eds.), (1998). Endless forms: Species
and speciation New York: Oxford University Press.
Howarth, F. G. (1980). The zoogeography of specialized cave animals:
A bioclimatic model. Evolution, 34, 394—406.
Howarth, F. G. (1983). Ecology of cave arthropods. Annual Review of
Entomology, 28, 365—389.
Howarth, F. G. (1993). High-stress subterranean habitats and evolu¬
tionary change in cave-inhabiting arthropods. American Naturalist,
142, S65-S77.
Medeiros, M. J., Davis, D., Howarth, F. G., & Gillespie, R. (2009).
Evolution of cave living in Hawaiian Schrankia (Lepidoptera:
Noctuidae) with description of a remarkable new cave species.
Zoological Journal of the Linnean Society, 156, 114—139.
Oromi, P., Martin, J. L., Medina, A. L., & Izquierdo, I. (1991). The
evolution of the hypogean fauna in the Canary Islands. In E. C.
Dudley (Ed.), The unity of evolutionary biology (Vol. 1). Portland,
OR: Dioscorides Press, pp. 380-395.
Rivera, M. A. J., Howarth, F. G., Taiti, S., & Roderick, G. K. (2002).
Evolution in Hawaiian cave-adapted isopods (Oniscidea:
Philosciidae): Vicariant speciation or adaptive shift? Molecular
Phylogenetics and Evolution, 25, 1—9.
Peck, S. B., & Finston, T. (1993). Galapagos islands troglobites: The
questions of tropical troglobites, parapatric distributions with
eyed-sister-species, and their origin by parapatric speciation.
Memoires de Biospeologie, 20, 19—37.
ANCHIHALINE (ANCHIALINE)
CAVES AND FAUNA
Boris Sket
Univerza v Ljubljani, Slovenia
Anchihaline (or anchialine) habitats are water bodies in
hollows along the sea coasts where the influence of the
sea may be felt and which are inhabited by some subter¬
ranean species. This usually indicates an underground
connection of the cave or pool with the sea. Such a habi¬
tat may contain seawater, but it primarily has layers of
different brackish salinities. In exceptional cases, it has
freshwater but is inhabited by some animals of the
recent marine provenance (Sket, 1996). We must, how¬
ever, exclude genuine sea caves, "spaces ... containing a
high number of speleophilous forms of the rocky littoral
with exclusion of any speleobiotic species" (Riedl, 1966).
In principle, anchihaline habitats can occur everywhere
along the coast, but hardly any have been noted and
investigated outside the tropical or warm moderate cli¬
matic zones. Geologically, two very different types of
coasts are often of interest. Some anchihaline caves are in
karstified limestones, while others are in lava fields.
Because the sea level along most coasts during the Ice
Ages was approximately 100 meters lower than it is now,
the coasts may be hollowed out to such a depth. They
often contain dripstone formations (such as stalactites and
stalagmites), indicating their own previous continental
and terrestrial nature. The limestones may be of different
ages; some being old and primarily compact, with others
being quite recently formed and still primarily porous
coral limestones or little transformed coral reefs. Similarly,
anchihaline habitats are also formed in lava flows, either
in systems of cracks or in the tunnel-shaped lava tubes.
GEOGRAPHY AND SOME
HISTORICAL DATA
Although well known since ancient times, the so-
called blue caves in the Mediterranean ( e.g ., Capri,
Italy; Bisevo, Croatia) were never investigated speleo-
biologically, as their morphology does not promise
extensive anchihaline habitats in them. A number of
localities with anchihaline caves have been discovered
and investigated along tropical seas and in the Medi¬
terranean. Since the middle of the nineteenth century,
brackish anchihaline habitats of some fish and shrimp
species in Cuba have been known. Another historical
site of this kind is Jameos del Agua, a segment of a
lava tube on the island of Lanzarote (Islas Canarias).
The twentieth century began with the discovery of
interesting fauna in some south Italian caves. Impor¬
tant anchihaline caves were later sampled in Libya and
in islands of the Caribbean.
Riedl (1966), who systematically studied sea caves
of the Mediterranean, conceived the concept of the
marginal cave (Randhohle in German), which was similar
to the present-day concept of the anchihaline cave.
Riedl defined the marginal cave by its position at the
sea coast, the presence of brackish water, and the pres¬
ence of troglobiotic species in it, but he also supposed
that the origin of such caves was due to wave action
and their gradual loss of connection to the sea. Thus,
Encyclopedia of Caves.
© 2012 Elsevier Inc. All rights reserved.
18
A
they should be a vector of the stygobiotic fauna origi¬
nating in the sea and making it limnic.
An important study by Holthuis (1973) followed
Riedl's studies and dealt with a number of shrimp spe¬
cies, including the now well-known red shrimps, as
well as with a new concept of anchihaline pools that was
free of suppositions about the past and the future of the
habitat and therefore easier to apply. However,
Holthuis obtained his animals from open pools in
Hawaii and on the Red Sea coast, which contained
some troglomorphic animals along with a very scarce
selection of normal marine animals. In the period that
followed. Stock and his numerous colleagues investi¬
gated a number of anchihaline caves in the Caribbean,
mainly in regard to their faunistic and taxonomic fea¬
tures but also biogeographically. This region appeared
to be particularly rich in such caves and fauna
(e.g., Stock, 1994). Stock suggested and substantiated
the spelling change of anchialine to anchihaline.
Sket studied the caves along the Adriatic coast and
gave the first detailed quantitative data about ecological
conditions in them (Sket, 1986). Later, he also specified
the concept of the anchihaline cave habitat. Iliffe (2001)
is noted for his sampling of a great number of caves in
the Caribbean region as well as in the tropical islands of
the Indo-Pacific. His studies revealed a high number of
new animal species representing an array of groups. In
the meantime, a number of other field researchers and
taxonomists were either sampling caves more locally or
describing and studying collected animals.
Yager (1981) discovered the new crustacean group
Remipedia, probably the most astonishing discovery in
anchihaline waters. About the time of this discovery,
research intensified into the blue holes and inland
anchihaline caves in the Bahamas. Por, Fosshagen,
and Boxshall studied Copepoda and Tantulocarida;
Kornicker and Danielopol studied Ostracoda; Stock
and Holsinger studied Amphipoda; Holthuis and
others studied Caridea; and Stock and Botosaneanu
studied Isopoda. Jaume noted euhaline anchihaline
caves in the Mediterranean (e.g., Jaume and Boxshall,
1996), and Por described hyperhaline pools in the
Arabian Peninsula. A number of authors discussed the
models and hypotheses regarding the colonization of
subterranean habitats by animals in which the anchi¬
haline caves played a non-negligible role.
CAVE MORPHO LOGY AND HYDROLOGY
(See Por, 1985; Sket, 1986, 1996; Iliffe, 2001.) The
morphology of the anchihaline caves and other cavities
containing anchihaline habitats is of low importance
for their fauna; however, it may be extremely impor¬
tant for its accessibility (particularly for the time before
SCUBA techniques were introduced) and to gain an
understanding of its composition and distribution, as
well as relations between species. Horizontal caves
that are either filled with water along the entire corri¬
dor or that exhibit series of pools with different con¬
nections between themselves and with the sea are
comparatively scarce. Such caves can be either karst
caves or lava tubes. Horizontal coastal caves may,
however, be very extensive inundated cave systems,
such as those being investigated along the Mexican
coasts.
More common are shallow vertical caves. They may
be of small dimensions like the natural wells on the
Adriatic islands in Croatia, which are sometimes artifi¬
cially elaborated as watering places for sheep. Similar
are grietas in some Caribbean islands or cracks in lava
fields of the Galapagos; however, accessible vertical
parts are always connected with horizontal corridors or
fissures that may connect groups of caves. Cenotes have
large dimensions and are found mainly in Mexico's
Yucatan Peninsula; these, again, are often extended by
systems of horizontal corridors. An important character¬
istic of such habitats is that a significant part of the
aqueous body is illuminated by daylight.
Even more open are anchihaline pools in lava or in
coral reefs. Such pools must be connected to the sea
and may be interconnected by systems of more or less
dark corridors or fissures, although the open space may
prevail. Such pools are particularly well developed in
Hawaii.
Due to a U-shaped tube effect (phreatic loop), in
caves close to coasts under the influence of tides, the
tides may also be felt underground, although with a
certain delay and decline. This means that in the mild
cases, such as in the Adriatic coastal caves, the entire
stratified column of water moves twice a day up and
down, carrying the floating organisms with itself and
remarkably changing the life conditions for substra¬
tum-bound animals, particularly in the reach of the
halocline. The tides in the open Adriatic are only
approximately 0.5 meters, but more extreme conditions
may occur elsewhere. In the Bahamian blue holes, the
effects of very important tidal currents (dangerous for
a diver) have not yet been studied for their influence
on living conditions. In the horizontal St. Paul's
Underground River corridor (Palawan, Philippines),
the seawater moves with the tides for some kilometers
in and out along the bottom. It is difficult, however,
to imagine that any anchihaline fauna exists in such
a cave.
The aqueous body is seldom purely marine (euha¬
line) but such are present in some deeply set subma¬
rine caves that have to be very extensive to contain
any troglobionts. However, at least the water in lower
portions of the cave systems in the underground of the
ENCYCLOPEDIA OF CAVES
ANCHIHALINE (ANCHIALINE) CAVES AND FAUNA
19
Bahamas (Porter, 1993), as well as some caves in the
Mediterranean islands (Jaume, 1996) is euhaline. When
influenced by freshwater, the situation may be very
different for a cave on a small island than one that is
on the mainland coast. In any case, the relation
between the fresh and marine part of the column is
governed by the Ghyben—Herzberg principle. In a
porous aquifer, if the freshwater overlies seawater, the
interface will be below the sea level approximately 40
times the height of the water above sea level. On an
island, the freshwater body will take on the form of a
lens (Ghyben—Herzberg lens), the depth of which
depends on the precipitation at the site and the speed
of the water outflow toward the sea. The shape of the
freshwater lens and the height of the underlying layers
of seawater will change slowly during the year and
from year to year. Less predictable and more complex
is the situation on the coast of the continent (or of a
larger island), where canalized (thus concentrated)
water inflow from inland may be the result of net-like
systems of corridors. This may complicate the salinity
distribution and will sometimes change the salinity
relations very rapidly.
The transition from euhaline to limnic salinities in
the column is never perfectly gradual. The depth of
very gradually stratified or even homogeneous waters
is regularly divided by a halocline that can be a few
decimeters to a couple of meters thick. The halocline is
a layer where the salinity abruptly changes (sometimes
from polyhaline to nearly limnic). In genuine caves—
such as in caves in general—the water temperature
may be constant and equal throughout the column,
approximately equal to the local yearly average. If the
water body is exposed to the surface temperature
influences, this may also cause temperature stratifica¬
tion, changing its direction throughout the year.
During the winter in moderate climates, the surface
cooling in freshwater lakes causes the loss of stratifica¬
tion and penetration of cooler (i.e., heavier) and well-
aerated surface water toward the bottom. In anchihaline
waters, the density changes caused by varying tempera¬
tures are much lower than the density differences
caused by a small salinity difference. With cooling from
20°C to 4°C, the density of pure water changes from
998 to 1000 g L -1 , while with the increase of salinity
from 10 to 20 parts per thousand, the density changes
from 1009 to 1019 g L _1 . This makes the water column
in mixohaline anchihaline caves very stable and pre¬
vents mixing of layers, with consequences as explained
below.
Although the source of organic input can appear
rather obvious, some sources have also been traced by
sophisticated isotope methods (Pohlman et ah, 2001).
In caves of Mexico, the soil particles may originate
from a tropical rain forest, while in the Mediterranean
these particles may be from the poor remains of soil in
deep karren. The use of the holes as watering places,
however, may cause the introduction of very high
quantities of debris which is deposited on the bottom.
Another source of organic matter is production in illu¬
minated parts of caves. Rich planctonic algal commu¬
nities may be present there, while rocky walls can also
be overgrown by microscopic algae. The virtual contri¬
bution of autotrophic bacteria has only been supposed.
Rich sources of oxidizable substances can be used in
such a way. Some nitrate production (which may be
by nitrite oxidation) has been noted within the halo¬
cline in a Mexican cave. On the other hand, remarkable
hydrogen sulfide (H 2 S) concentrations in many anchi¬
haline caves are a strong basis for chemoautotrophy;
clouds of bacteria within the halocline and mats cover¬
ing the bottom have been observed in some caves, and
the sulfide-oxidizing mixotrophic bacteria of the gen¬
era Beggiatoa and Thiothrix have been identified.
Wherever the input of organic substances is remark¬
able, they are deposited on the bottom where they
putrefy, often causing total consumption of oxygen and
an accumulation of hydrogen sulfide. If communication
with the open sea is feeble, a thick lower layer of the
water may be infested by this gas. A 250-mmol L -1
concentration was established in a Mexican cave, while
such remarkable concentrations have only been con¬
firmed by smell in Adriatic caves. Such an accumula¬
tion can be identified up to the water surface in
extreme cases.
The oxygen that is consumed in cave waters must
be renewed from air on the surface, by an influx of
precipitation water and water jets from inland, at the
surface, or by mixing or diffusion from the open sea in
deeper layers. Therefore, the surface layer is usually
comparatively well aerated while the oxygen concen¬
tration diminishes toward the halocline; below the hal¬
ocline it may rise again or decrease further. It is the
rule that dissolved oxygen does not reach its saturation
in anchihaline waters. Its depletion may be remark¬
able, probably total, although no analysis until now
has been accurate enough to establish this. Such are
the conditions close to the debris-covered bottom
under a deep layer of hydrogen sulfide—rich water. In
some other caves, the oxygen saturation is the lowest
or close to zero within the halocline, the most
stable water layer. Characteristically, the pH has also
always been lowest in the halocline when measured.
NONCRUSTACEAN GROUPS IN THE
ANCHIHALINE FAUNA
Sponges are represented very richly in true marine
caves, sometimes with some deep-sea elements, such as
ENCYCLOPEDIA OF CAVES
20
A
the calcified Calcarea pharetronida. This is particularly
evident in some sack-like Mediterranean caves that cap¬
ture and hold cold water throughout the year and also
harbor small glass sponges (Hexactinellida) and the
exceptional carnivorous Cladorhizidae ( Asbestopluma
hypogea Vacelet and Boury-Esnault) (Vacelet et al.,
1994). Some species are claimed to be true troglobionts
of anchihaline caves, such as the haplosclerid Pellina
penicilliformis Van Soest et Sass in marine caves in
the Bahamas and the axinellid Higginsia ciccaresei Pansi-
ni et Pesce in southern Italy; the latter is a typical
anchihaline animal living in brackish water.
Very little is known about gastropods in anchihaline
caves. The presence of the tiny and mostly interstitial
marine Caecum spp. is most probably in connection
with some sand deposits within the cave. Only some
of the apparently hydrobioid snails in Adriatic caves
have yet been identified.
Among the few filter feeders are some tube worms
occurring quite regularly at high salinities, such as
Filogranula annulata (O.G. Costa) in Adriatic caves. Some
stygobiotic errant polychaetes have been described, such
as, for example, Gesiella jameensis (Hartmann—Schroeder)
from the Canaries. Probably more sediment-dependent
are the tiny members of the family Nerillidae. It is
remarkable that the Dinaric cave tube worm ( Marifugia
cavatica Absolon and Hrabe) has never been found in a
cave within the influence of the sea. Some oligochaetes
occurring in such caves do not seem to be closely related
to them.
Some species of sessile chaetognaths (family Spadel-
lidae) occur in euhaline anchihaline waters, and at
least the eyeless and pigmentless Paraspadella anops
Bowman et Bieri from the Bahamas is supposed to
be stygobiotic; some other species from caves are
oculated.
Fishes are among the longest known anchihaline
animals; approximately 10 anchihaline species are
known today. The family Ophidiidae is exapted for
such habitats by their shy behavior, nocturnal habits,
and/or deep-sea prevalence. The stygobiotic Lucifuga
subterraneus Poey and Stygicola dentatus Poey have
been recognized in Cuban grietas and similar localities
since the middle of the nineteenth century. New,
related species have recently been found in the region.
The stygobiotic and euryhaline Ogilbia galapagosensis
(Poll et Leleup) from underground waters of the
Galapagos Islands has variably reduced eyes and a
closely eyed relative in the species O. deroyi (Van Mol),
which dwells in littoral crevasses. In Yucatan caves,
the ophidiid Ogilbia pearsei (Hubbs) is a neighbor of
Ophisternon infernale (Hubbs), representing another
family of often cryptically living fishes, Synbranchidae.
The synbranchid Ophisternon candidum (Mees), at up to
370 mm in length, is probably the longest cave animal.
It inhabits brackish waters in northwestern Australia,
accompanied by the gobiid Mylieringa veritas Whitley.
It is remarkable that related (congeneric) species occur
in anchihaline caves in mutually very remote areas.
ANCHIHALINE CRUSTACEA
Generally, Crustacea are by far the richest and most
diversely represented group in subterranean waters,
and the same holds for anchihaline cave waters. The
predominance of crustaceans may also be felt within a
particular cave where the specimen number or the bio¬
mass relation between crustaceans and noncrustaceans
may be 10:1 or higher. Among Crustacea, by far the
richest in species are Amphipoda and Copepoda,
although Decapoda (due to their size) are probably the
most obvious. For Remipedia, the anchihaline caves
are the only known habitat, and it seems to be the
central one also for Thermosbaenacea.
Remipedia are crustaceans with by far the most ple-
siomorphic trunk and appendages but with a highly
specialized (and diversified) anterior end of its body, a
very nice case of mosaic evolution. They have up to now
only been found in anchihaline caves in euhaline salin¬
ities, or close to them, usually below the halocline in
poorly aerated water (typically less than 1 mg L 1 oxy¬
gen). Remipedia are hermaphrodites. Up to 45-mm-long,
centipede-shaped swimmers have been found in the
tropical belt connecting Mexico, the Bahamas, and the
Canaries and along the northwestern Australian coast;
they seem to be absent in the Mediterranean. In some
Bahamian caves, a number of species may occur
together; Speleonectes lucayensis Yager was the first found
and described.
Copepods are the richest group of Entomostraca,
which include approximately 25% of crustacean species
either in the whole or in subterranean habitats, but only
14% of those in anchihaline waters. In subterranean
freshwaters, Harpacticoida are several times richer than
Cyclopoida, while Calanoida are negligible (approxi¬
mately 440:150:10 as known in 1986); in anchihaline
waters, the harpacticoids are not so important, while
calanoids are comparatively well represented. Each of
these groups has some ecologically endemic families, or
nearly so, such as Superornatiremidae in Harpacticoida,
Epacteriscidae in Calanoida, and Speleoithonidae in
Cyclopoida. According to existing data, a family may
be also geographically extremely limited, such as the
calanoid Boholinidae in the Philippines. Of particular
interest is the small and relatively plesiomorphic group
Misophrioida, an essentially (but not exclusively) deep-
sea group discovered recently in Atlantic and Indo-
Pacific tropical anchihaline caves.
ENCYCLOPEDIA OF CAVES
ANCHIHALINE (ANCHIALINE) CAVES AND FAUNA
21
Ostracods are easily recognizable as tiny crustaceans
with an extremely shortened body hidden between
two valves. The order Halocyprida is well represented
with species of very diversely shaped and sculptured
valves. Among them, the genus Danielopolina is inter¬
esting for its presence in both, the anchihaline waters
and in the deep sea.
Leptostracans are small malacostracans with a
bivalved carapace on the anterior part of the body and
stalked eyes. Speonebalia cannoni Bowman, Yager, and
Iliffe from a cenote-like cave in the Caicos Islands is
eyeless.
Decapods are the best-known crustaceans repre¬
sented in anchihaline caves, primarily by different
kinds of shrimps. Members of the small family
Procarididae have only been found in such habitats in
the tropical Atlantic and in Hawaii; Agostocarididae seem
to be restricted to the Caribbean region. Atyidae are a
huge family of freshwater shrimps, represented at the sea
nearly exclusively by some genera in anchihaline waters.
The most remarkable is the genus Typlilatya, which is
found with some closely related genera exhibiting a range
of polyhaline to limnic species distributed circumtropi-
cally; they also exhibit different degrees of troglomorph-
ism. The atyid Antecaridina lauensis (Edmondson) alone is
very widely distributed—Madagascar, Red Sea, Hawaii—
which is testament to the dispersion possibilities of some
shrimps outside the anchihaline habitat. This has been
also demonstrated for Palaemonidae, represented by the
genus Macrobrachium, for example, which repeats some
distributional characteristics of the family Atyidae. The
palaemonid Typhlocaris letlwea Parisi from Libya is proba¬
bly the largest cave invertebrate at all. A series of anchiha¬
line genera belong to the essentially marine family
Alpheidae. The alpheid Barbonria cubensis (Von Martens)
in the Caribbean region has been known since the begin¬
ning of the twentieth century; the Indo-Pacific Parhippolyte
uveae Borradaile, again one of the famous red shrimps,
has been known from the end of the nineteenth century.
Some Grapsidae crabs, such as Orcovita spp., have been
found recently, although they are mostly not very
troglomorphic. The only anomuran, the historical
Munidopsis polymorpha Koelbel (family Galatheidae), from
the inundated lava tube in the Canary Islands, belongs to
an essentially deep-sea genus.
Thermosbaenaceans are small (up to 5 mm long)
and delicate crustaceans with a short carapax
under which the female broods her eggs; all are stygo-
biotic. Although the first species was found in hot
(approximately 45°C) freshwater springs, and some
species inhabit fresh continental waters, most are
nevertheless anchihaline. The group is distributed in
the Tethyan fashion, which will be described later.
As an example, Halosbnena acanthura Stock is widely
spread in the Caribbean, inhabiting interstitial and cave
waters of brackish to hyperhaline salinities. Related
species are known from the Canaries and northwestern
Australia in almost freshwater. Monodella halophila
S. Karaman, found in caves along the northeastern
Adriatic coast, often most densely populates the least
aerated water layers of a polyhaline salinity.
Mysids are small and delicate shrimp-like crusta¬
ceans with a ventral brood chamber. It is essentially a
marine group that inhabits anchihaline caves circum-
tropically, with some exceptional species in southern
Italy. Such is also the distribution of the genus
Spelneomysis, which resembles the atyid shrimp Typhlatya
in its geographical and ecological distribution and
the span of morphological adaptations. The widely dis¬
tributed and nontroglomorphic species S. cardisomae
Bowman seems to penetrate the entire variably brackish
part of the groundwater body in the Caribbean island of
San Andres, occurring even in the putrefied and proba¬
bly deoxygenated waters in the land crab burrows.
Mictaceans slightly resemble the thermosbaenaceans.
This is a small group of some deep-sea and two marine
cavernicolous species in the Bermudas ( Mictocaris halope
Bowman et Iliffe) and the Bahamas.
Isopods are one of the major groups of Malacostraca,
characterized mainly by their foliaceous abdominal
appendages (pleopods); they are very diverse in shape
and size. Since early in the twentieth century, some
Typhlocirolana spp. from the Mediterranean basin have
been known; some are anchihaline while some are from
freshwaters. Similarly old is the history of Annina lacus-
tris Budde—Lund from east Africa, which is a very eur-
yoecious, nontroglomorphic animal that also occurs in
mixohaline anchihaline caves. All of these animals
belong to the family Cirolanidae, which is essentially
marine, and the great majority of freshwater species
is stygobiotic. A number of genera and species have
been found in anchihaline waters, particularly in the
Caribbean (such as Bahalana mayana Bowman in Mexico
and the Yucatan), but also in the Mediterranean and in
Indo-Pacific areas and elsewhere, such as Haptolana pho-
leta Bruce and Humphreys in western Australia. Similar
is the frame habitat selection of the genus Cyathum
(Anthuridae), although these animals are strongly sedi¬
ment bound and burrow, while cirolanids are free walk¬
ers and even occasional swimmers. The stygobiont
anthurids are only present in the tropics.
The presence of tiny species of the families
Microcerberidae and Microparasellidae in anchihaline
waters such as artificial wells is mostly due to sand
deposits, as they are essentially interstitial animals.
Genuine cave animals are evidently some members of
the related (i.e., asellote) families of Janiridae ( Trogloia-
niropsis Uoberai Jaume) in the Balearic Islands; Stenetrii-
dae ( Neostenetroides schotteae Ortiz, Lalana, et Perez) in
Cuba; and Gnathostenetroididae ( Stenobermuda mergens
ENCYCLOPEDIA OF CAVES
22
A
Botosaneanu et Iliffe) in the Bahamas, which were
found in Caribbean to Bermudian anchihalines, as well
as in Mediterranean caves. The family Atlantasellidae
seems to be endemic to anchihaline caves of the tropi¬
cal Atlantic (Bermuda and the Caribbean).
Amphipoda seem to be by far the richest anchiha¬
line group (more than twice the species number of any
other crustacean group or three times all noncrusta¬
cean species), although the effect of numerous
researchers cannot be excluded. The family level tax¬
onomy of this group (counted to 6000 or even 8600
world species) is not yet settled, but it has to be used
for the sake of surveying. The most characteristic
group is comprised of Hadziidae and Melitidae with
particularly high diversities of genera in the Caribbean
region (e.g., Metaniphargus, Bahadzia) and Mediterranean
(e.g., Hadzia) and with a number of scattered taxa else¬
where (e.g., Liagocemdocus spp.); Western Australia
seems to represent the third such diversity center (with
the genus Nedsia, for example). The group as a whole,
as well as some genera, are represented by species of
different salinity preferences, from euhaline to limnic.
On a smaller scale and limited to the Mediterranean
and Atlantic, is the genus Pseudoniplmrgus- the majority
of its species are freshwater but never very far from the
sea. Taxonomically remote is the family Ingolfiellidae;
although including some marine species, it also includes
ancient freshwater species in the center of Africa, for
example. This family also has some anchihaline species.
Some taxonomically and geographically isolated
species seem to be more or less direct descendants of
marine relatives, such as Antronicippe serrata Stock and
Iliffe of Pardaliscidae, from anchihaline caves in the
Galapagos Islands.
At least most species or groups mentioned above are
primarily marine, conquering continental waters with
more or less success. Different are some amphipods
of the Dinaric (i.e., Adriatic) anchihaline caves. Some
species of Niphargus are present there; N. hebereri
Schellenberg is moderately widely distributed, although
not obligate, while the narrowly endemic N. pectencor-
onatae Sket seems to be an obligate anchihaline species
(Fig. 1). Niphargids are a large group of essentially
freshwater animals and only secondarily did they adapt
to the brackish water; N. hebereri is very euryhaline and
also in other respects a very tolerant species.
BIOGEOGRAPHY
Different patterns of geographical distribution can be
seen in anchihaline cave faunas. One of the most remark¬
able patterns is the so-called Tethyan distribution
(Humphreys, 2001), if historical grounds are emphasized,
or circumtropical distribution, if we suppose ecological
connections. The Tethys Sea encircled the globe in the
Mesozoic period, when the supercontinent Pangaea split
into a northern Laurasia and southern Gondwanaland. It
was only in the Tertiary when southern and northern
continents came again so close to each other that they
disconnected this circumtropical seaway; note that even
the Mediterranean was completely dry for some period
in the Miocene. A number of animal groups, even genera
and some species, have distribution ranges that reflect
the possible circumglobal distribution within the Tethys.
We may presume that particularly in the tropical belt a
number of Tethyan inhabitants have survived and many
of them succeeded in colonizing marginal waters, avoid¬
ing in this way competition with modern sea fauna.
These include the Remipedia, Thermosbaenacea, hadziid
amphipods, the Typhlatya-related group of shrimp gen¬
era, and even the mysid genus Spelaeomysis, the thermo-
sbaenacean genus Halosbaem, and some others.
However, it is difficult to believe that some circumtropi-
cally spread species, such as the shrimp Antecaridina
Iciuensis, could conserve its morphological unity without
a recent or nearly recent gene flow among populations;
its genetic unity, however, has not been proven. An open
question is also what occurred during the Messinian
salinity crisis. Did the Tethyan elements temporarily
change for freshwaters? Maybe this is connected with
the existence of some freshwater populations of some
elements such as Hadzia fragilis S. Karaman in the
Adriatic region (Fig. 2).
A certain number of those Tethyan elements character¬
istically skip the Mediterranean in their west—east distri¬
bution, such as Remipedia and the genus Halosbaena;
therefore, it is also possible that some Tethyan elements
disappeared from the dried-out Messinian Mediterranean,
which was later colonized again by only some of them.
The third possibility is their disappearance during the
colder climates in place after the Tertiary.
Other distribution patterns are more locally restric¬
ted. An instructive one is the paralittoral distribution
FIGURE 1 The amphipod Niphyargus pectencoronatae. Photo h/
Boris Sket.
ENCYCLOPEDIA OF CAVES
ANCHIHALINE (ANCHIALINE) CAVES AND FAUNA
23
FIGURE 2 The amphipod Hadzia fragilis. Photo by Boris Sket.
area along the northeastern (Dinaric) coast of the
Adriatic. Typical anchihaline taxa (such as Hadzia fragi¬
lis and Monodella halophila) are present along the coast,
including the islands, in brackish waters. They are,
however, absent in such waters within the gulf of
Kvarner (Quarnero) which was not yet marine prior
to the Pleistocene. These species are replaced in the
Kvarner anchihaline caves by generally continental
species, such as Niphargus arbiter G. Karaman, while
Hadzia is represented there by freshwater populations.
This is clear evidence of a historical—not a purely eco¬
logical—basis for the species distribution in anchiha¬
line habitats.
BIOLOGY AND ECOLOGICAL
DISTRIBUTION OF INHABITANTS
Unfortunately, for most anchihaline species the
ecological inclinations are virtually unknown. Some
groups are evidently represented only by euhaline or
at least polyhaline species which are all unquestion¬
able immigrants from the deep sea or close relatives
of deep-sea animals: the glass sponges and carnivo¬
rous sponges in the Mediterranean sea caves, the
anomuran Munidopsis polymorpha in the Canary
Islands, and the mictacean Mictocaris halope in the
Bermudas. There are also unique representatives of
some marine groups such as chaetognaths or leptos-
tracans (Paraspadella anops and Speonebalia cannoni );
however, the completely anchihaline group Remipe-
dia also seems to be polyhaline.
Most anchihaline animals are to some degree eury-
haline, with different optima more or less expressed.
The distribution of shrimp species in a series of pools
in Hawaii seems to depend mostly on different salin¬
ities. They are able to adapt their position in the water
body to other local conditions, particularly the pres¬
ence of competitors or predators. The physically deli¬
cate thermosbaenacean Monodella halophila, along the
Adriatic coast, is limited to comparatively high salin¬
ities, although its very close relative M. argentarii Stella
inhabits freshwaters in the Italian Monte Argentario.
The Adriatic caves are inhabited by some (most of
them by two) amphipod species which may be preda¬
tory on Monodella. This might also explain the (sup¬
posed) absence of Monodella in some Caribbean islands
where caves were probably not investigated meters
deep and where hadziid amphipods were regularly
found. Some essentially limnic species, such as the
copepod Diacyclops antrincola Kiefer or the shrimp
Troglocaris sp. may be present at higher salinities
than the supposed marine derivatives Hadzia fragilis and
Salentinella angelieri Ruffo et Delamare—Deboutteville,
while the primarily limnic Niphargus hebereri Schellenberg
reaches 30 parts per thousand salinity, which is close to
marine values.
It was observed in some Adriatic anchihaline waters
that at least some highly troglomorphic animals do not
avoid sunlit parts of the water body if they are not
forced to do so by competing surface animals. The pop¬
ulation of Niphargus hebereri was the most dense at the
water surface in a shaft-like cave. Such a behavior (as
well as similar situations in continental cave waters)
teaches us that light and dark relations are not of direct
importance for all cave species. This also excuses us
when we treat the open anchihaline pools jointly with
anchihaline caves. The so-called red shrimps are very
characteristic inhabitants of those pools, and they show
quite different relations toward sunshine. The atyid
Halocaridina rubra Holthuis also appears in open anchi¬
haline pools during sunshine, while Antecaridina lauen-
sis is more photophobic but is able to change its color
from red to white translucent and back in response to
light and disturbance. The hippolytid Parhippolyte uveae
Borradaile hides in cracks after sunset and during clou¬
dy days, while it reappears in ponds in masses when
the sun strikes the water; it is also able to rapidly
change its coloration. One has to add, however, that
these observations were made in single localities, and
we do not know if such was the behavior of the species
in any conditions.
Again, very characteristic is the sometimes appar¬
ently positive (in fact, probably neutral) relation of
anchihaline animals toward low oxygen concentrations
in water. This discussion has unfortunately some very
serious limitations. First, due to some analysis prob¬
lems at very low oxygen concentrations, we do not
know if cave animals have been found in totally anoxic
or just deeply suboxic water; however, the presence of
H 2 S in a thick layer speaks in favor of a total lack of
oxygen. Second, particularly high densities of some
animal colonies within the anoxic (or suboxic) water
layers are most probably a result of competition (or
predation) avoidance; however, this speculative
ENCYCLOPEDIA OF CAVES
24
A
conclusion has never been subjected to an experimen¬
tal verification. Anyway, at extremely low oxygen con¬
centrations and in the presence of H 2 S (not
quantitatively evaluated) euhaline Remipedia were
found commonly or regularly; the mesohaline
Monodella halophila, often; and the originally limnic
Niphargus hebereri, quite so. Resistance to extremely
low oxygen concentrations has been experimentally
studied and verified in some freshwater cave crusta¬
ceans; it may be explained by their characteristically
low metabolism.
Thus, based on the evidence available, anchihaline
animals are in general very euryecious (generalist)
toward a number of ecological parameters. One of the
most durable among stygobionts is Niphargus hebereri,
which has been found in limnic conditions (kilometers
far inland) and in mixohaline waters of up to 30 parts
per thousand salinity; in very food-poor habitats and in
deposits of decaying organic debris; in well-oxygenated
waters as well as in apparently anoxic layers; in waters
smelling strongly of H 2 S; and in a sink polluted with
mineral oil.
TROPHIC RELATIONS
Supposedly, the richest source of food is input
from the Earth's surface, while some food may be
produced in the cave itself, as mentioned above. It is
safe to say that the currents reaching into a euhaline
marine cave are bringing food particles as well as
pelagic larvae. Therefore, such caves are usually pop¬
ulated with a depauperate marine fauna, and the sty¬
gobionts are to some degree excluded by competition.
As the most modest users of organic matter and as
ecological generalists (as shown above), stygobionts
are able to populate the rest of the coastal under¬
ground. Some interesting studies of the organic
matter transport within genuinely marine and anchi¬
haline caves have been made, but generalizations are
still not possible. The sessile suspension (filter) fee¬
ders are nearly absent; some sponges and tube worms
are on the border between the marine and anchihaline
realm. Very few observations have been done about
the behavior of those crustaceans that possess puta¬
tive filtering structures, such as atyid shrimps and
thermosbaenaceans. They seem to be predominantly
on the bottom and collect organic detritus particles or
bacteria from the sediment. Probably more pelagic is
the rich assortment of anchihaline calanoid copepods.
On the other hand, some inhabitants are proven pre¬
dators, such as the procaridid shrimp Procaris ascen-
sionis Chace, that feed on other shrimps. The
maxillipeds of Remipedia are explicitly predatory,
while on the fang tip of their maxilla I, a poison gland
discharges. Most probably, the majority of anchiha¬
line amphipods are generalistic, feeding on detritus
and carcasses as well as on live animal prey.
THEORETICAL IMPORTANCE OF
ANCHIHALINE HABITATS
The position of anchihaline habitats on the doorstep
between the epigean and hypogean realms, as well as
between marine and freshwater environments, makes
them very intriguing for theoretical exploitation. Several
theories about the colonization of the underground
include anchihaline habitats, either implicitly or explic¬
itly. However, the first question that has to be answered
regards the origin of the anchihaline fauna itself. The
marine provenance of the majority is out of the question
if we look at the taxonomic composition. Only one local
fauna is known for which a higher percentage of mem¬
bers are freshwater by origin (anchihaline fauna along
the Adriatic coast). The fact that the band inhabited by it
was out of the sea during the last glaciation is not pecu¬
liar to that area. Probably, it was the particularly rich
continental cave fauna that succeeded to rule anchihaline
waters and thus prevented intrusion of other biota.
A much-discussed question has been the relation of
anchihaline fauna to the deep sea. Their phylogenetic
proximities have been discussed previously. At one
time it was thought that anchihaline species were
directly derived from deep-sea species. Later, data
about deoxygenation of deeper sea layers during the
geological past cast doubts on this possibility, and some
researchers came to regard the deep-sea species and
anchihaline species as two parallel lineages originating
from littoral ancestors. However, cladistic analyses of
some ostracod and some copepod groups have revealed
that at least some anchihaline species in some regions
may originate from deep-sea faunas (Danielopol, 1990).
The anchihaline habitat may also be regarded as a door
into continental waters. The regression model and the
two-step model suggest that a marine benthic animal colo¬
nizes the marine cave, adapts to the cave environment
and to the less saline water, and becomes isolated from
the sea after the sea regresses, which results in a limnic
stygobiont. These and other models differ in some details
which become theoretically negligible (making all models
complementary) if we consider the ecologically and genet¬
ically sound supposition of an active immigration. Active
immigration into a new habitat would be furthered by a
diversified gene pool and wide ecological potential of the
species and would be forced to occur only by its own nor¬
mally excessive reproduction and population expansion.
Sea regression (stranding) in this context is not an act of
repression; it implies only interruption of the gene flow
ENCYCLOPEDIA OF CAVES
ANCIENT CAVERS IN EASTERN NORTH AMERICA
25
(with the marine or surface part of the population) and
therefore faster specialization. Of course, this is equally
important for understanding the origin of the surface
freshwater fauna, which may have occurred the same
way on the surface, and the origin of continental cave
fauna, which occurred by immigration from surface fresh¬
water underground.
ANCHIHALINE FAUNA AND HUMANS
It has always been a challenge to humans to enter
the dark underworld; sometimes such a venture reveals
the beauty of speleothems or offers the promise of a
cool atmosphere in the heat of summer. So, many karst
and lava caves have been exploited for tourism, and
those that are close to coastal tourist resorts are particu¬
larly subject to such a possibility. Some of them contain
anchihaline pools.
Sipun cave in Cavtat (near Dubrovnik, Croatia) was
adapted for such visits decades ago, with the anchihaline
pool at its end being shown as a special gem. Its compara¬
tively very rich fauna of tiny crustaceans is not particu¬
larly interesting for most nonprofessionals but has
nonetheless been greatly endangered by pollution pro¬
duced by the visitors. Anchihaline caves play an impor¬
tant role in tourism and the economy of Bermuda;
nevertheless, a number of them have been damaged or
destroyed—along with their diverse fauna—by pollution
or by quarrying activities. On the other hand, the lava
cave Jameos del Agua in Lanzarote (Islas Canarias, Spain)
has also been exploited for tourism and its jameito or can-
grejo ciego (Munidopsis polymorpha) is an official symbol of
the island that is used in various promotional campaigns.
The gorgeous hypogean estuary St. Paul's Underground
River (Palawan, Philippines) is the central attraction of St.
Paul's Subterranean National Park; unfortunately, there
are no data on anchihaline fauna in it. Very imposing is
the 7-cm-long anchihaline shrimp Typhlocaris lethaea from
the cave Giok-Kebir, close to Benghazi (Libya); this might
be the world's largest troglobiotic invertebrate and could
be an interesting subject for interested tourists. Although
smaller, the beautifully colored anchihaline red shrimps
can be remarkable enough to be given local vernacular
names such as opaeula (Halocaridina rubra) in Hawaii.
They may be a subject of local taboos, such as pulang
pasayan (Parhippolyte uveae) in the Philippines, or even of
a kind of worship, such as ura buta (the same species) on
the Fiji Island of Vatulele. No doubt, red shrimps might
also become an attraction for ecotourists.
Apart from species protection to serve tourist needs
or religious habits, modern humans have become aware
of the biological importance of such habitats. A system
of anchihaline pools on the Hawaiian island of Maui
has been protected as a natural reserve particularly for
the rich anchihaline fauna. A flooded crack in the Sinai
Peninsula has (or used to have) a similar status.
See Also the Following Article
Coastal Caves
Bibliography
Danielopol, D. L. (1990). The origin of the anchialine fauna: The
"deep sea" versus the "shallow water" hypothesis tested against
the empirical evidence of the Thaumatocyprididae (Ostracoda).
Bijdragen tot de Dierkunde, 60(3—4), 137—143.
Holthuis, L. B. (1973). Caridean shrimps found in land-locked salt¬
water pools at four Indo-west Pacific localities (Sinai Peninsula,
Funafuti Atoll, Maui and Hawaii Islands), with the description of
the new genus and four new species. Zoologische Verhandelingen,
128, 1-48.
Humphreys, W. F. (2001). Relict faunas and their derivation.
In H. Wilkens, D. C. Culver & W. F. Humphreys (Eds.),
Subterranean ecosystems (pp. 417—428). Amsterdam: Elsevier.
Iliffe, T. M. (2001). Anchialine cave ecology. In H. Wilkens, D. C.
Culver & W. F. Humphreys (Eds.), Subterranean ecosystems
(pp. 59— 73). Amsterdam: Elsevier.
Jaume, D., & Boxshall, G. A. (1996). The persistence of an ancient
marine fauna in Mediterranean waters: New evidence from miso-
phrioid copepods living in anchialine caves. Journal of Natural
History, 30, 1583-1595.
Pohlman, J. W., Cifuentes, L. A., & Iliffe, T. M. (2001). Food web
dynamics and biogeochemistry of anchialine caves: A stable isotope
approach. In H. Wilkens, D. C. Culver & W. F. Humphreys (Eds.),
Subterranean ecosystems (pp. 345—356). Amsterdam: Elsevier.
Por, F. D. (1985). Anchialine pools: Comparative hydrobiology.
In G. M. Friedman & W. E. Krumbein (Eds.), Hypersaline ecosystems:
The gavish sabkha. Ecological studies: Analysis and synthesis (Vol. 53,
pp. 136—144). New York: Springer.
Porter, M. (1993). The formation of biota of inland and oceanic blue
holes. Pholeos, 13(2), 1—8.
Riedl, R. (1966). Biologie der meereshohlen. Hamburg/Berlin: Parey.
Sket, B. (1986). Ecology of the mixohaline hypogean fauna along the
Yugoslav coast. Stygologia, 2(4), 317-338.
Sket, B. (1996). The ecology of the anchihaline caves. Trends in
Ecology and Evolution, 11(5), 221—225.
Stock, J. H. (1994). Biogeographic synthesis of the insular groundwater
faunas of the (sub)tropical Atlantic. Hydrobiologia, 287, 105—117.
Vacelet, J., Boury-Esnault, N., & Harmelin, J. G. (1994). Hexactinellid
cave, a unique deep-sea habitat in the scuba zone. Deep-Sea
Research, 41(7), 965-973.
Yager, J. (1981). Remipedia, a new class of Crustacea from a marine
cave in the Bahamas. Journal of Crustacean Biology, 1(3), 328—333.
ANCIENT CAVERS IN EASTERN
NORTH AMERICA
Patty Jo Watson
Washington University
INTRODUCTION
The longest cave in the world (more than 625 km, and
still being explored and mapped) is the Mammoth
Encyclopedia of Caves.
© 2012 Elsevier Inc. All rights reserved.
26
A
Cave System in west central Kentucky in the United
States. Beginning about 4000 years ago, the people liv¬
ing in the Mammoth Cave region began exploring
some portions of this immense subterranean labyrinth.
Indigenous peoples continued to use parts of the
Mammoth Cave System until about 2000 years ago.
The materials discarded or lost underground many
millennia ago preserve detailed information, not only
about prehistoric activities in the cave, but also about
subsistence practices and other lifeways aboveground.
CAVE ARCHAEOLOGY IN EASTERN
NORTH AMERICA
The Mammoth Cave System was not the only cave
entered by aboriginal inhabitants of the Americas. To
the contrary, any cave opening large enough for a per¬
son to squeeze into was probably explored, and well
known. The earliest trip documented far into the dark
zone of any cave in the Western Hemisphere is that of
a 45-year-old man whose ca. 8000-year-old skeletal
remains were preserved in Hourglass Cave, high in
the Southern Rocky Mountains (Mosch and Watson,
1997). In Eastern North America, the oldest detailed
archaeological record of dark zone cave exploration
dates to approximately 5000 years ago when eight or
nine people made one or two trips far back inside an
11-km-long cave in what is now northern Tennessee.
They left a scattering of charcoal from their cane
torches, and—in the damp clay floor of one passage¬
way—multiple impressions of their feet (Watson et al.,
2005).
For many millennia prior to the time of European
contact, indigenous peoples of the Americas entered
caves, explored them, and used them for a variety of
purposes: for example, as storage places, as sources of
valuable minerals or of magical powers, as shrines, as
cemeteries, or as places to contact the spirit world. The
archaeology of the Mammoth Cave System illustrates
several of those activities for the two millennia between
4000 and 2000 years ago.
ARCHAEOLOGY OF THE MAMMOTH
CAVE AREA
Like many other regions of the eastern United
States, west central Kentucky was first inhabited by
Paleo-Indian hunter-gatherers, who were present by
approximately 14,000 years ago. Archaeological evi¬
dence for this time period {ca. 14,000 to 11,500 years
before present) is sparse, but data from here and else¬
where in the midcontinent indicate to archaeologists
that Paleo-Indian communities were small, and were
dispersed fairly widely across the landscape.
In archaeological terminology, the Paleo-Indian
period in Eastern North America is followed by the
Archaic period {ca. 11,500 to 3000 years ago), which is
succeeded by the Woodland period (3000 to 1100 years
ago), and—in some places—the Mississippian period
(1100 to 500 years ago). The explorers who left their
footprints in that Tennessee cave 5000 years ago
are spoken of by archaeologists as Archaic people. The
earliest explorers of the Mammoth Cave System were
also Archaic folk, as were the first people to enter
Fisher Ridge Cave just east of Mammoth Cave
National Park, Lee Cave and Bluff Cave (both within
Mammoth Cave National Park but not known to be
connected to the Mammoth Cave System), and Adair
Glyph Cave some distance east of Mammoth Cave
(Crothers et al., 2002).
During most of the Archaic period, all societies in
the Eastern Woodlands of North America hunted,
fished, and gathered the abundant resources of the for¬
ests where they lived. By 3000 years ago, however, at
the end of the Archaic and beginning of the Early
Woodland period, some local groups had become
farmers growing several species of small-seeded
plants: sunflower, a relative of sunflower called sump-
weed or marshelder, goosefoot, maygrass (also known as
canary grass), bottle gourd, and a gourd-like form of
squash (Browman et al., 2009). These Early Woodland
folk were among the first agriculturists in Eastern
North America, and some of them were also highly
skilled cavers who confidently made their way
through many miles of passages in what is now
known to be the world's longest cave, the Mammoth
Cave System. Since the 1950s, archaeologists and spe¬
leologists have studied remains left by these early
cavers to discover where they went and what they did
underground, and also to investigate the evidence they
left of their early farming economy.
PREHISTORIC ARCHAEOLOGY IN THE
WORLD’^LONGEST CAVE
Dry cave passages are excellent repositories of
archaeological and paleontological remains because
they lack factors of weathering and decay that quite
rapidly alter or destroy all perishable materials
aboveground. Some of the earliest Euroamerican
explorers of the Mammoth Cave System noted well-
preserved traces of prior human presence there.
When the National Park Service acquired the land
above Mammoth Cave, Salts Cave, and a series of
ENCYCLOPEDIA OF CAVES
ANCIENT CAVERS IN EASTERN NORTH AMERICA
27
other caves (known since 1972 to form a single cave
system; see Brucker and Watson, 1987), archaeologist
Douglas Schwartz at the University of Kentucky
was asked to prepare a series of reports on Mammoth
Cave area archaeology. Subsequently, other investiga¬
tors affiliated with the Cave Research Foundation
and the National Park Service took up the work,
which is still continuing (Crothers, 2001; Crothers
et al., 2002; Watson, 1969, 1997). Results are summa¬
rized below.
Chronology
Several dozen radiocarbon dates have been obtained
on a wide variety of material from various parts of the
prehistoric archaeological record in the Mammoth Cave
System (Crothers et al., 2002, Table 1 and Figure 1
therein). As already noted, the dates span some 2000
years, from a little before 4000 years ago to a little after
2000 years ago. Most of them, however, cluster around
2500 years before present, plus or minus a few hundred
years. Two archaeological periods are represented:
Late Archaic and Early Woodland. Both Archaic and
Early Woodland cavers explored several miles of
passages in the upper and lower levels of Mammoth
Cave and Salts Cave, but the Early Woodland record is
much more abundant and much more extensive than is
the Archaic one.
Archaeological Evidence
The most prominent and pervasive category of pre¬
historic material in dark zone cave passages is torch
and campfire debris. Everywhere the ancient cavers
traveled, they left a scatter of cane and dried weed
stalks, as well as charred fragments of these torch
materials. Charcoal smudges left by torches on cave
walls, ceilings, and breakdown boulders are also com¬
monly visible. River cane seems to have been the pre¬
ferred torch material, but dried stems of goldenrod
and false foxglove (Fig. 1) were also used. In addition
to the torch remains, fragmentary bindings and cord¬
age made from plant fibers or the inner bark of certain
trees are strewn along the passage floors.
Less abundant items include wooden and gourd
containers, mussel shell scrapers or spoons, basketry
and vegetal fiber artifacts (including remains of bags,
cordage, and moccasin-like footwear), digging sticks,
scaling or climbing poles, prints of human feet in dust
or damp clay floored passages (Fig. 2), dried human
excrement (paleofeces), and the physical remains
("mummies," actually simply desiccated bodies) of
two of the ancient cavers themselves.
FIGURE 1 Torch of dried weed stalks ( Gerardia sp., false fox¬
glove) with inner-bark tie, in situ on the floor of Indian Avenue, Salts
Cave, Mammoth Cave National Park, Kentucky. Cave Research
Foundation photo by Pete Lindsley. Used with permission.
(B)
FIGURE 2 (A) Outline of a prehistoric caver's foot in the dust
covering a breakdown boulder in Indian Avenue, Salts Cave,
Mammoth Cave National Park, Kentucky. (B) Imprint of a prehistoric
caver's foot in mud floor of the Upper Crouchway, Unknown Cave,
Mammoth Cave National Park, Kentucky. Cave Research Foundation
photos by William T. Austin. Used with permission.
ENCYCLOPEDIA OF CAVES
28
A
In most places within the dry cave passages entered
by the Early Woodland people there is evidence that
they removed cave minerals from the walls, break¬
down boulders, ceilings, and cave sediments. These
minerals consist of various sulfate compounds, most
prominently gypsum (hydrous calcium sulfate) and
mirabilite (hydrous sodium sulfate), with epsomite
(hydrous magnesium sulfate) probably also included,
although it is much less abundant in the cave than are
gypsum and mirabilite.
Archaeological excavation in the entrance areas
of both Mammoth Cave and Salts Cave yielded a con¬
siderable quantity of charred plant material, and of
animal bone (deer, turkey, raccoon, opossum, squirrel,
rabbit, turtle, fish, rodents), as well as tools made of
bone, chipped stone, and ground stone. The Salts Cave
excavations also recovered a considerable quantity of
human bone (Robbins, 1997). Physical remains of
prehistoric people are known from a few cave sites
where they died by accident ( e.g ., an adult man in
Mammoth Cave and a young boy in Salts Cave),
or were purposely buried (Crothers et al., 2002). Use
of sinkholes and vertical shafts as mortuary facilities,
into which the dead were lowered or dropped, was
much more common than cave interments across the
Midwest, Midsouth, and Southeast.
Interpretations
The somewhat scant evidence dating to the earliest
human presence in Mammoth Cave, Salts Cave, Lee
Cave, and Fisher Ridge Cave, as well as the Tennessee
footprint cave, indicates that Archaic people who ven¬
tured into these underground locales seem to have
been exploring rather than working or spending long
periods of time there. Adair Glyph cave is quite differ¬
ent, however (DiBlasi, 1996). Here the ancient people
used sticks or their fingers on a mud-floored passage
to make a dense series of curvilinear and rectilinear
markings whose meanings are unknown to us, but
whose creation must have taken considerable time
and thought, and probably more than one trip into
the cave. At any rate, it is obvious that those respo¬
nsible for the markings were not simply exploring
this room, which is about one kilometer from the
entrance. At Fisher Ridge Cave the archaeological evi¬
dence primarily indicates exploration, but there is a
lattice or checkerboard design (scratched onto a large
breakdown boulder) that was present when the first
Euroamerican cavers entered that passage during the
early 1980s. Finally, several geometric and representa¬
tional figures (some scratched, some drawn with charcoal)
documented in both Salts Cave and Mammoth Cave
are probably prehistoric, either Late Archaic or Early
Woodland, or both (Crothers, 2001; DiBlasi, 1996).
In Wyandotte Cave, Indiana, and 3rd Unnamed
Cave, Tennessee, Late to Terminal Archaic people
quarried and worked chert derived from underground
sources. At 3rd Unnamed Cave they also engraved rec¬
tilinear, curvilinear, and representational figures onto
the limestone ceilings, walls, and breakdown boulders
(Franklin, 2008), some of which are reminiscent of the
mud images in Adair Glyph Cave. Thus, it appears
that the first people to enter dark zones of several big
caves in Eastern North America were indeed exploring
these complexes, but they were also obtaining chert
from some of them, and in several they were carrying
out ritual activities as well.
During the subsequent. Woodland period (3000
to 1200 years ago), caves were important to many com¬
munities across the Midsouth and Southeast as cere¬
monial and mortuary foci. Dozens of examples
are known in Kentucky, Tennessee, the Virginias,
Alabama, Georgia, Florida, and Texas, but the prehis¬
toric archaeological record in the Mammoth Cave
System ceases 2000 years ago. For the centuries
between about 2800 and 2300 years BP, however. Early
Woodland cavers spent considerable amounts of time
in both Salts Cave and Mammoth Cave mining sulfate
minerals. They were obviously very familiar with sev¬
eral miles of complicated passage systems in each of
these caves, and were quite capable of navigating their
way through multiple levels of walking passages,
crouch ways, and crawl ways. They battered the gyp¬
sum crust off interior cave surfaces; they dug gypsum
crystals (satin spar and selenite) from crevices in cave
walls, and/or from sediment on cave floors and
ledges; they scooped up mirabilite from those areas
where it forms so abundantly that it piles up in and on
the breakdown and on passage floors; they brushed
mirabilite crystals off cave walls into gourd, wood, or
basketry containers; and they removed mirabilite-bear-
ing cave sediments, presumably so that the salty,
medicinal sulfate deposits could be leached from them
outside the cave.
The technology used for all this work was simple
but effective, consisting of cane and weed stalk
torches; twigs and small branches for campfire fuel;
digging sticks, climbing poles (Fig. 3) and ladders; bas¬
kets (Fig. 4), fiber bags, and gourd and wood contain¬
ers. The minerals removed from the cave were
probably sought for their medicinal (mirabilite, and
perhaps epsomite as well) or magical attributes.
Battering the gypsum crust off cave walls produces
gypsum powder, which can be mixed with water or
grease to make white paint, known ethnographically to
be important in ritual contexts. Gypsum crystals may
have been thought to have special powers that would
enable curers and wisdom-keepers to diagnose and
treat illness, to foretell the future, or to exercise other
ENCYCLOPEDIA OF CAVES
ANCIENT CAVERS IN EASTERN NORTH AMERICA
29
FIGURE 3 Prehistoric scaling pole, Indian Avenue, Salts Cave,
Mammoth Cave National Park, Kentucky. Radiocarbon age of
2760 ± 40 years (uncalibrated; Beta-87915). Cave Research Foundation
photo by Frangois-Marie Callot and Yann Callot. Used with permission.
FIGURE 4 Prehistoric split-cane basket in Canter Avenue, Mammoth
Cave, Mammoth Cave National Park, Kentucky. Radiocarbon age of
2630 ± 55 years (uncalibrated; Beta-47292). Cave Research Foundation
photo by Roger W. Brucker. Used with permission.
magical abilities (all these activities are attested ethno-
graphically among indigenous North American popu¬
lations). Early Woodland mining of cave sulfates is
most extensive and best known in Salts Cave and
Mammoth Cave, but evidence of gypsum mining has
also been documented recently in Hubbards Cave,
Tennessee, and probably took place in Big Bone Cave,
Tennessee, as well (Crothers et al., 2002).
The pictographs and mud glyphs created in some
caves as early as the Late Archaic period, and present
in many other subterranean locales dating from
Woodland and Mississippian periods up to the time of
European contact (Simek, 2008), are clear evidence
that many, if not all, activities underground took
place within settings that were very special to the
ancient cavers. We cannot know exactly what they
thought about the world underground, but there is
information from ethnographic and ethnohistorical
records indicating general concepts that help us specu¬
late about prehistoric belief systems in Eastern North
America ( e.g Diaz-Granados, 2008; Hudson, 1976;
Simek and Cressler, 2008). Many North American
Indians believed in a multilayered universe, wherein
there was another world (or several, the details differ
from group to group) below the surface world, and
another (or several) above the surface world. Each
layer was inhabited by an array of creatures specific
to it. The ancestors of humankind, having originated
in the underground world, had climbed onto the sur¬
face world and established themselves there in the pri¬
mordial past. Sky beings (e.g., falcon or owl spirits)
and subterranean beings (the most powerful and fear¬
some being a kind of griffin, or dragon-like creature
known variously as the Uktena, Michibichi, the Water
Panther, and the Long Tailed Beast) could be contacted
and engaged in human enterprises by specially gifted
individuals (e.g., shamans) who knew how to carry out
the dangerous procedures required: in particular, how
to journey to those realms, negotiate with their inhabi¬
tants, obtain magical substances and objects, and
return safely to the surface world. It seems likely that
the enigmatic cave pictographs, petroglyphs, and mud
glyphs were created by specialists in the esoteric
knowledge necessary to ensure safe journeys to and
from portions of the world underground.
CONCLUSION
The first human inhabitants of America were also
the first North American cavers. They found and
explored hundreds of caves, especially in the midconti¬
nental karst region that extends from the Midwest
through the Midsouth to the southeastern United
States, including what is now known to be the longest
cave in the world. Archaeologists and speleologists
have investigated only a small sample of indigenously
utilized caves, but the resulting information makes it
clear that subterranean landscapes were well known to
aboriginal communities who lived in these karstic
areas. As was the case with other features of their sur¬
roundings, prehistoric people were familiar with the
physical and spiritual aspects of the world under¬
ground, knew how to get there, and once there how to
travel to their objectives. They also knew how to nego¬
tiate with the supernatural beings that inhabited these
ENCYCLOPEDIA OF CAVES
30
A
realms of darkness, and what to do to ensure safe
return to the surface world. Because of the size and
complexity of the caves they frequented, it can
quite justifiably be claimed that they were the best
cavers in the world, a distinction held by Archaic and
Woodland peoples of Eastern North America until less
than a century ago.
See Also the Following Articles
Cave Dwellers in the Middle East
Maya Caves
Bibliography
Browman, D. L., Fritz, G. J., Watson, P. J., & Meltzer, D. J. (2009).
Origins of food producing economies in the Americas.
In C. Scarre (Ed.), The human past (pp. 306—349). London:
Thames and Hudson.
Brucker, R. W., & Watson, R. A. (1987). The longest cave (2nd ed.).
Carbondale: Southern Illinois University Press.
Crothers, G. M. (2001). Early Woodland mineral mining and perish¬
able remains in Mammoth Cave, Kentucky. In P. Drooker (Ed.),
Fleeting identities: Perishable material culture in archaeological research
(pp. 314—334). Carbondale: Southern Illinois University Press.
Crothers, G. M., Faulkner, C. H., Simek, J. F., Watson, P. J., &
Willey, P. (2002). Woodland cave archaeology. In D. Anderson &
R. Mainfort (Eds.), The woodland southeast (pp. 502—524).
Tuscaloosa: The University of Alabama Press.
Diaz-Granados, C. (2008). Picture cave. In D. H. Dye (Ed.), Cave
archaeology in the Eastern Woodlands (pp. 203—215). Knoxville:
University of Tennessee Press.
DiBlasi, P. J. (1996). Prehistoric expressions from the central Kentucky
karst. In K. C. Carstens & P. J. Watson (Eds.), Of caves and shell
mounds (pp. 40—47). Tuscaloosa: University of Alabama Press.
Franklin, J. D. (2008). Big cave archaeology in the East Fork Obey
River Gorge. In D. H. Dye (Ed.), Cave archaeology of the Eastern
Woodlands (pp. 141—155). Knoxville: University of Tennessee Press.
Hudson, C. (1976). The southeastern indians. Knoxville: University of
Tennessee Press.
Mosch, C. J., & Watson, P. J. (1997). An ancient Rocky Mountain
caver. Journal of Cave and Karst Studies, 59, 10—14.
Robbins, L. M. (1997). Prehistoric people of the Mammoth Cave area.
In P. J. Watson (Ed.), Archeology of the Mammoth Cave area
(pp. 137—162). Reprint of the 1974 Academic Press publication,
with a new preface by the editor. St. Louis, MO: Cave Books.
Simek, J. F. (2008). Afterword: Onward into the darkness. In D. H.
Dye (Ed.), Cave archaeology of the Eastern Woodlands (pp. 261—270).
Knoxville: University of Tennessee Press.
Simek, J. F., & Cressler, A. (2008). On the backs of serpents.
In D. H. Dye (Ed.), Cave archaeology of the Eastern Woodlands
(pp. 169—191). Knoxville: University of Tennessee Press.
Watson, P. J. (Ed.), (1969). The prehistory of Salts Cave, Kentucky
Springfield: Illinois State Museum.
Reprint of the 1974 Academic Press publication, with a new preface
by the editor. Watson, P. J. (Ed.), (1997). Archeology of the
Mammoth Cave area. St. Louis, MO: Cave Books.
Watson, P. J., Kennedy, M. C., Willey, P., Robbins, L. M., & Wilson,
R. C. (2005). Prehistoric footprints in Jaguar Cave, Tennessee.
Journal of Field Archaeology, 30, 25—43.
A SELLUS AQUATICUS: A MODEL
SYSTEM FOR HISTORICAL
BIOGEOGRAPHY
Rudi Verovnik
University of Ljubljana
INTRODUCTION
The isopod Asellus aquaticus is a common and wide¬
spread freshwater crustacean in most parts of Europe. It
prefers slow-flowing streams and lakes and is often
found in ponds and ditches shaded over by trees—the
decaying leaves being its favorite food, but it is extremely
eurytopic and eurythermic, surviving even in brackish
waters. Young specimens are often found in interstitial
hyporheic habitats, possibly due to avoidance of preda¬
tion. Being so eurytopic the species can be considered as
preadapted to the subterranean life. Thus, many hypo-
gean populations are found scattered throughout Europe
(Fig. 1). In the northern part of its range these habitats
are commonly of anthropogenic origin including mines,
water supply pipes, and waterworks, indicating very
recent hypogean invasion. Indeed the specimens from
such habitats as well as from caves in northern Europe
are only slightly troglomorphic with their slight body and
sometimes eye depigmentation. Highly troglomorphic
populations are confined only to two relatively small
areas in southeastern Europe.
The first such area is the vicinity of Mangalia
in the Dobrogea region in eastern Romania with
the famous Movile Cave. This region is known for
its highly sulfur ous mesothermal waters (21°C)
and well-established subterranean food chain based
on chemoautotrophic bacteria. From here the
troglomorphic subspecies A. aquaticus infernus was
described. It is commonly found in the hand-
excavated wells and sulfurous springs in the karstic
aquifer around Mangalia, but much sparser in the
Movile Cave itself. It is characterized by several
troglomorphic traits including K-strategy traits.
The second region is in the northwestern part of the
Dinaric karst, a well-known "hot spot" of subterranean
biodiversity. Here several caves are inhabited by partially
or highly troglomorphic populations, which were
described as separate subspecies. In addition, several
karst poljes with isolated surface water drainages are also
inhabited by the morphologically distinct subspecies,
A. aquaticus carniolicus. The common trait linking these
surface populations with subterranean subspecies is the
presence of a large respiratory area on posterior pleo-
pods—the unsclerotized part of the expodite (Fig. 2). The
respiratory area is much smaller in all other populations
Encyclopedia of Caves.
© 2012 Elsevier Inc. All rights reserved.
ASELLUS AQUATICUS: A MODEL SYSTEM FOR HISTORICAL BIOGEOGRAPHY
31
throughout Europe, including the A. a. infernus. Due to
this common feature, it has been presumed that the sur¬
face populations in karst poljes are the ancestors of the
subterranean populations. As in most troglobionts the
epigean ancestors are either extinct or unknown, and this
gives us a unique opportunity to study cave invasions
through historical biogeography.
DEVELOPMENT OF TROGLOMORPHIE S
Even those subterranean populations of A. aquations
that are not considered extremely troglomorphic
can be easily distinguished from the surface popula¬
tions by their lack of body pigmentation and reduction
of eyes and some cuticular structures (Fig. 3). The pro¬
gressive troglomorphies present in A. a. cavernicolus
include enlargement of the body and elongation of
appendages, in particular of the antennae II which can
be linked with enhancement of the extraoptical sensory
structures, K-strategy consequences in this subspecies
seem to be limited to different paedomorphoses
including more isometric articulation and more sym¬
metric shape of some appendages. In subspecies A. a.
infernus, the smaller number of eggs produced by
females compared to the surface-dwelling subspecies
is the only observed adaptation in reproductive biol¬
ogy, indicating that A. aquaticus is a relatively young
immigrant to the subterranean waters.
Even from this brief description of the troglo¬
morphic traits, it is evident that they differ in different
subterranean populations. It is expected that subterra¬
nean populations in Mangalia and those from the
Dinaric karst do not share a common ancestor, being
separated by more than 1000 km distance. It is.
FIGURE 1 Approximate distribution of Asellus aquaticus in
Europe. Black dots denote nonhomogeneous troglomorphic popula¬
tions; the squares, areas with specialized troglomorphic populations.
however, much more surprising to see that differences
occur also among subterranean populations in a small
area of the northwestern part of the Dinaric karst and
even within the same cave system. In the Planina
Cave, populations from the Rak River and the Pivka
River channels differ in troglomorphic traits. The
Rak channel population is considered to be more
troglomorphic, characterized by multiplication and
shortening of setae. This implies that even single caves
were invaded on several occasions, strongly supporting
the hypothesis of active cave invasions of A. aquaticus.
POSTOJNA-PLANINA CAVE SYSTEM
The main theater of the historical biogeographic
studies is the Postojna-Planina Cave System (PPCS), an
approximately 25-km-long system comprising several
caves with accessible river passages of hypogean flow
of the Pivka and Rak rivers. Upstream from the PPCS
the plain of the Postojna basin is inhabited by the
nominotypic subspecies A. a. aquaticus, while the short
but spectacular Rakov Skocjan Polje is inhabited by the
FIGURE 2 The shape of the nonsclerotized respiratory area of the
fifth pleopod of Asellus aquaticus. The large respiratory area form is
distributed along the Dinaric karst.
7
\ , U .,
*
/ J
- '■
FIGURE 3 Troglomorphic and surface populations of Asellus
aquaticus are easy to distinguish. Photo courtesy of Boris Sket. Used with
permission.
ENCYCLOPEDIA OF CAVES
32
A
large respiratory area subspecies A. a. carniolicus. In the
hypogean Pivka River the first cave is the famous tour¬
istic Postojna Cave where only slightly depigmented
specimens of the A. a. acjuaticus are found at the
entrance. The Pivka River then flows through caves
Crna jama and Pivka jama where a mixture of fully
pigmented, partially depigmented, and fully depig¬
mented and eyeless specimens is present. Pivka Cave
is also the type locality of A. a. cavernicolus.
Morphometric studies revealed that the type speci¬
mens collected 80 years ago are more troglomorphic
than the recent inhabitants of the Pivka Cave. This
strange phenomenon could perhaps be explained by
the eutrophication of the Pivka River due to organic
pollution, which was particularly severe in the second
half of the twentieth century. This enabled several epi-
gean species, including A. a. aquaticus, to penetrate
deeper into the cave system. Downstream from the
Pivka Cave, there are several kilometers of inaccessible
river passages after which the river reappears in the
Pivka channel of Planina Cave. Highly troglomorphic
A. a. cavernicolus is abundant here; however, single
pigmented specimens can also be found after floods.
In the Rak channel only A. a. cavernicolus is present.
Downstream of the PPCS the epigean river Unec in the
large Planina Polje is stretching, inhabited by the large
respiratory area subspecies A. a. carniolicus.
In order to understand the connections between
these (sub)populations and subspecies and to put them
into the phylogeographic context, different molecular
studies were performed. With an RAPD (Random
Amplified Polymorphic DNA) study several interest¬
ing patterns emerged: first, the grouping of the popu¬
lations is congruent with subspecific status; second,
the mixed population from the Pivka Cave is geneti¬
cally homogeneous despite morphological heterogene¬
ity, and is more closely related to the upstream surface
populations than to downstream A. a. cavernicolus;
third, the genetic diversity was significantly lower in
the Pivka Cave and the Pivka channel of the Planina
Cave, possibly due to bottlenecks caused by the recent
organic pollution of the river. The gene flow estimates
confirm the discontinuity between the Pivka Cave and
the Pivka channel (Fig. 4); however the differences
between gene-flow estimates among populations in the
PPCS were not significant. Molecular analyses also
confirmed clear separation between A. a. cavernicolus
populations from the Pivka and the Rak channels
which is in line with observed morphological differ¬
ences. The results of RAPD analyses have currently
been corroborated also by a preliminary microsatellite
study where the gene-flow estimate across 10 microsa¬
tellites showed the lowest gene flow between these
two populations. All above-mentioned analyses con¬
firm separate status of all populations of A. aquaticus
in PPCS, but due to the anonymous nature of RAPD
DNA fragments, no historical biogeographic conclu¬
sions can be drawn.
To add the historical component to the molecular
differentiation, a wider scale mt-DNA sequence study
was employed. One of the most important issues in
phylogeographic studies is the calibration of the
molecular clock: the speed with which the changes in
DNA sequences evolved. To do the calibration, either
well-developed series of fossil records or good timing
geological events that have caused the split of the
populations and prevented gene flow between them
up to present are required. None of these was available
in A. aquaticus, as the only geological information that
could be potentially useful is the estimate of the begin¬
ning of the karstification and the speleogenesis in the
region of the study. This proved to be broadly esti¬
mated between 1 and 5 myr, thus providing only a
general timeframe for the evolution of troglomorphic
populations. Luckily, a molecular clock of a related
subterranean aquatic isopod Stenasellus spp. calibrated
by the separation of the Tyrrhenian Islands could be
used.
At the level of the PPCS populations, the sequence of
the first subunit of cytochrome oxidase mitochondrial
gene (COI) revealed no hydrogeographic or taxonomic
grouping, thus negating the RAPD-based conclusions.
The groups that appeared did confirm a higher level of
FIGURE 4 Analysis of gene flow between A. aquaticus popula¬
tions along two confluent sinking rivers in the PPCS. The values on
the arrows are estimated number of effective migrants per genera¬
tion. Black symbols indicate populations with pigmented individuals
belonging to A. a. aquaticus (Planina surface) and A. a. carniolicus
(Planina and Rak Polje). White symbols indicate troglomorphic
populations belonging to A. a. cavernicolus; the black-and-white sym¬
bol stands for the morphologically mixed population. The darker
area represents the subterranean river sections.
ENCYCLOPEDIA OF CAVES
ASELLUS AQUATICUS: A MODEL SYSTEM FOR HISTORICAL BIOGEOGRAPHY
33
isolation of the Rak channel population where two pop¬
ulation-specific COI haplotypes were dominant. On the
other hand, a closer relationship between the Pivka
Cave and the Pivka channel was also revealed opposing
the RAPD-based conclusions. However, these results
should be treated with caution as the presence of wide¬
spread and common haplotypes in the PPCS indicates
incomplete lineage sorting due to recent splits and possi¬
bly also due to sporadic introgression. Since the
mt-DNA is transmitted only maternally, even a single
introgression event involving a female from a different
population could be retained in the observed haplotype
diversity of the receiver population. To understand the
phylogeographic patterns, a broader geographic scale,
which in this case provides also a wider timeframe,
should be addressed.
BROADER VIEW
In order to explain the results at a broader geographic
scale, further details on the presence of cave populations
in the northwestern part of the Dinaric karst should be
discussed. Apart from the PPCS, the subspecies A. a.
cavernicolus is also distributed upstream from the polje
Rakov Skocjan and also downstream from the Planina
Polje. Additionally, a disjunct population of this troglo-
morphic subspecies was considered to be present in the
Labodnica/Trebiciano Cave in hypogena Reka River
which is a part of the Adriatic drainage; all other stud¬
ied rivers in the northwestern Dinaric karst drain
toward the Danube. Other less troglomorphic popula¬
tions of the subspecies A. a. cyclobmnchialis are present
in the Krka drainage farther east from the PPCS. The
large respiratory area A. a. carniolicus is present also at
the polje of the Cerknica Lake east of the PPCS. Two
more subspecies are described from isolated karst poljes
farther eastward, both sharing the large respiratory area
character with A. a. carniolicus.
At this broader scale, including the northeastern
Dinaric karst, mt-DNA analysis reveals structuring
that could be linked to recent and also to the presumed
historical hydrography of the region. It also provides
additional confirmation of separate cave invasions by
A. aquaticus in three geographically and currently also
hydrologically separate areas: the Krka drainage, the
Ljubljanica drainage (PPCS included), and the Reka
drainage. In the Ljubljanica drainage there is a further
split between the PPCS and the cave populations
downstream from the Planina Polje, providing evi¬
dence of an additional separate cave invasion.
With the use of the stenasellid molecular clock the
splits identified by the phylogenetic and phylogeo¬
graphic analyses can be traced down to the geographic
isolation of the north western Dinaric karst populations
between 2.7 and 4.3 mya, thus within the supposed time-
frame of the speleogenesis in the region. Soon after that,
the first fragmentation followed within the Dinaric karst
area separating the Reka River population from the rest
(Fig. 5). At that time the historical drainage of the
Ljubljanica River was dominant, receiving water from
the regions much farther eastward than at present. This
situation changed approximately 1 mya, when the cur¬
rent extent of the Ljubljanica drainage was established.
At that time the eastern karst poljes were still connected,
draining more directly toward north. The present frag¬
mented hydrology of the eastern drainage was estab¬
lished around 0.8 mya also with a shift of part of the
drainage to the Krka River (Fig. 5). These splits also pro¬
vide the approximate ages of the cave invasions, as most
of the changes of drainages can be linked to the ongoing
karstification. Thus the first invasion of the subterranean
environment was possibly completed in the Reka River
drainage, followed by the Ljubljanica and Krka River
drainages. This corresponds well with the decreasing tro-
glomorphy of the populations, the one from the Reka
River in the Labodnica/Trebiciano Cave being the most
troglomorphic with slender and elongated appendages
and highly paedomorphic shape of the pleopods.
ASELLUS KOSSWIQI
The troglomorphic population from the Labodnica/
Trebiciano Cave in the Reka River splits basally within
the northwestern Dinaric karst group in mt-DNA phy¬
togeny; therefore its previous designation as belonging
to the subspecies A. a. cavernicolus was not supported.
This is confirmed also by a large number of popula¬
tion-specific RAPD fragments found in the Reka River
population, indicating a long period of isolation. In
addition, the troglomorphic population from the Reka
River is characterized by a distinct nuclear 28S rDNA
sequence, which is otherwise highly conserved
throughout the European range of the species (Fig. 6).
The phylogeny based on 28S rDNA puts the Reka
River population firmly into the northeastern Dinaric
karst clade, making the historical scenario described in
the previous chapter more plausible.
The solution of the ambiguous taxonomic position
of the Reka River troglomorphic population came with
the discovery of the syntopic occurrence of the troglo¬
morphic and nontroglomorphic specimens of A. aquati¬
cus in the resurgence of the Reka River near
Monfalcone in Italy. Despite syntopic occurrence, no
intermediate forms were found, indicating that there is
an effective reproductive barrier in place. The subterra¬
nean population was therefore described as separate
ENCYCLOPEDIA OF CAVES
34
A
FIGURE 5 Graphical representations of a hypothetical chain of events that best explains the current distribution of haplotypes of Asellus
aquaticus in the northwestern Dinaric karst. A zigzag line indicates fragmentation events that separated drainages. The pale line denotes sub¬
terranean river passages; white-filled circles subterranean populations; and brown-filled circles isolated karst polje surface populations. The
maximum age of the events was estimated by applying the COI clock rate of stenasellids.
species Asellus kosswigi, honoring the first discoverer of
this cave population. The species status was confirmed
also by molecular studies, as no traces of gene
exchange between the cave and the surface popula¬
tions was observed, both for mitochondrial and
nuclear genes. One should note that there are several
troglobiotic species described within genus Asellus, but
all are from the Far East, mostly Japan. Asellus kosswigi
is currently the second species of the genus recognized
in Europe and illustrates the common side effect of
phylogeographic studies: description of new species.
99 -
i
NW Dinaric karst
77L Asellus kosswigi
>A. a. aquaticus
CONTINENTAL SCALE
One must understand that the processes within the
range of Asellus aquaticus in the northwestern part of the
Dinaric karst were not isolated, and therefore an even
wider continental scale phylogeographic study was per¬
formed. To get an insight into the historical biogeogra¬
phy at the continental scale, more than 70 populations
were sampled throughout the species range in Europe,
also covering the rest of the Dinaric karst. Based on the
COI molecular clock calibration for stenasellids, the
expansion of the species to central Europe was estimated
89
100
^ -Asellus sp. ? - Ukraine
-Asellus hilgendorfii
100
//
Proasellus sp.
0.1
FIGURE 6 A simplified 28S rDNA phytogeny of the European
populations of the Asellus aquaticus. The posterior probability of each
node (in percent) is shown on branches.
ENCYCLOPEDIA OF CAVES
ASELLUS AQUATICUS: A MODEL SYSTEM FOR HISTORICAL BIOGEOGRAPHY
35
at approximately 9 mya when the southern Russia region
was connected with the Pannonian Basin by the partially
brackish Parathetys Sea, which provided dispersal routes
for the ancestral A. aquaticus. The Pannonian basin was
the origin of the first expansion in Europe and its south¬
western edge still hosts the highest genetic diversity,
indicating the presence of a very old refugium in this
region (Fig. 7). Between 4 and 8 mya the species dis¬
persed throughout most of its current southern range,
forming secondary refugia in the Balkan Peninsula,
Apennine Peninsula, and in the newly forming Dinaric
karst. All these refugia are confirmed by the presence of
geographically defined groups of the COI phylogeny
and host large haplotype diversity (Table 1, Fig. 7).
In the COI phylogeny the northwestern Dinaric
karst clade is positioned as a sister group of the central
European clade and an isolated clade from the Baske
Ostarije karst polje in the central Dinaric karst. This
provides a reasonable assumption that the Dinaric
karst region was colonized from the Pannonian basin.
Due to karstification, the populations in the northwest¬
ern parts of the Dinaric karst, and the Baske Ostarije
became isolated from the rest of the species range,
which is also evident from the 28S rDNA phylogeny
(Fig. 6). As the 28S rDNA nuclear gene is highly repeti¬
tive, homogenization through concerted evolution over
most of the species range could be achieved with very
limited gene flow. Due to long-term isolation, the
northwestern Dinaric karst escaped that process,
retaining several specific 28S rDNA sequences. The
highly troglomorphic population from Mangalia in
Romania was found to be identical in its 28S rDNA
sequence to the other populations in Europe, and had
COI haplotypes similar to those from the nearby sur¬
face populations.
Through distribution of COI haplotype diversity,
several additional questions were addressed. The first
was whether the species survived Pleistocene glacia¬
tions in the presumed permafrost areas north of the
Alps. Several populations with high haplotype diver¬
sity, similar to the observed diversity in the popula¬
tions in southern refugia, were found as far north as
the Netherlands and northern Germany, indicating that
such microrefugia did exist (Table 1). The second com¬
parison was between habitat types where large rivers
were found to be most diverse, providing the core hab¬
itat of the species. It was more surprising to find that
haplotype diversity in the subterranean populations
was twice as high as in the karstic springs, the pre¬
sumed origin of some of the cave invasions (Table 1).
Several hypotheses were formulated to explain that,
including relaxed selection in caves due to reduced
competition and predation. The second is that caves
^ Labodnica - Trcbiciano Cave
£ BaSkc Ostarije Polje
0 Drava River
FIGURE 7 Geographic distribution of clades of the COI mito¬
chondrial DNA haplotypes in European Asellus aquaticus. The north¬
western Dinaric karst region is magnified for clarity. The possibly
oldest refugium and origin of dispersion situated in the western
Pannonian region (Drava River) is denoted by a black circle. Areas of
overlapping distribution of clades are indicated by darker color.
TABLE 1 ANOVA (Analysis of Variance) Design for Two
Hypotheses Explaining the Geographic Pattern of Nucleotide
Diversity (%) of Asellus aquaticus in Europe
Hypothesis
ANOVA groups
N
7T
SD
Classic refugia
Dinaric karst
17
0.00314
1.03E-05
Apennine
6
0.01259
0.000506
Pannonian
14
0.01012
0.000147
S. Balkan
12
0.00965
0.000332
N. Europe
12
0.01129
0.000299
Habitat
river
21
0.01736
0.00039
stream
17
0.00185
6.65E-06
lake
5
0.00189
3.79E-06
subterranean
7
0.00902
0.00011
spring
11
0.00487
8.86E-05
ENCYCLOPEDIA OF CAVES
36
A
served as shelters during Pleistocene climatic oscilla¬
tions, which reduced or even eliminated the surface
populations during extremely cold or dry conditions.
The most obvious explanation is that caves in general
are a much larger habitat compared to springs, there¬
fore hosting larger and more stable populations.
PROSPECTS FOR THE FUTURE
Many questions were answered and many stories
unfolded, but there is still much more to be discov¬
ered using the model species Asellus aquaticus. One of
the unresolved issues is the taxonomic status of the
isolated populations from the northwestern Dinaric
karst. Despite all evidence, the process of speciation
within this group seems to be at such an early phase
that the isolation and number of separate species
could not be unequivocally resolved by the studied
molecular markers. Microsatellites could provide
these answers as they allow much more accurate
detection of gene flow between populations. This will
provide opportunities to test A. aquaticus as a model
species for parallel speciation, possibly of a common
speciation process in many subterranean taxa. It is
also not clear whether the formation of the troglo-
morphic populations was a result of a complete isola¬
tion from the surface source population or it could
have been achieved in sympatry through ecological
speciation. These are certainly very promising pro¬
spects for the future and A. aquaticus still has many
important questions to answer.
See Also the Following Article
Invasion, Active versus Passive
Bibliography
Avise, J. C. (2000). Phylogeography: The history and formation of species.
Cambridge, MA: Harvard University Press.
Coen, E., Strachan, T., & Dover, G. (1982). Dynamics of concerted
evolution of ribosomal DNA and histone gene families in the mel-
anogaster species subgroup of Drosophila. Journal of Molecular
Biology, 158, 17—35.
Culver, D. C. (1982). Cave life. Cambridge, MA: Harvard University
Press.
Culver, D. C. & Sket, B. (2000). Hotspots of subterranean biodiversity
in caves and wells. Journal of Cave and Karst Studies, 62, 11—17.
Ketmaier, V., Argano, R., & Caccone, A. (2003). Phylogeography and
molecular rates of subterranean aquatic Stenasellid isopods with
a peri-Thyrrenian distribution. Molecular Ecology, 12, 547—555.
Lynch, M. & Milligan, B. G. (1994). Analysis of population genetic
structure with RAPD markers. Molecular Ecology, 3, 91—100.
Prevorcnik, S., Blejec, A., & Sket, B. (2004). Racial differentiation in
Asellus aquaticus (L.) (Crustacea: Isopoda: Asellidae). Archiv fur
Hydrobiologie, 160, 193—214.
Sket, B. (1994). Distribution of Asellus aquaticus (Crustacea: Isopoda:
Asellidae) and its hypogean populations at different geographic
scales, with a note on Proasellus istrianus. Hydrobiologia, 287,
39-47.
Turk-Prevorcnik, S. & Blejec, A. (1998). Asellus aquaticus infernus, new
subspecies (Isopoda: Asellota: Asellidae), from Romanian hypo¬
gean waters. Journal of Crustacean Biology, 18, 763—773.
Verovnik, R., Sket, B., Prevorcnik, S., & Trontelj, P. (2003). Random
amplified polymorphic DNA diversity among surface and subter¬
ranean populations of Asellus aquaticus (Crustacea: Isopoda).
Genetica, 119, 155—165.
Verovnik, R., Sket, B., & Trontelj, P. (2004). Phylogeography of sub¬
terranean and surface populations of water lice Asellus aquaticus
(Crustacea: Isopoda). Molecular Ecology, 13, 1519—1532.
Verovnik, R., Sket, B., & Trontelj, P. (2005). The colonization of
Europe by the freshwater crustacean Asellus aquaticus (Crustacea:
Isopoda) proceeded from ancient refugia and was directed by
habitat connectivity. Molecular Ecology, 14, 4355—4369.
ASTYANAX MEXICANUS : A MODEL
ORGANISM FOR EVOLUTION AND
ADAPTATION
William R. Jeffery
University of Maryland
EVOLUTION, ADAPTATION, AND
MODEL ORGANISMS
Biological evolution is the change in inherited traits
over successive generations in populations of organ¬
isms. Evolutionary modification of traits occurs when
variation is introduced into a population by gene
mutation or genetic recombination or is removed by
natural selection or genetic drift. Adaptation is a key
evolutionary process in which the fitness of traits and
species are adjusted by natural selection to become
better suited for survival in specific ecological habi¬
tats. The environment acts to promote evolutionary
change through changes in development, but little is
known about the underlying mechanisms. One of the
reasons is that there are only a few suitable model
organisms that can shed light on this topic. A model
organism is a species that is studied extensively with
anticipation that the results can be applied to biologi¬
cal phenomena in general. Cave animals can serve as
excellent models to study the relationships between
the environment, evolution, adaptation, and develop¬
ment because of the relative simplicity of their habi¬
tat. Troglomorphic (cave-related) traits, such as
elongated appendages, lowered metabolism, special¬
ized sensory systems, and loss of eyes and pigmenta¬
tion have evolved as a response to the effects of
perpetual darkness. Here, we describe the characid
fish Astyanax mexicanus, one of the few model
Encyclopedia of Caves.
© 2012 Elsevier Inc. All rights reserved.
ASTYANAX MEXICANVS: A MODEL ORGANISM FOR EVOLUTION AND ADAPTATION
37
organisms that is available for studying the molecu¬
lar, developmental, and genetic basis of evolution and
adaptation.
ASTYANAX MEXICAN US AS A MODEL
ORGANISM
Astyanax mexicanus is a single species with a
surface-dwelling form (surface fish) and many dif¬
ferent cave-dwelling forms (cavefish) (Fig. 1A). The
divergence between surface fish and cavefish is
relatively recent, having occurred less than three
million years ago. Cavefish have evolved construc¬
tive features, such as new methods of feeding and
physical orientation, which adapt them to life in
caves. They have lost or substantially reduced their
eyes, pigmentation, and other regressive traits
whose evolutionary significance is not completely
understood (Table 1).
The biology of Astyanax mexicanus makes it an excel¬
lent model organism. It is well suited for laboratory
research because it can be raised on a simple diet, pro¬
duces a large number of embryos, and has a relatively
short generation time, about 4—6 months. Surface fish
and cavefish embryos are large, clear, and suitable for
experimental manipulations, such as tissue transplan¬
tations. Gene inhibition and overexpression methods
have been developed in Astyanax mexicanus, and phe¬
notypic changes can be studied by powerful genetic
analysis (Jeffery, 2009a). Surface fish and cavefish are
interfertile. Moreover, different cavefish populations
can be crossed to perform genetic complementation
tests. Therefore, cavefish can be compared to surface
fish in essentially the same way that mutants are com¬
pared to wild-type phenotypes in model organisms
such as Drosophila melanogaster. Finally, some cavefish
populations have been derived independently from
surface fish ancestors, providing a rich foundation for
studying convergence or parallel evolution.
NATURAL HISTORY OL ASTYANAX
MEXICANUS
The ancestors of Astyanax mexicanus surface fish first
entered Central America after the Isthmus of Panama
was formed about 3 million years ago. Since then their
range has gradually expanded north into Mexico and
the southwestern United States. During the Pleistocene
epoch there was probably periodic northern expansion
and contraction of this range due to temperature fluc¬
tuations. Surface fish entered limestone caves through
entrapment by stream capture during this period but
permanent colonization probably occurred only in a
FIGURE 1 Astyanax mexicanus cavefish populations in northeast¬
ern Mexico. (A) Photographs of cavefish populations named on the
left and located according to letters on the map. (B) Map showing the
distribution of 29 cavefish populations in Tamaulipas and San Luis
Potosi, Mexico. Each sphere represents one cavefish population. Filled
spheres labeled with a letter represent populations shown and named
on the left. Boxes indicate the location of independently evolved
Guatemala (blue box), El Abra (red box), and Micos (green box) cave¬
fish populations. Dashed lines representing zones of cavefish popula¬
tions that may have evolved some troglomorphic features separately
subdivide the El Abra box. Inset: Outline map of Mexico showing
northeastern region in B as a rectangle and a filled sphere indicating
the location of the Granadas cavefish population in Guerrero.
TABLE 1 Constructive and Regressive Troglomorphic Traits in
Astyanax mexicanus
Constructive Traits
Regressive Traits
Jaws
Eyes
Taste buds
Melanophores
Feeding posture behavior
Melanin
Forebrain
Optic tectum
Hypothalamus
Rib-bearing vertebrae
Orbit bones
Aggressive behavior
Superficial neuromasts
Alarm behavior
Fat deposition
Growth control
Maxillary teeth
Scales
Nares
Olfactory interneurons
Vibration attraction behavior
Fin rays
restricted portion of their range. The hotspot of cave¬
fish distribution is the Sierra Madre Orientale (Sierra
de El Abra, Sierra de Guatemala, and Micos regions)
in the northeastern Mexican states of Tamaulipas and
ENCYCLOPEDIA OF CAVES
38
A
San Luis Potosi (Fig. IB; Mitchell et al., 1977). There is
a single outlying cavefish population located across
the continental divide in the southern Mexican state of
Guerrero. No troglomorphic cavefish have been dis¬
covered in other areas of Mexico or Central America.
This suggests that the caves of northeastern Mexico
may have unknown features that are ideal for cavefish
evolution.
Originally only three Mexican characid cavefish
populations were known, and these were classified as
members of a distinct genus consisting of three species:
Anoptichthys jordani, A. antrobius, and A. hubbsi. Based on
molecular studies and the ability to produce fertile off¬
spring, all cavefish populations and nearby surface fish
are now considered to be a single species: Astyanax mexi-
canus. Thirty Astyanax mexicamis cavefish populations
with varying degrees of troglomorphic features have
been identified (Fig. 1). At least four separate cavefish
lineages were established in the El Abra, Guatemala,
Micos, and Guerrero areas and evolved troglomorphic
traits independently (Jeffery, 2009a). Further troglo¬
morphic evolution occurred as the founders of these
lineages dispersed underground and became isolated in
separate cave systems ( e.g ., El Abra, Fig. IB).
The 30 Astyanax mexicanus cavefish populations are
named after their caves of origin (Mitchell et al., 1977).
For example, Pachon, Molino, Rio Subterraneo, and
Granadas cavefish are found in La Cueva de El Pachon
(Tamaulipas), El Sotano de Molino (Tamaulipas),
La Cueva de la Rio Subterraneo (San Luis Potosi), and
La Cueva de las Granadas (Guerrero), respectively
(Fig. 1). The Rio Subterraneo and Granadas cavefish are
considered as examples of phylogenetically young popu¬
lations, based on minimal reductions in eye size and
amount of pigmentation, whereas the Pachon and
Molino cavefish are considered phylogenetically old,
with eyes and pigmentation regressed to the maximal
extent. Troglomorphic traits range between these
extremes in other cavefish populations.
In some cases, most famously Chica cavefish, there
has been subsequent introgression between cavefish
and surface fish to form hybrid populations. La Cueva
Chica is located near a large surface stream, Rio
Tampaon, from which surface fish enter the cave via
underground conduits and hybridize with cavefish.
Chica cavefish have been collected and sold in pet
stores as blind Astyanax cavefish, but are not recom¬
mended for most research purposes because of their
recent hybrid background.
TROGLOMORPHIC TRAITS
evolved in cavefish (Table 1, Fig. 3). The most conspicu¬
ous troglomorphic traits are the loss or reduction of
eyes and pigmentation. Eye and pigment regression can
be considered developmental defects. Although the eye
primordia and body pigment cell precursors initially
develop in cavefish embryos, they fail to complete dif¬
ferentiation and eventually degenerate (Fig. 2). Some
parts of the cavefish embryonic eye are more degenerate
than others, suggesting a degree of independence in
their troglomorphic evolution. The lens is reduced to a
tiny vestige and the cornea and ciliary body are absent.
Although a layered retina is initially formed, retinal
growth eventually ceases and the layering becomes dis¬
organized. Photoreceptor cells initially differentiate but
subsequently disappear and are not replaced in adults.
Neurites grow from the retinal ganglion cells to the
1 Day
3 Days
1 Week
2 Weeks
3 Weeks
1 Month
Adult
Troglomorphic traits are the sum of constructive and
regressive changes {e.g., gains and losses) that have
FIGURE 2 Astyanax mexicanus surface fish (SF) and Pachon cave¬
fish (CF) shown from top to bottom at selected developmental stages
and as adults.
ENCYCLOPEDIA OF CAVES
ASTYANAX MEXICANVS: A MODEL ORGANISM FOR EVOLUTION AND ADAPTATION
39
Neuromasts Jaws
FIGURE 3 Constructive traits. (A,B) The numbers and diameters
(insets) of superficial neuromasts (green dots) are increased in a
region below the orbit (white outlined areas) in Pachon cavefish (B)
relative to surface fish (A). Views are of the head from a lateral side.
(C,D) Jaw span and protrusion is increased in Pachon cavefish (D)
relative to surface fish larvae (C). Views are of the head from the
ventral side with jaw and pharyngeal cartilage stained blue.
Reprinted with permission from Yamamoto et at. (2009).
optic tectum through the optic nerve but they also prob¬
ably disappear in adults. In concert with visual decay,
cavefish have smaller optic tecta, the paired regions of
the brain that transmit visual information from the eyes.
There are also differential effects on three types of pig¬
ment cells: black melanophores, silver iridophores, and
orange xanthophores. The number of melanophores is
reduced to various extents in different cavefish popula¬
tions. Furthermore, black melanin pigment is absent in
most Pachon and Molino cavefish, resulting in albinism.
Changes in iridophores and xanthophores, if any, are
less severe than those affecting melanophores.
Other regressive traits in cavefish are a reduced
metabolic rate, which is presumably beneficial for
energy conservation, reductions in the size of scales
and number of fin rays, and loss of one or two rib¬
bearing vertebrae. Fewer vertebrae lead to a shorter
body axis relative to surface fish. The advantage of a
shorter body is not clear. Finally, some of the beha¬
viors typical of surface fish are absent in cavefish.
These include schooling behavior, which is based on
vision, the scattering (or alarm) reaction, which is trig¬
gered by a body secretion warning conspecific fish
schools of predator approach, and aggressive behavior.
The absence of large predators in caves could facilitate
these regressive behavioral changes.
Constructive traits include those that are beneficial
for feeding, such as larger jaws (Fig. 3C,D), more taste
buds, and additional maxillary teeth, or to meet the
challenges of navigation in the dark, hallmarked by an
enhanced sensory neuromast system (Fig. 3A,B).
Changes in jaw shape promote feeding or prey capture
in different ways: small jaws are useful for biting, which
occurs in surface fish, whereas large protruding jaws are
used for sucking, which is prevalent in cavefish. Taste
buds are employed to distinguish food from nonfood
items as cavefish graze on the sediment of cave pools.
There is also a dramatic change in posture that is related
to substrate feeding. Cavefish glide smoothly across the
substrate at a 45° angle using their shovel-like jaws to
efficiently scoop and sample detritus. In contrast, surface
fish feed primarily in the water column guided by
vision. When surface fish are forced to search for food
on the bottom of a darkened laboratory tank, they
approach the substrate at a 90° angle, awkwardly swivel
on their body axis, and pick up food inefficiently.
Competition between surface fish and cavefish for lim¬
ited food in lighted or darkened tanks has shown that
surface fish can outcompete cavefish in the light, but
cavefish can obtain food more successfully in the dark.
In northeastern Mexico, there is a rainy season,
when food supply in caves may be abundant, and a
dry season, when food may be minimal. Probably to
survive such fluctuations, cavefish have accumulated
fat deposits under the skin throughout the body and
especially in the empty eye sockets. Thus, the ability to
detect and consume different types of food and to
store energy may be required for adaptation to the
cave environment.
Navigation without vision in a dark cave is accom¬
plished by increasing the size and number of motion¬
detecting neuromasts. As in other teleosts, Astyanax
mexicanus has canal neuromasts located in the lateral
lines and superficial neuromasts distributed freely in the
head and body (Fig. 3A,B). An increase in superficial
neuromasts is the cause of a behavior that attracts cave¬
fish to water vibrations (Yoshizawa et al., 2010).
Vibration attraction behavior is strong in cavefish but
weak and rarely seen in surface fish, probably because
of the risk in attracting large predators. Attraction to
vibrations confers an advantage for feeding in darkness
and could have been one of the first evolutionary
changes that resulted in adaptation to the cave habitat.
An enhanced neuromast system also allows cavefish
to "memorize" their complex physical environment.
Astyanax mexicanus perform exploratory swimming
behaviors when challenged with new environments that
later allow them to avoid obstacles, and this spatial ori¬
entation behavior is enhanced in cavefish.
The ventral forebrain and hypothalamus are enlarged
in cavefish compared to surface fish (Retaux et al..
ENCYCLOPEDIA OF CAVES
40
A
2008). One of the reasons for this increase may be to
match gustatory processing centers in the central ner¬
vous system with the development of additional taste
buds. Ventral forebrain changes also involve an increase
in the number of GABAeric olfactory interneurons,
which are crucial for olfactory learning and discrimina¬
tion. This constructive change in olfaction is coupled
with a greater size, and presumably receptor capacity,
of cavefish nasal areas.
The adaptive significance of some constructive char¬
acters is more difficult to understand. For example, the
surface fish craniofacial skeleton includes six flat bones
surrounding each orbit. Cavefish have a different
craniofacial organization: their orbital bones expand
over the empty eye sockets and are subdivided into
numerous smaller bones. The changes in craniofacial
structure are due to the absence of a fully differenti¬
ated eye, which negatively influences orbital bone
growth in surface fish. Surface fish-like craniofacial
features can be restored in cavefish by swapping its
embryonic lens for that of a surface fish (see below).
As a by-product of eye degeneration, craniofacial reor¬
ganization may have no adaptive significance.
DEVELOPMENTAL BASIS OF
TROGLOMORPHIC TRAITS
The primary focus of developmental studies in the
Astyanax mexicanus model system has been to determine
the molecular mechanisms of eye and pigment degener¬
ation (Jeffery, 2009b). These studies have mostly been
done with Pachon cavefish.
As in other teleosts, the embryonic eyes of Astyanax
mexicanus consist of two major parts: the lens, which is
derived from the surface ectoderm, and the optic cup,
an out-pocketing of the brain. The optic cup is the pre¬
cursor of the retina and its melanin-pigmented epithe¬
lium (the retinal pigment epithelium or RPE). Eyes
with a smaller lens and optic cup relative to surface
fish initially develop in cavefish embryos but they sub¬
sequently cease growth, degenerate, and disappear
into the orbits (Fig. 2). The pax6 gene, which encodes a
transcription factor, is a key regulator of eye develop¬
ment in vertebrates. The small cavefish optic cup
shows reduced pax6 expression relative to surface fish,
which is caused by upregulation of the sonic hedgehog
(shh) gene along the anterior embryonic midline. Shh
does not affect pax6 expression directly. Instead Shh
proteins diffuse from the midline, resulting in upregu¬
lation of the pax2 and vaxl genes, which in turn nega¬
tively regulate pax6 in the optic cup.
The expression of pax6 is also downregulated in the
cavefish lens. The cavefish lens is formed just as in sur¬
face fish, but lens fiber cell differentiation does not
occur. Instead of differentiating, the cavefish lens
undergoes programmed cell death. The aA-crystallin
gene, which encodes a cell survival protein, is downre¬
gulated, and the heat shock protein 90a ( hsp90a ) gene,
which encodes a cell death-promoting protein, is upre-
gulated in the cavefish lens. Together, these changes in
gene expression trigger lens cell death in cavefish. The
lens normally induces the anterior parts of the eye,
including the cornea and ciliary body. Therefore, in
the absence of a functional lens these optic components
are missing in cavefish.
Differentiation of the retina and RPE are also initi¬
ated in cavefish. Neural and glial cells differentiate
normally during retinal development but this is not
the case for rod photoreceptor cells. Cavefish rod cells
are formed with abnormally shortened outer seg¬
ments, but they subsequently disappear and are not
replaced. Parts of the cavefish retina also undergo cell
death and its layers become extremely disorganized.
Melanin pigmentation appears in the RPE in some
cavefish populations, although their RPE does not
seem to differentiate normally.
Eye growth continues throughout life in teleosts due
to the division of stem cells in the lens and the outer
edge of the retina/RPE (the ciliary marginal zone or
CMZ). The rate of cell proliferation in the cavefish lens,
however, is insufficient to replace the dying cells, and
consequently the lens becomes progressively smaller
during larval development. In contrast, there is a high
rate of cell division in the cavefish CMZ. However, the
absence of net growth in the retina is due to death of
the newly born cells before they can contribute to its
structure. Why the cavefish retina continues to produce
new cells, presumably at high energetic cost, only to
subsequently destroy them, is a paradox. The final size
of the vestigial (internalized) eye is variable both
within and between different cavefish populations.
The critical role of Shh in eye degeneration has been
demonstrated by shh gene overexpression in surface
fish. Similar to cavefish, Shh upregulation causes
increases in pax2 and vaxl expression and a decrease
in pax6 activity, reducing the size of the surface fish
optic cup (Yamamoto et al., 2004). In addition,
Shh overexpression induces cell death in the surface
fish lens. Thus, a surface fish mimic of cavefish eye
degeneration can be made by Shh overexpression.
The process of eye regression in cavefish is summa¬
rized as follows. First, Shh is over expressed along the
embryonic midline. Second, the increase in Shh signal¬
ing down regulates pax6, which reduces the size of the
optic cup and induces lens cell death. Third, the non¬
functional lens is either not able to induce other eye
tissues or to protect them from cell death and thus
allows them to mount a sufficient growth rate to match
the increasing body size.
ENCYCLOPEDIA OF CAVES
ASTYANAX MEXICANVS: A MODEL ORGANISM FOR EVOLUTION AND ADAPTATION
41
There are three ways in which the cavefish eye can
be rescued from degeneration. First, eyes and vision can
be restored in hybrids produced by crossing members
of independently evolved cavefish lineages. Rescue by
genetic complementation shows that some of the genes
responsible for eye degeneration are different in these
lineages. Second, inhibition of lens cell death with the
Hsp90 antagonist radicicol can partially rescue eye
development, indicating that the cell death program is
reversible. Third, the cavefish eye can be rescued by lens
transplantation (Yamamoto and Jeffery, 2000). A surface
fish lens transplanted into the cavefish optic cup evades
cell death and survives to influence the development of
other eye parts, including the cornea, ciliary body, and
retinal photoreceptor cells. Moreover, in the presence of
a surface fish lens, cell death is prevented in the cavefish
retina, allowing it to resume growth and form a large
external eye, indicating that cavefish have retained the
instructions for optic development.
Reduction or elimination of melanophores is another
major regressive trait. Melanophores and other body
pigment cells are formed from the neural crest, which
originates in surface ectoderm adjacent to the neural
plate. During neural tube formation, neural crest cells
leave the surface and migrate long distances through the
embryo to eventually differentiate into many different
adult tissues, including spinal ganglia, the peripheral
nervous system, facial bones, and body pigment cells. It
seems unlikely that such an important cell type could
be modified without lethality. Accordingly, no defects
in neural crest cell formation or migration have been
discovered in cavefish embryos. Instead, the block in
pigmentation is beyond this point at the stage of mela-
nophore differentiation.
During melanophore differentiation, melanin is syn¬
thesized within the melanosome, a membrane-bound
organelle, via a series of well-known biochemical reac¬
tions (Fig. 4A). The amino acid L-tyrosine is trans¬
ported into the melanosome, where it is converted to
l-DOPA by tyrosinase. l-DOPA is then converted into
melanin by a cascade of enzymatic reactions, catalyzed
either by tyrosinase or other enzymes. In albino cave¬
fish, melanin synthesis can be rescued by supplying
exogenous l-DOPA (Fig. 4B), but not L-tyrosine.
Therefore, cavefish melanophores lack melanin not
because they are deficient in tyrosinase or downo-
stream enzymes but instead because they do not con¬
tain L-tyrosine. A defect at the first step in the melanin
synthesis pathway has evolved independently in
several different cavefish populations.
Developmental studies of constructive traits have
focused on jaws and taste buds, which require shh
gene expression (Yamamoto et al., 2009). During
mouth development, shh is first expressed throughout
the oral epithelium and later attenuated to the
developing taste buds. Experimental downregulation
of shh in cavefish or surface fish embryos results in
smaller jaws with fewer taste buds. Reciprocally, shh
upregulation results in larger jaws with more taste
buds. Thus, the pleiotropic shh gene functions in both
feeding apparatus enlargement and eye degeneration,
revealing an important antagonistic linkage between
these constructive and regressive traits.
INHERITANCE AND GENETIC BASIS OF
TROGLOMORPHIC TRAITS
Eye degeneration is controlled by mutations in 6—12
different genes, each with a small role in the process
(Jeffery, 2009a). Quantitative trait locus (QTL) analysis
has been used to map eye regression traits and to iden¬
tify candidate genes. One of the eye candidate genes is
aA-crystallin (see above) whose downregulation may
affect the survival of the cavefish lens. Other candidate
genes for eye degeneration are ROM1, which is
required for normal development of rod photoreceptor
cells, and Shroom2, which regulates melanosome locali¬
zation in the RPE. The over expressed shh gene is not
(A)
L-Tyrosine
_1
L-DOPA
1
DOPAquinone
l
LeucoDOPAchrome
l
a n DOPAchrome
V \
DICA DHI
1 1 _
lndole-5,6-quinone Indole-
carboxylic acid 5,6-quinone
E3\ /
Melanin
FIGURE 4 (A) Melanin biosynthetic pathway showing substrates
and positions of enzyme and Oca2 function. (B) Albino Pachon cave¬
fish embryo provided with exogenous l-DOPA substrate rescues
melanin-containing pigment cell precursors (TPC).
ENCYCLOPEDIA OF CAVES
42
A
represented by a QTL, suggesting that upstream regu¬
latory genes, rather than shh itself, have mutated in
cavefish (Protas et al., 2008). The identification of addi¬
tional candidate genes remains challenging in the
absence of a sequenced Astyamx mexicanus genome.
Pigmentation changes are inherited independently
from eye regression, as indicated by segregation of these
two traits in the F2 progeny of a cavefish X surface fish
cross (Jeffery, 2009a). Multiple genes with additive
effects also decrease melanophore number. Similar to
eyes, the genes responsible for melanophore reduction
can be different in various cavefish populations. The
number of genes controlling changes in pigmentation is
potentially high: 18 different QTL affect melanophore
pigmentation (Protas et al., 2007). Included in the many
QTL and predicted pigmentation genes are two traits
with relatively simple inheritance: albinism and the
brown eye phenotype.
Cavefish albinism is inherited as a recessive trait con¬
trolled by oculocutaneous albinism 2 local), a single gene
that also has a major role in mouse and human albin¬
isms (Protas et al., 2006). Loss of oca2 function has
occurred in both Pachon and Molino cavefish, but
through different mutations: large deletions encompass¬
ing two distinct regions in the protein coding sequence.
Thus, for an unknown reason, the oca2 gene is repeat¬
edly targeted for mutation, resulting in independent ori¬
gins of albinism. The Oca2 protein may control
accessibility of L-tyrosine to tyrosinase for conversion to
l-DOPA, perhaps by serving as a transport channel in
the melanosome membrane.
Mutations in the brown gene decrease the number of
melanophores and change pigmentation from black to
brown in the (degenerate) eyes of many different
cavefish populations. Genetic analysis predicted a
single gene responsible for the brown phenotype, and
a single QTL was identified (Gross et al., 2009). The
broum gene encodes the melanocortin 1 receptor
(Mclr), a component of the Melanosome Stimulating
Hormone-a (MSH-a) signaling system, which is
involved in changing pigmentation in many different
vertebrates. Furthermore, different mutations in the
Mclr gene, either in the coding or regulatory regions,
are responsible for the brown eye phenotype in vari¬
ous cavefish populations, and Mclr is another example
of a gene that has been mutated repeatedly to generate
the same troglomorphic phenotype.
Several mutated genes are responsible for taste-bud
amplification but as in the case of eyes, none of these
is likely to be shh (Protas et al., 2008). This indicates
that the effects of Shh on taste-bud development also
occur downstream of the mutated genes. In contrast to
the large number of genes involved in eye and pig¬
ment degeneration, only three gene loci, one major
gene and two modifiers, are required to change
feeding behavior. This is surprising because of the
complexity thought to be associated with behavioral
traits.
EVOLUTION OF TROGLOMORPHIC
TRAITS
Constructive troglomorphic traits are likely to have
evolved by natural selection. The best example is the
attraction to vibrations, which is rare in surface fish
but robust in cavefish. Vibration attraction behavior
has a genetic basis and is beneficial for feeding, allow¬
ing it to be influenced by natural selection. The evolu¬
tionary forces controlling regressive traits are less
clear. Two major hypotheses have been proposed: (1)
accumulation of neutral mutations and genetic drift;
and (2) negative selection based either on energy con¬
servation or indirect pleiotropy involving a tradeoff
with constructive traits.
In the case of eye loss, selection seems to be sup¬
ported over neutral mutation, although all cavefish
researchers do not agree with this conclusion. All known
eye QTL decrease eye size (Protas et al., 2007). In con¬
trast, the accumulation of random neutral mutations
might be expected to increase as well as decrease the
size of the eyes. However, decreases in eye function
could also be explained by the fact that most natural
mutations are recessive. An additional consideration is
that genes such as shh and hsp90a increase rather than
decrease expression activity during cavefish eye devel¬
opment. Particularly in the case of shh, eye loss could be
the result of indirect selection involving a tradeoff for
the development of larger jaws and more taste buds at
the expense of eyes. Because many genes are mutated to
give rise to eye loss, it is also possible that one or more
evolutionary forces are acting in concert on this complex
troglomorphic trait.
Reduced pigment cell number is probably caused
by neutral mutation and genetic drift (Protas et al.,
2008). Genetic analysis shows that individual QTL can
increase or decrease melanophore abundance. When
albinism is considered, however, the picture becomes
less clear. On one hand, neutral mutations and subse¬
quent genetic drift could have occurred if oca2 is not
pleiotropic and thus does not affect other developmen¬
tal pathways. On the other hand, the position of Oca2
function at the beginning of the melanin biosynthesis
pathway suggests another possibility (Fig. 4A). This is
a critical branch point in which L-tyrosine can be
converted to either (1) l-DOPA and melanin or (2) to
dopamine, which could have adaptive significance in
changing feeding behaviors. Therefore, inhibition
of melanin synthesis at a step that would provide
more L-tyrosine substrate for dopamine synthesis
ENCYCLOPEDIA OF CAVES
ASTYANAX MEXICANVS: A MODEL ORGANISM FOR EVOLUTION AND ADAPTATION
43
could be another example of the role of indirect natu¬
ral selection via pleiotropic tradeoff as a driving force
for evolution and adaptation.
See Also the Following Articles
Adaptation to Low Food
Adaptive Shifts
Gammarus minus: A Model System for the Study of
Adaptation
Niphargus: A Model System for Evolution and Ecology
Bibliography
Gross, J. B., Borowsky, R., & Tabin, C. J. (2009). A novel role for
Mclr in the parallel evolution of depigmentation in independent
populations of the cavefish Astyanax mexicanus. PLoS Genetics, 5
(1), el 0000326.
Jeffery, W. R. (2009a). Regressive evolution in Astyanax cavefish.
Annual Reviews of Genetics, 43, 25—47.
Jeffery, W. R. (2009b). Evolution and development in the cavefish
Astyanax. Current Topics in Developmental Biology, 86, 191—221.
Mitchell, R. W., Russell, W. H., & Elliot, W. R. (1977). Mexican eye¬
less characin fishes, genus Astyanax: Environment, distribution.
and evolution. Special Publication of the Museum of Texas Tech
University, 12, 1—89.
Protas, M., Hersey, C., Kochanek, D., Zhou, H., Wilkens, H., Jeffery,
W. R., et al. (2006). Genetic analysis of cavefish reveals molecular
convergence in the evolution of albinism. Nature Genetics, 38(1),
107-111.
Protas, M., Conrad, M., Gross, J. B., Tabin, C. J., & Borowsky, R.
(2007). Regressive evolution in the Mexican cave tetra, Astyanax
mexicanus. Current Biology, 17(5), 452—454.
Protas, M., Tabansky, I., Conrad, M., Gross, J. B., Vidal, O., Tabin,
C. J., et al. (2008). Multi-trait evolution in a cave fish, Astyanax
mexicanus. Evolution & Development, 10(2), 196—209.
Retaux, S., Pottin, K., & Alunni, A. (2008). Shh and forebrain evolu¬
tion in the blind cavefish Astyanax mexicanus. Biology of the Cell,
110(3), 139-147.
Yamamoto, Y. & Jeffery, W. R. (2000). Central role for the lens in
cave fish eye degeneration. Science, 289(5479), 631—633.
Yamamoto, Y., Stock, D. W., & Jeffery, W. R. (2004). Hedgehog sig¬
nalling controls eye degeneration in blind cavefish. Nature, 431
(7010), 844-847.
Yamamoto, Y., Byerly, M. S., Jackman, W. R., & Jeffery, W. R. (2009).
Pleiotropic functions of embryonic sonic hedgehog expression link
jaw and taste bud amplification with eye loss during cavefish
evolution. Developmental Biology, 330(1), 200—211.
Yoshizawa, M., Goricki, S., Soares, D., & Jeffery, W. R. (2010).
Evolution of a behavioral shift mediated by superficial neuro¬
masts helps cavefish find food in darkness. Current Biology, 20
(18), 1631-1636.
ENCYCLOPEDIA OF CAVES
BATS
Thomas H. Kunz, Susan W. Murray,
and Nathan W. Fuller
Boston University
INTRODUCTION
Bats (order Chiroptera) are an ecologically diverse and
geographically widespread mammalian taxon. With
1232 extant species currently described (Simmons, per¬
sonal communication), members of two suborders
(Yinpterochiroptera and Yangochiroptera) constitute
approximately one-fifth of all living mammals. They
range in size from the tiny 2-gram Kitti's hog-nosed
bat (Craseonycteris tlionglogyai) to the large Malayan fly¬
ing fox (Pteropus vampyrus) weighing about 1200
grams, with a wingspan of approximately 1.5 meters.
Bats are characterized by several life-history traits that
make them unique among mammals. For example, com¬
pared to small terrestrial species, bats of similar size
have very few young and long periods of pregnancy and
lactation, and some are known to live up to 42 years in
the wild (Kunz and Parsons, 2009). Differences in
life-history traits between bats and other small mammals
are often attributed to the evolution of flight and echolo-
cation (Crichton and Krutzsch, 2000; Kunz and Fenton,
2003).
Bats are the only mammals that have evolved pow¬
ered flight. Thus, along with echolocation, these two
traits have made it possible for bats to exploit many
different types of roosting habitats ( e.g ., foliage, tree
cavities, caves, rock crevices) that are generally inac¬
cessible to most terrestrial mammals. Echolocation and
flight have also made it possible for bats to exploit an
unprecedented variety of food sources, including fruit,
nectar, pollen, leaves, seeds, invertebrates {e.g., insects,
spiders, crustaceans, and scorpions), small vertebrates
{e.g., fish, frogs, lizards, birds, and other small mam¬
mals), and blood (Kunz and Fenton, 2003).
Although echolocation has evolved independently
in birds and mammals, and even within different
groups of mammals {e.g., bats, shrews, cetaceans), the
most sophisticated and diversified form of echoloca¬
tion can be found in laryngeal-emitting species of bats.
Echolocation is used for prey detection and capture,
for navigation, and in some instances for communica¬
tion. In additon to echolocation, several species may
rely on a combination of vision, olfaction, and prey¬
generated sounds to locate food (Kunz and Fenton,
2003).
The roosting habits of bats are often highly special¬
ized, with different species occupying tree cavities;
spaces beneath exfoliating bark; unmodified foliage;
leaves modified into so-called tents; abandoned ant,
termite, and bird nests; large and small caves; rock cre¬
vices; and a wide variety of manmade structures,
including mines, old and new buildings, ancient stone
ruins, and bridges (Kunz, 1982; Kunz and Fenton,
2003). Caves alone provide a variety of structural
substrates for roosting, including crevices, cavities,
textured walls and ceilings, massive rooms, rock
outcrops, and rock rubble on floors. In addition, the
microclimates of caves occupied by bats can vary enor¬
mously, depending on latitude, altitude, depth, and
volume, as well as the number, size, and position of
openings to the outside. These variables can influence
the amount of airflow, the presence of flowing and
standing water, and daily and seasonal variations
in atmospheric pressure, temperature, and humidity.
Thus, the environmental conditions within caves may
be hot, cold, dry, humid, still, or windy, although dif¬
ferent species appear to select parts of caves and cave¬
like structures that minimize energy expenditure and
water loss, and that facilitate an array of social
interactions.
This article highlights the biology of bats that roost in
caves and cave-like structures. Specifically, we briefly
Encyclopedia of Caves.
45
©2012 Published by Elsevier Inc.
46
B
discuss why bats live in caves, where they are found,
their roosting requirements, threats from moving wind
turbines and emerging diseases, as well as conservation
and management issues important to protecting
cave-roosting species.
CAVE BATS^AND THEIR DISTRIBUTION
Cave bats are defined as trogloxenes, species that
complete part of their life cycle outside of caves. Their
ability to fly and echolocate has allowed many species
to exploit caves and similar subterranean habitats for
roosts and to forage for food often at considerable dis¬
tances from these structures. All bats that belong to the
suborder Yangochiroptera and orally echolocating spe¬
cies belonging to the suborder Yinpterochiroptera are
able to navigate in dark spaces and to feed in habitats
ranging from open fields, high altitudes, and dense for¬
ested areas to over lakes and streams. Rousette fruit bats
(family Pteropodidae; suborder Yinpterochiroptera)
may also roost in caves, but instead of using vocal cords
to produce echolocation calls, they rely on tongue clicks
to produce audible sounds to facilitate navigation in the
dark. The handful of other pteropodids found roosting
in caves are restricted to areas where enough ambient
light makes it possible for them to navigate in and out,
as they largely rely on vision and not echolocation while
in flight (Kunz and Fenton, 2003).
Bats are virtually ubiquitous; they are known on all
continents, except Antarctica, and from many archipel¬
agos and oceanic islands (Kunz, 1982; Fleming and
Racey, 2010). Most bats occur in tropical regions,
where they are often the most diverse and abundant
mammals present. The diversity of bats generally
increases as one travels from the poles toward the
equator, a pattern that is largely attributable to an
increase in habitat complexity as latitude decreases
(Kunz and Fenton, 2003).
The geographic distribution of cave-roosting bats
not only depends on the presence of caves (Palmer
and Palmer, 2009), but is also a consequence of specific
roosting requirements (Kunz, 1982). For example,
although the ghost bat ( Macroderma gigas), the orange
leaf-nosed bat ( Rhinonycteris aumntius), and the large
bent-wing bat ( Miniopterus schreibersii) are all found in
Australian caves, their roosting requirements and
hence geographic distributions are markedly different.
The orange leaf-nosed bat selects caves that are
extremely hot and humid (28 to 30°C and greater than
94% relative humidity) and are known from only 10
caves in Australia. In contrast, the large bent-wing bat
can be found roosting at a broader range of temperature
and humidity and has one of the widest reported distri¬
butions of cave-roosting bats, encompassing southern
FIGURE 1 Townsend's big-eared bat (Corynorhinus toumsendii).
Photo by J.S. Altenbach. Used with permission.
Europe, Africa, Southeast Asia, Japan, and Australia
(Zubaid et al., 2006). Similarly, different species of bats
in the West Indies select caves whose temperatures may
range from 18 to 40°C (Fleming and Racey, 2009). The
presence and abundance of caves with characteristics
allowing the creation of gradients in temperature will
strongly influence the presence and abundance of
different bat species.
Bats are found almost everywhere subterranean
spaces exist. The distributions of cave-dwelling bats are
determined largely by species-specific roosting require¬
ments that vary depending on their ecology and evolu¬
tionary history. Local and global distributions and
densities of bats that rely on caves for at least part of
their life cycle are in turn determined largely by the dis¬
tribution, quantity, and characteristics of available
caves (Palmer and Palmer, 2009). For example, the
Townsend's long-eared bat ( Corynorhinus toumsendii),
endemic to North America (Fig. 1), almost exclusively
roosts in caves and cave-like structures (Kunz, 1982).
FUNCTIONS OF CAVE ROOSTS
Bats roost in caves for a variety of reasons, including
courtship and mating, raising young, and hibernation.
They seek shelter during the day and disperse from
these sites to forage for food at night. During the day,
bats typically sleep, rest, groom, or interact with their
roost-mates. For example, a typical day for lactating
female lesser long-nosed bats ( Leptonycteris cnrasoae)
involves sleeping or resting quietly for up to 12 hours,
interspersed with periodic grooming and nursing.
Although female L. curasoae usually roost together in
caves during the day, they seldom interact with
males, except during a brief mating season in November
and December (Munoz-Romo and Kunz, 2009). The com¬
mon vampire bat ( Desmodus rotundus) forms long-term
ENCYCLOPEDIA OF CAVES
BATS
47
social bonds, and individuals groom one another as they
interact socially while occupying both cave and tree
roosts (Crichton and Krutzsch, 2000; Kunz and Fenton,
2003). In addition, many insectivorous species retreat to
caves between feeding bouts, where they may cull the
wings and heads of insects that were captured while
foraging. Frugivorous species sometimes transport large
fruits to caves where they cull soft pulp and where they
can reduce the risks of predation. Some species that roost
in foliage or tree cavities during warm months typically
hibernate in caves and cave-like structures during the
winter (Kunz, 1982; Kunz and Fenton, 2003; Zubaid
et ah, 2006).
Courtship and Mating
Several types of mating systems have been described
for bats (Crichton and Krutzsch, 2000). Mating systems
of bats and other mammals are often classified into
three general categories: promiscuity, polygyny, and
monogamy. Mating systems of bats, however, cannot
be easily categorized into one of these groups, as
they often exhibit a continuous spectrum of mating
behaviors.
Promiscuity is a type of mating system in which
both males and females have multiple partners.
Flowever, such a system is almost always highly struc¬
tured, with some males siring more offspring than
others. Promiscuity is common among temperate cave-
roosting species, possibly because of the limited time
available for mating in autumn before individuals
enter hibernation. Males of most temperate species
generally do not roost with females during warm
months but instead roost alone or in small groups.
Assemblages of males and females from temperate
regions often gather at caves and mines in the autumn
(referred to as swarming behavior), a time that also
coincides with courtship and mating behavior. During
the swarming period, bats are active in caves and
mines at night, where males can often be observed dis¬
playing and chasing females. Male and female greater
horseshoe bats ( Rhinoloplius ferrumequinum) have a
mating system in which males establish territorial sites
inside caves and mines in early autumn. Females
gather at these sites and selectively visit and mate with
different males in their territories (Crichton and
Krutzsch, 2000).
Polygyny, a mating system that is most common in
bats, is characterized by one male mating with several
females (Crichton and Krutzsch, 2000). An example of
this type of mating system can be observed in the
cave-roosting greater spear-nosed bat ( Phyllostoinus
lwstatus). In this species, females roost in small
stable groups, often remaining together for 10 years or
more. Females often form discrete roosting groups in
solution cavities or "potholes" on cave ceilings, where
a dominant male typically defends a group of females
from intrusions by other males. By defending the
females, or the roost cavity, a so-called harem male is
often able to mate with several females. Sometimes
these harem males are accompanied by a subordinate
male who positions himself in the harem to assume a
dominant role if the harem male should become
injured or die. The risks and costs associated with
mate-guarding behavior can be substantial. For exam¬
ple, a harem male greater spear-nosed bat may incur
some injuries while defending the females or roost
cavity and may only sire 60 to 90% of the young born
to those females. A similar pattern of mate guarding
and courtship has been observed in the Jamaican fruit
bat ( Artibeus jamaicensis) , which commonly roosts in
caves on many of the islands in the West Indies and
throughout Central and South America (Crichton and
Krutzsch, 2000).
Monogamy occurs when males and females form
long-term pair bonds. This type of mating system has
been described for only a few species of bats. Two
examples are the African false vampire bat ( Cardioderma
cor) and the American false vampire bat ( Vampyrum
spectrum), both of whom are carnivorous, sit-and-wait
predators. An extended period of parental involvement
in which males provision both females and young may
have contributed to the evolution of monogamy in these
and other species. To our knowledge, lek mating behav¬
ior, in which males assemble and display simulta¬
neously to attract females, has not been observed in
cave-roosting bats (Crichton and Krutzsch, 2000).
Rearing Young
During the pregnancy and lactation periods, females
form maternity colonies, which are often located in
separate places from roosts used by males. In most
species of bats, the responsibility of raising young lies
solely with females. Pregnancy and lactation are both
energetically expensive events; thus females and their
young can benefit from the heat generated when they
form dense clusters in partially enclosed spaces often
found in tree cavities, buildings, caves, and cave-like
structures. Roosting together in large clusters may
reduce the energy expenditure of some individuals by
up to 50%. When lactating females disperse from
roosts in the evening to feed, they often leave their
pups in a warm, incubator-like environment. Females
incur high energy costs when they forage and return
to the roosts one to three times each night to find and
suckle their dependent young. Thus, assembling in
warm places can help reduce the energy needed by
small bats to remain euthermic (Kunz, 1982; Crichton
and Krutzsch, 2000; Kunz and Fenton, 2003). The
ENCYCLOPEDIA OF CAVES
48
B
formation of multispecies assemblages of bats may
also contribute to reduce energy expenditure. By roost¬
ing in the same cave, species with different activity
patterns can form larger assemblages, and thus con¬
tribute to a higher ambient temperatures, and reduced
energy expenditure (Fleming and Racey, 2009).
Each spring, Brazilian free-tailed bats ( Tadarida brasil-
iensis ) migrate from Mexico to the southwestern United
States to form large maternity colonies in caves and
sometimes other structures such as buildings and
bridges. This species is thought to form some of the larg¬
est aggregations of mammals known to humankind,
where a single colony may exceed several million indi¬
viduals in single cave (Fig. 2). Each time a female
Brazilian free-tailed bat returns from a feeding bout to
suckle her young, she faces the daunting task of finding
her own pup among the millions of babies that are left
on the ceilings and walls of the cave. A mother bat
begins this adventure using spatial memory by returning
to the area in the cave where she left her pups before
emerging to feed. She then uses vocal and olfactory
cues to ultimately identify her own pup among the thou¬
sands or more present in the immediate surroundings
(Crichton and Krutzsch, 2000). Hungry pups will some¬
times attempt to nurse from almost any female, although
lactating females usually guard against milk stealing
from unrelated individuals. The investment that a
mother bat makes in her pups is substantial, often
requiring quantities of food intake equal to her entire
body mass each night during peak lactation (Kunz and
Parsons, 2009).
Young Brazilian free-tailed bats grow rapidly from
a diet of energy-rich milk. Mothers nourish their
young with milk for several weeks, because young
bats cannot fly and feed on their own until their wings
have almost reached adult dimensions. Within 6 weeks
of birth, young Brazilian free-tailed bats are able to fly
FIGURE 2 Emerging Brazilian free-tailed bats ( Tadarida brasiliensis )
from a maternity cave in south-central Texas. Photo by M.D. Tuttle,
Bat Conservation International. Used with permission.
and forage on their own. In contrast to most other
mammals that typically wean their young at about
40% of adult size, most insectivorous bat species suckle
their young until they are about 90% of adult size
(Crichton and Krutzsch, 2000; Kunz and Fenton, 2003;
Kunz and Parsons, 2009).
HIBERNATION
Bats have evolved behavioral and physiological
mechanisms to avoid long periods of adverse weather
and low food or water availability. Some species
migrate to more suitable areas where food is readily
available, but others use daily torpor, a controlled low¬
ering of body temperature to conserve energy when
food resources are limited or unavailable. Only tem¬
perate species in the families Vespertilionidae and
Rhinolophidae are known to hibernate in caves and
mines (Fig. 3) (Kunz, 1982; Kunz and Fenton, 2003).
Hibernating bats rely on stored fat as their primary
energy source during hibernation and are sustained on
these reserves for upwards of 6 to 8 months. Hibernation
is an energy-saving strategy that is strongly influenced
by the ambient conditions in a cave. When a bat is hiber¬
nating, low ambient temperatures lead to a decrease
body temperature and metabolic rate. When the ambient
temperature is too cold or too warm, bats typically
arouse and move to other parts of a hibernaculum. It is
important for hibernating bats to occupy caves and
mines that provide a variety of temperatures, because
individuals often change roosting positions as the season
progresses (Kunz and Fenton, 2003; Zubaid et al., 2006).
During hibernation, bats lower their body tempera¬
ture to within a few degrees of the ambient tempera¬
ture, but individuals arouse periodically by employing
nonshivering thermogenesis to increase their body
temperature. Bouts of hibernation can last anywhere
from a few days to several weeks. In areas with mod¬
erate winters, for example, bats such as the greater
horseshoe bat (Rhinolophus ferrumequinum) may arouse
FIGURE 3 Hibernating cave bats ( Myotis velifer) in a gypsum
cave in south-central Kansas. Photo by T.H. Kunz.
ENCYCLOPEDIA OF CAVES
BATS
49
periodically and feed on insects on warm winter
nights. Arousals from deep hibernation are energeti¬
cally costly, with a single arousal expending the
energy equivalent of a bat spending 30 days in deep
torpor. Thus, if hibernating bats arouse too often,
either because the microclimate is suboptimum or
from human disturbance, they may not have enough
fat reserves to survive the winter (Kunz and Fenton,
2003; Zubaid et al., 2006).
COSTS AND BENEFITS OF LIVING
IN CAVES
The decision about where to roost is critical to the
survival and reproductive success of bats. The type of
roost that a bat selects is influenced by its morphology,
ecology, and physiological requirements and often
reflects a compromise between the costs and benefits
associated with a particular type of roost (Kunz, 1982).
For cave-roosting species, the benefits of living in a
cave usually outweigh costs that they may otherwise
incur. In the following section, we discuss the major
costs and benefits considered critical for the selection
of roosts by cave-dwelling bats. It is important to note
that roosting requirements and relevant costs and ben¬
efits are not uniform for all species and may even vary
intraspecifically, depending on geographic location,
reproductive condition, and/or season (Kunz and
Fenton, 2003; Zubaid et al., 2006).
Benefits
Caves offer a wide range of benefits to bats includ¬
ing structurally and climatically stable environments,
and protection from predators and adverse weather.
Microclimate, specifically temperature and relative
humidity, is arguably the most important factor in roost
selection by cave-dwelling bats (Kunz, 1982; Zubaid
et al., 2006). Different bat species roost in a variety of
microclimates within caves and mines, and this varia¬
tion is often correlated with body size, diet, phylogeny,
and their ability to enter torpor. Metabolically conser¬
vative species on islands are also likely to derive bene¬
fits from the stable thermal microclimate of caves, and
hurricane-prone regions appear to favor the prevalence
of cave-dwelling bats (Fleming and Racey, 2009).
Compared to nonvolant mammals, bats experience
high rates of evaporative water and heat loss, due in
large part to their relatively high surface area-to-volume
ratio, enhanced by the large surface of their naked wing
membranes. At low relative humidity (<20%), bats may
lose up to 30% of their body mass per day from evapo¬
rative water loss alone. This rapid dehydration can be
lethal. Thus, many bat species select caves that have
high relative humidity to help conserve water during
the day (Kunz and Fenton, 2003).
Bats use at least four different strategies for conserving
energy while in their roosts (Kunz and Fenton, 2003;
Zubaid et al., 2006). Some species select roosts with ambi¬
ent temperatures within their thermal neutral zone.
For example, the California leaf-nosed bat ( Macrotus
californicus ) often exploits geothermally heated mines to
conserve heat during the winter. Other species form large
colonies in areas in caves that have little airflow ( e.g .,
high-domed ceilings or dead-end chambers), which con¬
tribute to an increase in roost temperatures as the meta¬
bolic heat generated by the bats becomes trapped. The
lesser long-nosed bat ( Leptonycteris curasoae), the large
bent-wing bat (Miniopterus schreibersii) the sooty mus-
tached bat (Pteronotus quadridens), and the Brazilian
free-tailed bat ( Tadarida brasiliensis ) are examples of cave¬
dwelling species that form colonies large enough to
substantially increase the temperature of their roost envi¬
ronment. Mormoopids, natalids, and many phyllostomids
in the West Indies form multispecies assemblages that
increase the temperature of the caves in which they roost
(Fleming and Racey, 2009). During the winter months,
hibernating species, such as big brown bats (Eptesicus fus-
cus), little brown myotis ( Myotis lucifugus), Indiana myotis
(M. sodalis), and the gray bat (M. grisescens), select colder
roost environments that allow them to reduce their body
temperature and thus become torpid. Torpor not only
reduces the amount of energy a bat expends, but also
helps reduce water loss. Finally, some species form dense
clusters in winter that buffer individuals from changes in
ambient temperature, a behavior that also reduces their
energy expenditure (Kunz, 1982; Zubaid et al., 2006;
Fleming and Racey, 2009).
In addition to the energy savings that cave-roosting
bats may experience, many species also benefit from
social interactions. For example, the environmental sta¬
bility of caves often facilitates social interactions such as
finding, attracting, and guarding mates; information
transfer; and interactions that evolve through kin selec¬
tion and/or reciprocal altruism (Crichton and Krutzsch,
2000). Females that roost together sometimes share infor¬
mation about food resources, such as the location of
flowering and fruiting trees. Information transfer, pre¬
sumably, is facilitated in the greater spear-nosed bats
(Phyllostomus hastatus ) through vocal contact, which may
help females coordinate efforts to defend food patches
from conspecifics. Also, bats often emerge synchro¬
nously from caves and cave-like structures, which may
decrease an individual's risk of predation (Fleming and
Racey, 2009) as a consequence of the so-called selfish
herd effect (Kunz, 1982).
The common vampire bat ( Desmodus rotundus),
another very social species, has evolved a system of
sharing blood with both relatives and unrelated female
ENCYCLOPEDIA OF CAVES
50
B
roost-mates. Vampire bats must obtain a blood meal at
least once every three days or they will invariably die
of starvation and dehydration. Females often share
blood with roost-mates that are at risk of starving, but
this sharing occurs only among individuals with
whom they are closely associated. This behavior is
referred to as reciprocity (or reciprocal altruism) and
occurs when the cost to the individual performing the
altruistic act is less than the benefit to the recipient
when such an act is later reciprocated (Crichton and
Krutzsch, 2000).
Costs
There are several potential costs associated with liv¬
ing in caves, most of which are related to living in
large groups. Large numbers of bats that live in close
physical contact with others may be more prone to
transmit certain diseases or increase the risk of para¬
sitic infestations. High mite infestations on a bat, for
example, may cause an increase in the amount of time
an individual spends grooming, and thus increase is
daily energy expenditure (Kunz, 1982). Large groups
of bats are more likely to transmit infectious pathogens
(Kunz and Fenton, 2003; Kunz and Parsons, 2009),
including fungal spores associated with white-nose
syndrome (Frick et al., 2010). Large concentrations of
bats in poorly ventilated chambers may also affect
the balance of certain gasses in a cave, requiring addi¬
tional physiological mechanisms to cope (Kunz, 1882;
Fleming and Racey, 2009).
Birds of prey, such as owls, hawks, and falcons, are
known to swoop down into columns of bats that
emerge nightly from caves and mines. Some predatory
birds, such as the bat hawk ( Machaerhamphus nlcinus)
in Africa and the bat falcon ( Falco rufigularis) in
Central and South America even specialize on bats.
Wintering merlins ( Falco columbarius) regularly congre¬
gate around large bat colonies in the West Indies,
influencing nightly departure patterns of bats during
the season when merlins are important predators
(Rodriguez-Duran and Lewis, 1985). Most predatory
birds are territorial, so their numbers at any one cave
are usually rather small.
Sometimes only a single mating pair of birds is
associated with one colony of bats, thus the impact on
local populations is probably minimal. Other animals
that prey on cave-roosting bats include snakes, rac¬
coons, skunks, opossums, and other bats; even frogs
have been observed preying on bats. Notwithstanding,
few studies have evaluated the impact that predators
have on bat colonies in caves.
The distribution of caves in most terrestrial landscapes
is highly variable (Palmer and Palmer 2009), and some
may be located at considerable distances from the food
resources on which bats depend. Bats that roost alone or
in small groups in tree cavities and in foliage can often
take advantage of food resources located near their
roosts, but cave bats, especially those that form large
aggregations, more often must commute considerable
distances to foraging sites. Because flight is energetically
expensive, bats must make compromises between colony
size and the amount of energy spent commuting to feed¬
ing sites and the energy that is conserved by selecting
roosts that have microclimates conducive to energy and
water conservation (Kunz, 1982).
Local food resources may not be sufficient to support
the energy and nutrient budgets of all individuals that
form large cave colonies; thus some individuals must
disperse considerable distances in order to secure their
daily energy and nutrient requirements. For example,
many maternity colonies of the Brazilian free-tailed bat
C Tadarida bmsiliensis) number in the millions, requiring
some individuals to fly upwards of 50 km each night to
obtain their food. This is the case for other species that
form large colonies in caves, such as the large bent¬
wing bat ( Miniopterus schreibersii), the lesser long-nosed
bat (Leptonycteris curasoae) (Kunz and Fenton, 2003), and
most members of the family Mormoopidae in the West
Indies (Fleming and Racey, 2009).
CONSERVATION AND MANAGEMENT
In recent years, reductions in the numbers of cave
bat populations have increasingly concerned conserva¬
tion biologists (Kunz and Racey, 1998; Kunz and
Fenton, 2003). One of the major problems that places
bat populations at risk is that they have relatively low
reproductive rates and thus are unable to recover
quickly from population declines (Kunz and Fenton,
2003). Cave bats face a variety of human threats that
may vary in different regions of the world. Some
threats reflect differences due to socio economic condi¬
tions, habitat types, and cultural attitudes toward bats.
Notwithstanding, several successful approaches in
recent years, such as habitat restoration and cave pro¬
tection, have been employed to protect bats, their
roosts, and their food resources. Increasingly, most if
not all geopolitical units (cities, states, countries) are
faced with issues related to increased surface and sub¬
surface mining operations, spelunking, ecotourism,
vandalism, sealing of caves and mines for safety rea¬
sons, and deliberate killing of bats. Other local threats
include guano mining and over-collection of bats for
scientific research, each of which can have adverse
affects on cave populations.
Bats are often portrayed by the media as being evil,
vile, or vectors of disease, but these negative attitudes
can best be overcome through increased educational
ENCYCLOPEDIA OF CAVES
BATS
51
efforts. Lack of basic information about the natural his¬
tory of most cave-roosting bats is a concern in many
regions of the world. Shared knowledge about the
ecology and behavior of different species is essential
for developing sound conservation strategies, and for
informing natural resource managers and decision
makers who will improve the well-being of bats and
the food and roost resources on which they depend
(Kunz and Fenton, 2003; Kunz et al., 2011).
ECOSYSTEM SERVICES PROVIDED BY
CAVE-ROOSTING BATS
Many cave-dwelling bats provide essential ecosys¬
tem services by helping to maintain forest diversity by
dispersing seeds and pollinating flowers (Kunz et al.,
2011). Changes in bat diversity or abundance due
to forest fragmentation or roost destruction can lead to
the dysfunction of forest ecosystems. In addition
to plant-visiting bats that disperse seeds and pollinate
flowers, many insectivorous bats that roost in caves
consume vast quantities of insects. Some insectivorous
species feed on insects that cause significant damage
to agricultural crops (Boyles et al., 2011). The Brazilian
free-tailed bat (Tadarida brasiliensis ), for example, is
known to feed on insects that cause millions of dollars
in damage to corn and cotton crops in the United
States each year. In addition, nearly everywhere that
large quantities of guano have accumulated in caves,
local communities have discovered its value as a fertil¬
izer. In some parts of the world, guano is mined
locally and sold commercially for fertilizer (Kunz and
Fenton, 2003; Kunz et al., 2011).
The organic input from bat guano (feces and urine)
is essential for sustaining the health of cave ecosystems
(Palmer and Palmer, 2009). Many cave-dependent
organisms (e.g., fungi, arthropods, fish, salamanders)
depend on bats to produce guano and thus provide
critical food resources in an environment where other
sources of organic nutrients are relatively scarce.
Typically, organic nutrients are provided to caves via
water flow. Subterranean streams wash coarse terres¬
trial detritus into caves, which is increased signifi¬
cantly following heavy rains, and groundwater drips
provide a source of dissolved organic compounds.
These processes can be rather infrequent or occur at a
relatively slow rate, whereas the seasonal nutrient deo-
sition with bat colonization can be very rapid.
The influx of organic nutrients from bat guano pro¬
vides a major source of carbon, nitrogen, and water,
especially into otherwise dry caves. In addition to car¬
bon and nitrogen, bat guano also contains important
minerals, including sodium, potassium, phosphorous,
iron, calcium, and magnesium, which are often limiting
nutrients in many ecosystems. The amount of energy
and inorganic nutrient that is provided by bat guano
depends on the energetic content of the food that bats
consume and how well the bats are able to assimilate
ingested food. For example, there are a number of
caves in south-central Texas that house large (400,000
to 1,000,000 individuals) maternity colonies of Brazilian
free-tailed bats. Within these caves, each lactating
female bat may produce 0.12 g dry mass of feces per
night, which contains 3.64 kj of free energy. Thus, a
colony of Brazilian free-tailed bats comprised of
1,000,000 individuals will produce enough feces to pro¬
vide a cave ecosystem with 3,640,000 kj of energy in a
given year (J.D. Reichard, unpublished data).
Bat guano provides the nutrient base on which com¬
plex food webs develop in caves (Palmer and Palmer,
2009). In particular, arthropod communities are the
most successful colonizers of bat guano deposits in
caves and form surprisingly rich communities com¬
prised of several trophic levels. Some arthropod
groups feed directly on bat guano or on the microor¬
ganisms that colonize the guano deposits. These
groups commonly include various species of mites,
coleopterans (beetles), diplopods (millipedes), bat ecto¬
parasite larvae and adults (fleas), collembolans (spring-
tails), isopods (woodlice), cockroaches, crickets, and
silverfish, among many others. These detritivores are
in turn fed on by various species of predators includ¬
ing chilopods (centipedes), hemipterans (true bugs),
arachnids, pseudoscorpions, carnivorous beetles and
mites, and vertebrates, including fish and salamanders,
some of which have become entirely dependent on the
cave fauna and flora derived from the organic input
from bats (Palmer and Palmer, 2009).
Most populations of bats in temperate regions do
not use caves for the entire year, using them for only a
few months while they are raising young or during
hibernation. For example, hibernating bats have little
impact on cave ecosystems because they only deposit
small amounts of guano; however, bats that form
maternity colonies incaves have considerable impact on
the cave ecosystems owing to high fecal deposition
associated with increased energetic demands of lacta¬
tion and parental care. The effects of seasonal bat colo¬
nies on cave ecosystems can be observed in organisms
that directly depend on guano for nutrients and energy.
For example, populations of dermestid beetles, which
feed on fallen bats and guano, closely track the pres¬
ence of bats. When bats are present, the beetles mostly
feed on bat carcasses. After the bats migrate from the
cave, dermestid beetles shift their diet to the accumu¬
lated guano; lack of new guano causes the population
of adult beetles to decline, leaving only eggs that will
hatch when the bats return in the spring. Cave-roosting
bats in tropical regions discard culled wings of insects.
ENCYCLOPEDIA OF CAVES
52
B
seeds, and bits of fruit and leaves dropped from feed¬
ing roosts, and thus provide a year-round supply of
energy and nutrients for a variety of cave organisms
(Kunz and Racey, 1998; Kunz et al., 2011).
Interestingly, organisms that are not typically asso¬
ciated with bat guano also depend on this nutrient-
rich organic source. For example, the larval stages of
the cave-adapted salamanders (Enrycea spelaea), which
inhabit caves in the Ozark Plateau that house large
maternity roosts of gray bats (Myotis grisescens), con¬
sume guano. Guano contains nearly the same amount
or more nutrients as the salamander's usual food (cave
amphipods). Thus, salamander populations increase
significantly in caves where bats are present for a
portion of the year (Palmer and Palmer, 2009).
Not only do bats provide nutrients to caves through
defecation, urination, and deposits of discarded food
items, they also influence the cave microclimate.
Metabolic heat is generated by bats and trapped in some
caves, although additional heat is generated from the
metabolic activities of microbes in guano deposits. This
heat can substantially increase the temperature of caves
beyond what can be expected from the influence of out¬
side ambient temperatures. Relative humidity, which is
typically high in most cave systems, can also be affected
by bats (Kunz and Fenton, 2003; Fleming and Racey,
2009). Water from feces and urine and exhaled air also
provide important sources of moisture that is especially
vital in dry caves.
With the appearance of white-nose syndrome (WNS)
in North America, cave ecosystems may face threats in
the near future as a result of widespread mortality
among cave-dwelling species (see below). If the mortal¬
ity of other hibernating species follows the effects
of WNS observed in the little brown myotis (Myotis luci-
fugus), then there could be major implications for cave
organisms that depend on the nutrients that bats import
into caves. Initially, the death of bats in a cave could
result in a nutrient boom from the accumulation of car¬
casses that decompose on the cave floor. However, such
a period of nutrient richness will be followed by a signif¬
icant decrease in nutrient input as fresh guano deposits
vanish with declining bat populations. While the cessa¬
tion or decline of guano deposition may not necessarily
lead to extinction of cave specialists, the loss of nutrient
input will most likely result in significant population
declines, loss of genetic diversity, and at worse the
collapse of some cave ecosystems.
Human Disturbance
Humans enter caves for various reasons, including
scientific research exploration, shelter, tourism, mining,
and even sometimes for collecting bats to eat. Whatever
the intentions might be, these activities can have adverse
consequences for bats (Kunz and Racey, 1998; Kunz and
Fenton, 2003). Disturbing bats during maternity periods,
whether they are handled or not, can cause pregnant
females to abort their young or cause young to fall to
the floor, leading to injury or death. Hibernating bats
are particularly vulnerable to disturbances from human
activities inside caves. Additionally, when humans dis¬
turb hibernating bats, they often respond by arousing,
which is energetically expensive. Nontactile stimuli,
such as light and noise, can also increase the activity of
bats that hibernate in caves and thus lead to unintended
arousals and the depletion of critical energy reserves
(Kunz and Fenton, 2003).
Habitat Destruction and Alteration
Surface and subterranean mining activities can have
adverse impacts on cave-roosting bats, because such
activities can modify the physical structure and microcli¬
mate of caves. Because many bats have very specific
roosting requirements, such changes may cause bats to
completely abandon these sites. Just as important is the
fact that mining operations frequently use toxic chemi¬
cals. For example, in some regions, modern gold-mining
methods use cyanide to extract gold from ore, and as a
consequence such practices have unexpectedly killed
enormous numbers of cave bats, by contaminating water
sources from which they derive their food and drinking
water. While increases in the number of abandoned
mines may have increased the abundance and distribu¬
tion of some cave-roosting species, reclamation of mines
and the closing of others have also led to an increase in
bat mortality when these structures are closed without
verifying their presence (Kunz and Fenton, 2003).
The habitat surrounding caves can be just as impor¬
tant for bats, as is the roosting environment within the
cave itself. Many hibernating species, such as the
endangered Indiana myotis (M. sodalis) typically roost
beneath exfoliating bark during the warm months but
hibernate in caves during winter. To survive a pro¬
longed period of hibernation, bats must be able replen¬
ish their fat reserves following autumn migration and
before entering hibernation, thus productive foraging
habitats located near hibernacula are essential for their
success. The vegetation around a cave not only sup¬
ports source populations of insect prey but may also
buffer the interior of the cave from severe changes in
wind flow and temperature.
Threats to Cave Bats from White-Nose Syndrome
Several pathogens and putative pathogens have been
linked to cave bats (Kunz and Parsons, 2009), some of
which are zoonotic (e.g., can be transmitted to humans).
Among the zoonotic viruses, rabies and Ebola are
ENCYCLOPEDIA OF CAVES
BATS
53
perhaps the best known to the general public, but there
is little to no evidence that these viruses have caused
mass mortality in bats. Recently, the emerging fungus
(Geomyces destructans, Gd) associated with white-nose
syndrome (WNS) has led to the death of over a million
hibernating bats in the northeastern United States, and
since its first discovery in New York State in 2006
(Blehert et al., 2009; Frick et al., 2010), has contributed to
one of the most precipitous declines of wildlife in
recorded history. Within four years of being reported,
this putative pathogen has spread from New York to 16
other states, south to North Carolina, west to Missouri
and Oklahoma, and northward to the provinces of New
Brunswick, Ontario, and Quebec in Canada. Based on an
analysis of reported mortality in the little brown myotis
(M. lucifugus), the species most affected by this condition,
regional extinction has been predicted in this species
within 20 years (Frick et al., 2010). A strain of Gd, the
same fungal pathogen that has been linked to WNS in
North America, has also been reported from hibernating
bats in European caves but, to date, there is no evidence
of WNS-related symptoms or mass mortality in these
species (Wibbelt et al., 2010).
Threats to Cave Bats from Wind Turbines
The proliferation of wind energy facilities through¬
out the world also places bats at risk of being killed
from being struck directly by turbines blades or indi¬
rectly from barotrauma (e.g., damage to the lungs on
being exposed to negative pressure associated with
moving turbine blades). To date, most bat fatalities
attributed to wind turbines in North America have
been tree-roosting species, but up to 24% of the bat
fatalities at some wind energy facilities have been cave-
roosting species (Kunz et al., 2007; Arnett et al., 2008),
including the big brown bat (Eptesicus fuscus) the little
brown myotis (M. lucifugus), the northern long-eared
myotis (M. septentrionalis) the tricolored bat ( Perimyotis
subflavus), and the Brazilian free-tailed bat ( Tadarida
brasiliensis). Several species of cave-roosting bats in
Europe are also being killed by wind turbines (Rydell
et al., 2010), a pattern that is likely to prevail throughout
the world where wind-energy facilities are being
installed and operated. Most importantly, wind-energy
facilities should not be installed near caves that serve as
maternity roosts, swarming sites, or hibernacula.
Fatalities to cave-roosting and tree-roosting bats
from wind turbines can be markedly reduced opera¬
tionally by increasing the rotor cut in speed of turbines
to 6.5 m s _1 (Arnett et al., 2010). The rationale for this
mitigation effort at sites where wind turbines have
been inappropriately located is that aerial insects, and
hence aerial feeding insectivorous bats, are most active
at low wind speeds, typically less than 6 m s _ .
Additionally, if wind turbines blades can be operation¬
ally feathered at certain time of year (e.g., especially
during fall migration), fatalities can be reduced signifi¬
cantly (Cryan and Barclay, 2009).
The Paradox of Vampire Bats
An endless source of myths and legends, vampire
bats offer a valuable lesson about the need to learn
more about these and other bat species before it is too
late to protect them from extinction. Three species of
vampire bats, each of which use caves as roosts, range
from northern Mexico through South America, but
only the common vampire bat (Desmodus rotundus) is
sufficiently abundant to be considered a nuisance to
humans and their livestock. All three species depend
on a diet of blood, but they feed on a variety of differ¬
ent animals. Most of our knowledge of vampire bats
comes from the common vampire bat, which specia¬
lizes on mammalian blood as its primary source of
food. Populations of vampire bats increased sharply
throughout Latin America following the introduction
of livestock by European settlers over 500 years ago.
Because the common vampire bats feed on cattle and
occasionally on humans, it is considered a pest in most
parts of Central and South America. The economic loss
due to cattle dying from bat-transmitted rabies alone is
a major concern in many regions of Latin America,
and thus massive eradication efforts are often imple¬
mented, leading to the death of other species as well.
Lack of education and misguided attempts to con¬
trol vampire bat populations have led to the mass
destruction of nontargeted species. Nonselective killing
techniques, such as fire and gas (fumigating caves),
have been used either because local landowners are
often unaware of the differences between vampire bats
and other species or because they are uninformed
about the ecological value of bats in general. Poisons,
such as strychnine, or anticoagulants are often applied
to the wounds on livestock because vampire bats
return to wounds that they made the previous night.
Selective approaches that concentrate on controlling
vampire bats should be used whenever possible.
Recent discoveries by researchers indicate that chemi¬
cals present in the saliva of common vampire bats
have important medical benefits (e.g., reducing the
risks of stroke and heart attacks in humans). Thus, a
bat species that is considered a nuisance or public
health threat by some segments of society may offer
enormous benefits to others.
Many local, national, and international organizations
have become engaged in efforts to support research on
bats and have helped to educate the public about the
ENCYCLOPEDIA OF CAVES
54
B
benefits of these flying mammals to humankind. Many
caving organizations have joined this effort to protect bats
in general, and cave-roosting bats in particular. Television
programs, newspaper articles, the Internet, and other
media must be used to promote the ecological value of
bats and the importance of caves for sustaining bat popu¬
lations on a worldwide scale. Organizations such as
Bat Conservation International (www.batcon.org). Bat
Conservation Trust (www.bats.org.uk), the Lubee Bat
Conservancy (www.lubee.com), the Organization for
Bat Conservation (www.batroost.com), and the National
Speleological Society (www.caves.org) are among a grow¬
ing number of nongovernment organizations that are con¬
tributing to these efforts. Notwithstanding, additional
efforts are needed to help promote and protect the 1232
species of living bats known worldwide. Given the threats
to cave-roosting bats from anthropogenic disturbances,
emerging diseases, misinformation, and human
ignorance, management practices that conserve cave-
roosting species are needed to ensure the survival of the
rare and unique cave ecosystems that these species help
to create and maintain.
Acknowledgments
We wish to acknowledge support from the American Society of
Mammalogists (S.W.M.), Bat Conservation International (S.W.M. and
N.W.F.), National Geographic Society (T.H.K.), National Science
Foundation (T.H.K.), National Speleological Society (T.H.K. and N.W.
F.), Lubee Bat Conservancy (T.H.K.), and U.S. Fish and Wildlife
Service (T.H.K.) for funding our research on bats. We also thank Tigga
Kingston, Wendy R. Hood, and Armando Rodriguez-Duran for mak¬
ing helpful suggestions on earlier drafts of this manuscript.
See Also the Following Article
White-Nose Syndrome
Bibliography
Arnett, E. B., Brown, W. K., Erickson, W. P., Fiedler, J. K., Hamilton,
B. L., Henry, T. H., et al. (2008). Patterns of bat fatalities at wind
energy facilities in North America. Journal of Wildlife Management,
72(1), 61-78.
Arnett, E. B., Huso, M. M. P., Schirmacker, M. R., & Hayes, J. P. (2010).
Altering turbine speed reduces bat mortality at wind-energy facili¬
ties. Frontiers of Ecology and the Environment, 9, 209—214.
doi:10.1890/100103.
Blehert, D. S., Hicks, A. C., Behr, M., Meyer, C. U., Berlowski-Zier,
B. M., Buckles, E. L., et al. (2009). Bat white-nose syndrome: An
emerging fungal pathogen? Science, 323(5911), 227.
Boyles, J. G., Cryan, P. M., McCracken, G. F., & Kunz, T. H. (2011).
Economic importance of bats in agriculture. Science, 332(6025),
41-42.
Crichton, E. G., & Krutzsch, P. H. (Eds.), (2000). Biology of bat repro¬
duction. San Diego: Academic Press.
Cryan, P. M., & Barclay, R. M. R. (2009). Causes of bat fatalities at
wind turbines: Hypotheses and predictions. Journal of Mammalogy,
90(6), 1330-1340.
Fleming, T. H., & Racey, P. A. (Eds.), (2009). Island bats: Evolution,
ecology, and conservation. Chicago: University of Chicago Press.
Frick, W. F., Pollock, J. F., Hicks, A., Langwig, K., Reynolds, D. S.,
Turner, G., et al. (2010). A once common bat faces rapid extinc¬
tion in the northeastern United States from a fungal pathogen.
Science, 329(5992), 679-682.
Kunz, T. H. (Ed.) (1982). Ecology of bats. New York: Plenum Press.
Kunz, T. H., & Fenton, M. B. (Eds.), (2003). Bat ecology. Chicago:
University of Chicago Press.
Kunz, T. H., & Parsons, S. (Eds.), (2009). Ecological and behavioral
methods for the study of bats. Baltimore: Johns Hopkins University
Press.
Kunz, T. H., & Racey, P. A. (Eds.), (1998). Bat biology and conservation.
Washington, DC: Smithsonian Institution Press.
Kunz, T. H., Arnett, E. B., Erickson, W. P., Johnson, G. D., Larkin,
R. P., Strickland, M. D., et al. (2007). Ecological impacts of wind
energy development on bats: Questions, hypotheses, and research
needs. Frontiers of Ecology and the Environment, 5, 315—324.
Kunz, T. H., Braun de Torrez, E., Bauer, D. M., Lobova, T. A., &
Fleming, T. H. (2011). Ecosystem services provided by bats.
Annals of the Nezv York Academy of Sciences, 1223, 1—38.
Munoz-Romo, M., & Kunz, T. H. (2009). Dorsal patch and chemical
signaling in males of the long-nosed bat, Leptonycteris curasoae
(Chiroptera: Phyllostomidae). Journal of Mammalogy, 90(5),
1139-1147.
Palmer, A., & Palmer, M. (Eds.), (2009). Caves and karst of the USA.
Huntsville, AL: National Speleological Society.
Rodriguez-Duran, A., & Lewis, A. R. (1985). Seasonal predation by
Merlins on sooty mustached bats in western Puerto Rico.
Biotropica, 17, 71—74.
Rydell, J., Bach, L., Dubourge-Savage, M.-J., Green, M., Rodrigues, L.,
& Hedenstrom, A. (2010). Mortality of bats at wind turbines links
to nocturnal insect migration. European Journal of Wildlife Research,
56, 823-827.
Wibbelt, G., Kurth, A., Hellmann, D., Weishaar, M., Barlow, A.,
Veith, M., et al. (2010). White-nose syndrome fungus (Geomyces
destructans ) in bats, Europe. Emerging Infectious Diseases, 16,
1238-1242.
Zubaid, A., McCracken, G. F., & Kunz, T. H. (Eds.), (2006). Functional
and evolutionary ecology of bats. New York: Oxford University
Press.
BEETLES
O ana Teodora Moldovan
Emil Racovitza Institute of Speleology, Cluj Department,
Romania
INTRODUCTION
As part of the Insect class, representatives of the order of
Coleoptera usually have a sclerotized body with sclero-
tized forewings that are leathery or horny and modified
to act as rigid covers ( elytra ) over the membranous,
reduced, or even absent hindwings. The mouthparts are
adapted for cutting, nibbling, and chewing, and the
antennae have usually 8 to 11 articles. The male genitalia
are retractable, and the females do not possess an
Encyclopedia of Caves.
© 2012 Elsevier Inc. All rights reserved.
BEETLES
55
ovipositor. Beetles are the most numerous of all insects,
with more than 400,000 described species—almost 40%
of all known animals (Hammond, 1992). Big or small,
they are everywhere, occurring in all environments,
including caves, lava tubes, cracks and fissures in a mas¬
sif in mesovoid shallow substratum (MSS) in limestone,
or in other different rocks such as schist, gneiss, grano-
diorites, basalts, quartzites, grits, and so on.
Of the 166 families of the order, only very few
have species in the underground world: the aquatic
Dryopidae, Dytiscidae (predaceous diving beetles),
Elmidae, Hydrophilidae (water scavenger beetles), and
Noteridae, as well as the terrestrial Carabidae (ground
beetles), Curculionidae (weevils), Histeridae, Leiodidae,
Pselaphidae, Staphylinidae (rove beetles), Endomychidae,
Ptiliidae, Scydmaenidae, and Tenebrionidae (darkling
beetles) (Fig. 1). Even underground they are the best
represented of all animals, with more than 2000 species.
It is not surprising that one of the first discovered
cave animals was a beetle, in Postojnska Jama (Slovenia).
In 1831, Cec observed specimens that looked very much
like ants on the beautiful stalagmites. Count Franz von
Hohenwart sent the material to the Austrian naturalist
Schmidt, who described Leptodirns hochenwartii, an
amazing terrestrial species displaying a high degree of
adaptation to cave life (Fig. 2).
ADAPTATIONS
Several adaptations of cave beetles to darkness, low
and heterogeneous in time and space food input (at
least in temperate regions, where the food is brought
inside caves through the cracks network, only in some
seasons, and mostly by water), and a relatively constant
climate are characteristic of cave life, but the degree of
morphological, anatomical, behavioral, and physiologi¬
cal adaptation is not similar in all species. One of the
FIGURE 1 Proportion of cave beetle families in the group of
Coleoptera. Source: Adapted from Juberthie and Decu (1998).
FIGURE 2 Leptodirus hochenwartii (Schmidt, 1832). Photo courtesy
of Valika Kustor, Slovenia. Used with permission.
first morphological changes to occur among beetles
during colonization of caves is a loss of pigmentation.
The cuticle becomes thinner, and the color of the cave
beetles given by the benzoquinones is a red brown. The
most evident morphological change is the reduction or
complete lack of eyes. The ocelles and rhabdomeres can
disappear completely, and reduction can also affect the
optical center, much more pronounced than for stygobi-
tic fishes and decapod crustaceans.
Elongation of the body and antennae, which become
slender and longer, is linked to their preferred habitat
(the network of cracks) and serves as compensation for
the lack of eyes. Longer antennae mean also longer
mechanical, gustatory, olfactory, and hygrosensitive
sensilla, as well as enhancement of the receptory sur¬
face. Cave beetles have a larger reception surface com¬
pared with their epigean relatives.
Life in caves makes the use of the wings impossible.
Wings are completely lost in some species, and the ely¬
tra are fused together. Moreover, in highly specialized
species, under the elytrae is located a compartment
containing air for humidity regulation that causes a
false physiogastry, similar to the bulging abdomen of
ants and termites filled with lipid reserves.
The internal anatomical modifications are especially
due to the scarcity of food and its uneven distribution
throughout the year; therefore, adapted species have
developed a fat body containing huge vesicles filled
with fat, proteins, and glycogen that allow survival dur¬
ing several months of fast. There is also a modification of
the exocrine glands, observed on some Leiodidae. The
ENCYCLOPEDIA OF CAVES
56
B
soil species have a big sternal exocrine gland (secreting
pheromones) that disappears completely in cave rela¬
tives, being replaced by smaller unicellular glands. The
secreted substances switch from a mixture of very vola¬
tile and less volatile ones for soil species to less volatile
ones, perceived only at small distances (the special artic¬
ular hydrocarbons), in subterranean species. In caves,
the presence of food can be an attractant at long dis¬
tances, while the small-distance pheromones act in place,
releasing the mating behavior and even the laying of
eggs. This behavior saves the energy needed for the pro¬
duction of offspring. Even the larvae save energy by not
feeding during the development, as they are protected
by small "houses" of clay.
Several breeding experiments were done in the
Pyrenean Cave laboratory of Moulis (France) (Deleurance-
Glafon, 1963) on beetles with different degrees of adapta¬
tion. The more adapted have low fecundity with a
reduction in the number of ovarioles. The females lay few
eggs or only one that is bigger, with more vitellinic
reserves; also, the time required for egg hatching is longer.
The larval stages and larval life are reduced, and the time
spent as a pupa increases (Fig. 3). The life cycle varies;
French species spend 4 to 5 years as adults, and the
American Ptomaphagus species spend only 2 years (Peck,
1986).
The activity of epigean species depends very much
on the day/night and seasonal alternations. For cave
beetles in the absence of light and other 24-hour
Eggs
FIGURE 3 Development cycles of underground beetles with dif¬
ferent levels of adaptations: 1, endogean; 2, less adapted; 3, very
adapted.
environmental cues, periods of activity and rest do not
have a daily rhythm.
Tropical caves with important food input are inhab¬
ited by less adapted beetles, especially those caves
with large deposits of guano. The process of adapta¬
tion of guanophilic and guanobiotic species is slowed
down. In these regions, at low altitude the species are
troglomorphic, and true troglobitic beetles typically
inhabit high mountain caves (Sbordoni et al., 1977).
SYSTEMATICS OF CAVE BEETLES
The family Carabidae includes the Trechinae, repre¬
senting more than 2000 subterranean species. These
have colonized all the subterranean habitats, from soil
to caves. The cavernicolous species have been classi¬
fied into two morphological types (anophthalm and
aphenopsian), corresponding to how advanced they
are in the adaptive evolutionary processes (Fig. 4). The
anophthalm type characterizes the endogean and some
cave species: depigmented, reduced or no eyes, ante¬
rior body and appendages slightly elongated. The
other type, aphenopsian, is a very evolved species
that is eyeless with a very long body, antennae,
and legs, and very thin cuticle. Typical for this last
group are Aphaenops in Europe and Neaphaenops and
Mexaphaenops in North America. But one of the most
amazing adaptations concerns the sensory equipment,
with development of olfaction through increased
numbers of antennae receptors and lengthening of
the mechanical trichobotries on the ellytrae. Some
Trechinae are polyphagous predators, such as the
French Aphaenops, which has a diet consisting of adults
and larvae of other beetles, springtails, flies, sometimes
(A)
FIGURE 4 Different Trechinae adaptations to caves: (A)
anophthalm and (B) aphenopsian. Source: Adapted from Ginet and
Decou (1977).
ENCYCLOPEDIA OF CAVES
BEETLES
57
millipedes, crickets, diplurans, and so on, and which
hunts in caves and MSS (Juberthie and Decu, 1998).
Others are very specialized on a single prey, such as
Rhadine subterranea from Great Onyx Cave (United
States), which feeds exclusively on the eggs and young
of the cricket Hadenoecus subterranea. The anatomy and
behavior of this beetle are surprisingly adapted for
locating and stealing the eggs deposited one at a time
beneath the surface of silt. Chemical substances left by
the cricket release the search behavior of the beetle, and
when the prey is found a hole is dug and the contents
of the eggs are devoured (Mohr and Poulson, 1966).
Some species can be rare in caves, only sometimes
being found on the sandy banks of subterranean rivers
or during flooding seasons. The Trechinae are very
mobile and active in the search for food, probably cov¬
ering large areas in fissured massifs. In this group, the
steps of evolution and the way in which adaptation has
occurred can be observed very well thanks to the pres¬
ence of many species outside, in moist or dark habitats
(such as under rocks or moss) and deep in the soil. The
troglomorphic features of the cave Carabidae are very
diversified; some of them are explained by adaptation
to different underground compartments, but others rep¬
resent the original contribution of some phyletical line
or the years spent as cave inhabitants (Fig. 5).
Another rich family is the Leiodidae, especially the
group of Leptodirini, with almost 200 genera and more
than 800 species. Leptodirini occur only in European
caves, being replaced in North America by smaller
but very interesting other subfamilies, such as the
Ptomaphagini. The Leiodidae have colonized all the sub¬
terranean habitats, but they consume organic or decom¬
posed matter and are not predators. Cave clay or
moonmilk that contains bacteria, fungi, and algae is
important to their diet, and these beetles can sometimes
FIGURE 5 Head, antennae, and first pair of legs of (A) Clivina
subterranea and (B) Italodytes stammeri showing the differences,
respectively, between new colonizers and old colonizers of caves.
be found in huge numbers on these substrates. Species
of this group have different morphologies according to
their degree of adaptation to cave life; four morphologi¬
cal types are accepted: (1) bathyscioid, for humicolous,
endogenous, and some less specialized cave species with
more or less globular forms of body and short appen¬
dages; (2) pholeuonid, characterizing specialized forms
with longer, slender bodies and appendages and false
physiogastry; (3) leptodiroid, for highly specialized spe¬
cies, such as Leptodirus, with extremely elongated legs
and antennae and a very small anterior part of the body;
(4) scaphoid (from scaphe, Greek for "boat"), which are
also highly specialized and very similar to the previous
group but with a different form of the body like a boat
(Fig. 6). Not enough data are available regarding preda¬
tors of these beetles. It is generally accepted that they do
have as many as their epigean relatives, but some cases
of predation from harvestmen and pseudoscorpions
have been reported and it is probably more pronounced
on the eggs and larvae than on the adults. Other
Coleoptera families have few troglobitic representatives
but do have interesting adaptations to subterranean life.
It was considered that the aquatic beetles were not as
successful in colonizing hypogean habitats as the terres¬
trial ones. The first stygobiotic beetle was mentioned in
France at the beginning of the twentieth century. Until
recently, 31 species were known to inhabit cave
streams, springs, or wells (Fig. 7A), usually in warmer
climates. Today, their number has increased. In
Australia alone are now more than 100 known species
of diving beetles (Dytiscidae) inhabiting the calcretes of
the arid parts of Australia (Watts and Humphreys,
2006). Besides the typical adaptations of cave beetles,
aquatic species have other adaptations, such as switch¬
ing between swimmers and crawlers, different methods
of obtaining air through cuticular respiration or trache-
ated elytral respiration, pupation at the bottom of sub¬
terranean waters, and smaller size (from 1.1 to 4.5 mm).
The Curculionidae is the best represented family of bee¬
tles on the surface, but there are few cave inhabitants, with
no typical cave adaptation. These species display different
degrees of specialization to a deep soil environment, such
as the eyeless Troglorhynchus monteleonii from a cave in
Central Italy (Osella, 1982). It is not a question of adapta¬
tion for cave life but rather one of a different degree of spe¬
cialization to a deep soil environment (Juberthie and
Decou, 1998). The cave in this case is more a trap.
Troglophilic and guanobiotic specialization (many
myrmecophilous and termitophilous species) of the
Histeridae can explain the low number of cave species;
the first cave specimen was discovered in Turkey
(Fig. 7B). More than 150 taxa of Pselaphidae are subterra¬
nean inhabitants (Fig. 7C). The origin of the temperate
troglobitic pselaphids can be traced to the Tertiary,
under similar conditions as of the humid and relatively
ENCYCLOPEDIA OF CAVES
58
B
FIGURE 6 Different types of Leptodirini as adaptations to caves: (A) bathyscioid; (B) pholeuonid; (C) leptodiroid; (D) scaphoid.
Source: Adapted from Ginet and Decou (1977).
FIGURE 7 Cave beetles: (A) aquatic Morimotoa gigantea (Ueno,
1957) (from Japan); (B) histerid Spelaeabraeus agazzli (Moro, 1957)
(from Italy); (C) pselafid Decumarellus sarbui (Poggi, 1994) (from
Romania); (D) staphylinid Domene vulcanica (Oromi and Hernandez,
1986) (from Canary Islands).
cold forests of the African mountains where forms with
small eyes and no wings are largely spread in the
humus.
Most Staphylinidae are from the Mediterranean region
(Morocco, Algeria, Spain, Italy) or nearby (Romania and
the Canary and Madeira Islands), but generally they are
troglophilic (Fig. 7D).
Other families of cave beetles are associated with
guano deposits and therefore do not share the same
morphological adaptations as the cave species.
GEOGRAPHICAL DISTRIBUTION
AND COLONIZATION
The subterranean beetles spread on all continents and
on some islands, not only in karstic areas (Fig. 8). The
richest regions are the temperate ones. The glaciated
areas usually do not have cave beetles, but some troglo-
bitic species have recolonized the underground habitats.
There are parts of the world where the beetle fauna is
very diversified, such as the Mediterranean karst. The
reasons for the separation of different populations with
wide ranges of distribution can be paleogeographical,
paleoclimatical, or ecological. European and North
American species are the best known, given the number
of specialists and long biospeleological tradition, but
also the favorable climatic conditions. The extreme north
and south areas were repeatedly covered with ice, and
the limits of the glaciers are very well reflected today in
the limits of the troglobitic beetle distribution.
The beetles that have colonized the underground
world were preadapted or exapted; they were terrestrial
and nocturnal, preferring moist habitats, frequently the
ENCYCLOPEDIA OF CAVES
BEETLES
59
FIGURE 8 World distribution of cave beetles (in gray). Source: Adapted from Juberthie and Decu (1998).
fissures of soil, bark, or entrances to caves. The moment
of cave occupation is still debated, but it is often related
with the Quaternary glacial—interglacial periods (the
climate relict hypothesis) or with the natural processes
of empty environment colonization (the habitat shift
hypothesis). One of these hypotheses can be verified by
using morphological features, characteristics of the male
genitalia, or molecular phylogenies.
Peck (1981) proposed the following scenario for colo¬
nization of caves in the Grand Canyon (between 1160-
and 1580-m altitude) by Ptonwphagus hirtus. The first
species arose during the interglaciary period after the
Illinois glaciation (235,000 to 185,000 BP) and dispersed
underground during the next glaciation. Then, again,
new cave species appeared in the next interglaciary
period (150,000 to 90,000 BP). Colonization with troglo-
philic or troglobitic species was followed by isolation
during the interglaciary periods which separated the
populations, and new cave species appeared. This
hypothesis has been tested and validated with alloen-
zyme paleodating and paleotemperature measurements.
More recently, molecular analyses on west Mediter¬
ranean cave beetles brought interesting explanations
about the origin and the speciation of cave inhabitants
(Ribera et al., 2010). The origin of Leptodirini (Leiodidae)
in this area is represented by perfectly adapted ancestors
to the subterranean domain. By combining obtained
phylogenies with a calibrated molecular clock, ages of
the major events in the phyletical line history, the origin
of species, and species diversification can be obtained
(Fig. 9). For the Mediterranean region one important
geological event was used to calibrate the molecular
clock together with estimation of the mutation rate for
the studied genes. Detachment of the Catalonian and
Sardinian continental plates was the determinant
of the subterranean species vicariance. Phylogeny and
molecular clock on subterranean Trechini from the
above-mentioned geographical area have shown that the
origin of the subterranean clade is placed at 10 Ma BP
and separation events occurred in the same geographi¬
cal area on species of the same morphological type
exclusively within the Pliocene and Pleistocene (Faille
et al., 2010). These two examples sustain the habitat shift
hypothesis.
Nevertheless, the opposing hypothesis (the climatic
relict) is sustained by phylogenies of diving beetles
from aquifers beneath the arid parts of Australia.
Lineages populated groundwaters during the late
Miocene or Early Pliocene and were caused by drastic
climatic changes. Humid habitats were replaced with
dry habitats, as can be seen today in most parts of this
continent (Leys et al., 2003).
ECOLOGY
Cave beetles generally live in relatively stable climates
with constant temperature, an atmosphere saturated with
ENCYCLOPEDIA OF CAVES
CENTRAL & EASTERN PYRENEES
Speonomus group
LBath^sciola ovata
Ovoboth^sciola^roup
TBathysciola zon’quieyi
Speonomiaius group
CENTRAL & WESTERN PYRENEES
CANTABRIAN MOUNTAINS
CENTRAL MEDITERRANEAN COAST
PYRENEES
SARDINIA
NORTHERN MEDITERRANEAN COAST
CANTABRIAN MOUNTAINS
Eocene Oligocene Miocene Pliocene Pleistocene
44 34 23
»
5.3 2.6
FIGURE 9 Phylogenetic tree of Leptodirini from the Western Mediterranean obtained from combined mitochondrial and nuclear
sequences. Source: Modified from Juan and Emerson (2010).
water vapor, and no air currents. There is a link between
the presence of beetles in caves and food input in spring
and autumn (for temperate regions). It has been observed
that populations migrate between the network of cracks
and the caves, depending on the presence of food and on
climatic parameters. The fissures and cracks offer a more
stable habitat than the big passages and rooms that have
large volumes of air and are more or less in direct contact
with the natural entrances. Very sensitive to any change
in the conditions of their environment, cave beetles
quickly initiate a behavior, even if it is only to run.
Interesting examples are provided by studies from
Transylvania (Romania), France, and the United States.
Variations of the number of individuals at different
times during a year were determined to be influenced
by air temperature or the level of the subterranean
stream. The example in Figure 10 relates the increase
of water level to the withdrawal of the local population
of Pholeuon (Racovifa, 1971).
Determining the size of cave beetle populations has
been a concern of coleopterologists. Individuals cap¬
tured at a given time have been marked and then recap¬
tured after a period of time. Estimates can then be
made by comparing the number of recaptured marked
individuals with the total number of all marked ones.
For example, the Neaphaenops tellkampfi in Mammoth
Cave has been estimated to be 750,000 individuals (Barr
and Kuehne, 1971). In France, the population of catopid
Speonomus in an MSS station was estimated at 1,000,000
individuals for two species, and the population of tre-
chine Aphaenops was estimated to be 100,000 individuals
(Juberthie and Decu, 1998).
The presence of two or more beetle species in the
same cave usually means that one of them is dominant,
especially if they compete for the same food resources.
The species can choose different places in the under¬
ground system and the predators' different niches. In
the Transylvanian caves, two genera of Leptodirinae
coexist: Pholeuon and Drimeotus. With very few excep¬
tions, Drimeotus is limited primarily near the entrances
and in small numbers, or in the network of cracks,
while the Pholeuon are numerous in the deep cave
passages. In Mammoth Cave, beside the numerous
Neaphaenops tellkampfi, there are other two dominant tre-
chine with different distributions: Pseudoanophthalmus
menetriesi, which is found in the upper level, and
P. striatus, which is found along the subterranean river.
Mammoth Cave has one of the largest and most com¬
plex subterranean biological communities of all known
caves, with no less than six trechine species, one psela-
phid, and one leiodid. In White Cave, the animal's
entire life is dependent on crickets and to a lesser extent
ENCYCLOPEDIA OF CAVES
BEETLES
61
FIGURE 10 Numerical variation of Pholeuon moczaryi in Vadu Crigului Cave (Romania), in the period 1968 to 1970 as a consequence of
flooding periods (indicated by arrows). Source: Adapted from Racovi\d (1971).
EPISEAN ARTHROPODS
FIGURE 11 Food web in White Cave (Kentucky). Source: Adapted
from Barr and Kuehne (1971).
on a few pieces of rotten wood (Barr and Kuehne,
1971); a generalized food web for this community is
shown in Figure 11. On the partially decomposed
cricket guano live saprophytic snails, isopods, milli¬
pedes, springtails, diplurans, and leiodid beetles. These
are eaten by spiders, harvestmen, pseudoscorpions, and
trechine beetles (it has not been determined whether or
not the predators are highly specialized).
IMPORTANCE AND PROTECTION
The diversity of species makes this group of cave
invertebrates one of the best for testing many of the
hypotheses concerning adaptation strategies during
colonization of empty places on Earth. Cave beetles are
also very precious to the natural biodiversity of the
world. One species can populate one cave or one mas¬
sif, so the degree of endemism is very high. On the
other hand, in many regions the biospeleological
explorations are only just beginning, and the task of
the coleopterologist is to find and describe newly dis¬
covered species. The importance of their studies lies
primarily in the development of knowledge regarding
conservation management measures to be taken for
more vulnerable or rare species. Experiments carried
out in a Romanian show cave indicate that the pres¬
ence of tourists has eliminated the troglobiotic beetles
from visited parts of the cave and has also influenced
the yearly dynamics of the leptodirine population
(Fig. 12) in the nonvisited part.
Rare and having a strange morphology, cavernico-
lous species have attracted the attention of collectors,
especially in Europe; however, laws to prevent over¬
collection of specimens from caves seem to have had
little effect on this trade.
ENCYCLOPEDIA OF CAVES
62
40
35
30
25
20
15
10
5
0
Z
o
C7
t
3'
re
o
-f
c
l/l
FIGURE 12 Variation in the number of
Drimeotus in Urgilor Cave (Romania) as a function
of tourist periods. Source: Modified after Moldovan
et al. (2003).
Months (1997-1998)
Bibliography
Barr, T. C., & Kuehne, R. A. (1971). Ecological studies in the mam¬
moth cave system of Kentucky. II. The ecosystem. Annales de
Speologie, 26(1), 47—96.
Deleurance-Glafon, S. (1963). Recherches sur les coleopteres troglo-
bies de la sous-famille des Bathysciinae. Annales des Sciences
Naturelles-Zoologie, 1(5), 1—172 (in French).
Faille, A., Ribera, L, Deharveng, L., Bourdeau, C., Garnery, L.,
Queinnec, E., & Deuve, T. (2010). A molecular phylogeny shows
the single origin of the Pyrenean subterranean Trechini ground
beetles (Coleoptera: Carabidae). Molecular Phylogeny and
Evolution, 54(1), 97—106.
Ginet, R., & Decou, V. (1977). Initiation a la biologie et a le'cologie souter-
raines. Paris: Delarge.
Hammond, P. M. (1992). Species inventory. In B. Groombridge (Ed.),
Global biodiversity. Status of the Earth's living resources, (pp. 19—36).
London: Chapman & Hall.
Juan, C., & Emerson, B. C. (2010). Evolution underground: Shedding
light on the diversification of subterranean insects. Journal of
Biology, 9(3), 17.
Juberthie, C., & Decu, V. (Eds.), (1998). Encyclopaedia biospeologica
(Vol. II). Paris: Societe de Biospeologie.
Leys, R., Watts, C. H. S., Cooper, S. J. B., & Humphreys, W. F. (2003).
Evolution of subterranean diving beetles (Coleoptera: Dytiscidae:
Hydroporini, Bidessini) in the arid zone of Australia. Evolution,
57(12), 2819-2834.
Mohr, C. E., & Poulson, T. L. (1966). The life of the cave. New York:
McGraw-Hill.
Moldovan, O. T., Racovitza, G., & Rajka, G. (2003). The impact of
tourism in Romanian show caves: The example of the beetle
populations in the Ursilor cave of Chiscau (Transylvania,
Romania). Subterranean Biology, 1(56), 73— 78.
Osella, G. (1982). I curculionidi cavernicoli italiani (riassunto). Lavori
Societa Italiana Biogeographia, 7, 337—338 (in Italian).
Peck, S. B. (1981). Evolution of the cave Cholevinae in North
America (Coleoptera: Leiodidae). In Proceedings 8th international
congress of speleology (Vol. 2, pp. 503—505). Kentucky: Bowling
Green.
Peck, S. B. (1986). Evolution of adult morphology and life-history
characters in cavernicolous Ptomaphagus beetles. Evolution, 40(5),
1021-1030.
Racovija, G. (1971). La variation numerique de la population de
Pholeuon (Parapholeuon) moczaryi Cs. De la grotte de Vadu-
Cri§ului. Travaux de I'Institut de Spe'ologie "E. Racovitza," 10,
273-278.
Ribera, I., Fresneda, J., Bucur, R., Izquierdo, A., Vogler, A. P.,
Salgado, J. M., & Cieslak, A. (2010). Ancient origin of a Western
Mediterranean radiation of subterranean beetles. BMC
Evolutionary Biology, 10(1), 29.
Sbordoni, V., Argano, R., Vomero, V., & Zullini, V. (1977). Ricerche
sulla fauna cavernicola del Chiapas (Messico) e delle regioni lim-
itrofe: Grotte esplorate nel 1973 e nel 1975. Criteri per una classifi-
cazione biospeleologica delle grotte. In Subterranean fauna of
Mexico (Part III). Quaderni Accademia Nazionale Lincei, 71 (pp.
5-74).
Watts, C. H. S., & Humphreys, W. F. (2006). Twenty six new
Dytiscidae (Coleoptera) of the genera Limbodessus Guignot and
Nirripirti Watts and Humphreys, from underground waters in
Australia. Transactions of the Royal Society of South Australia, 130(1),
123-185.
BEHAVIORAL ADAPTATIONS
Jakob Parzefall
Biozentrum Grindel, Zoologisches Institut, und Zoologisches
Museum der Universitdt Hamburg
Animals living in darkness have to compete for food,
mates, and space for undisturbed reproduction just as
their epigean conspecifics do in the epigean habitats.
Encyclopedia of Caves.
© 2012 Elsevier Inc. All rights reserved.
BEHAVIORAL ADAPTATIONS
63
but there is one striking difference: in light, animals can
use visual signals. Thus, important aspects of behavior
driven by visual signals cannot apply in darkness. The
question arises, then, of how cave dwellers compensate
for this disadvantage in complete darkness. This article
uses several examples to compare various behavior pat¬
terns among cave-dwelling populations with epigean
ancestors.
COMPARISON OF BEHAVIOR PATTERNS
IN CAVE-DWELLING ANIMALS AND
THEIR EPIGEAN RELATIVES
Potential cave dwellers must have the sense organs
and behavior necessary to find food and to reproduce
in caves. Such animals may be said to be preadapted to
cave life, and in fact some of these animals can survive
in the darkness without behavioral adaptations and
can reduce behavioral characters not necessary in the
caves. In contrast, some cave dwellers have improved
sense organs and have acquired behavior adapted to
their extreme habitats.
Food and Feeding Behavior
Suitable food sources and quantity vary from cave to
cave. In general, cave animals depend upon food
brought in from outside and are omnivorous. With
some exceptions, most caves do not have an abundance
of food compared to above ground habitats. Food
sources can be widely distributed or concentrated in
patches, and their occurrence is mostly unpredictable;
therefore, food-finding abilities have to be improved,
and food must be stored to ensure survival during long
starvation periods.
When the blind fish population of Astyanax fasciatus
from Pachon Cave, Mexico, was studied in competition
experiments conducted in darkness, it was found that
they retrieved 80% of small pieces of meat distributed on
the bottom of an aquarium, whereas the epigean fish got
only 20% (Hiippop, 1987). Among North American
amblyopsid fish, which comprise six species in four gen¬
era, the ability to detect invertebrates at low prey densities
in the dark is much better for the cave-living species
Amblyopsis spelea than the troglophile Chologaster agassizi.
When one Daphnia was introduced into a 100-L tank, the
A. spelea found the prey hours before the C. agassizi did. In
addition, the maximal prey detection distance is greater
in cave species. Daphnia was detected by Typhlichtys sub¬
terraneous within 30 to 40 mm and by C. agassizi within
10 mm (Poulson, 1963). Density-dependent cannibalism
in subterranean amblyopsids seems to account for
most of the mortality of the young (Poulson, 1963).
Cannibalism was also observed in Brazilian cave fishes
apparently subject to severe food shortage (Parzefall and
Trajano, 2010).
The reaction to prey by salamanders has been stud¬
ied in the facultative cave-living Pyrenean salamander
Euproctus asper, which has fully developed eyes, and
the blind Dinaric salamander Proteus anguinus. Both
species react to living and dead chironomids. Even in
light, where E. asper can use its visual sense, P. anguinus
required less time to initiate the first snapping response
to dead prey. When the time between the start of an
experiment and the first snap at prey was divided into
pre-approach and approach phases, it turned out that
the difference found could be attributed to the pre¬
approach phase (Fig. 1A). Living prey were detected
more quickly than dead prey in both species, but
E. asper needed more time in the darkness than did
P. anguinus (Fig. 1B,C). These data show that P. angui¬
nus is well adapted to search prey on the basis of chemi¬
cal and mechanical information. In contrast, E. asper
demonstrated a more directed, visually dominated
approach behavior with live prey in light and can
switch to a more active, widely foraging mode with live
prey in darkness and dead prey in light. This young
cave colonizer seems less adapted to the dark but is
capable of foraging successfully in both epigean and
hypogean habitats.
Compensating for the unpredictability of food qual¬
ity and quantity also results in physiological adapta¬
tions. Cave animals are able to survive for long periods
without food—nearly one year for invertebrates and up
to several years for caves fishes and salamanders
(Vandel, 1964). In A. fasciatus, the cavefish are able to
build up enormous fat reserves. A 1-year-old cavefish
fed ad libitum had a mean fat content of 37% fresh body
mass compared to 9% in epigean fish under the same
conditions (Hiippop, 1987).
Reproductive Behavior
Having found enough food to reach sexual matu¬
rity, the next problem to be solved by cave dwellers is
finding a sexual partner in the darkness. Subsequently,
they need behavior patterns that provide effective fer¬
tilization in the absence of any visual orientation. In
species with high population densities, it is easy for
the male to find a female. The male of the galatheid
crab Munidopsis polymorpha of the marine cave Jameos
del Aqua on Lanzarote in the Canary Islands receives
a chemical signal sent by the molting female ready for
reproduction (Parzefall and Wilkens, 1975).
In terrestrial invertebrates, a comparable situation
has been observed: The females of the cave crickets
Hadenoecus subterraneus and H. cumberlandicus that are
ready to mate release an olfactory attractant. Normally,
ENCYCLOPEDIA OF CAVES
64
Live prey,
darkness
Euproctus Proteus
FIGURE 1 Snapping response in Euproctus asper and Proteus
anguinus. The mean time interval between the start of the experi¬
ments and the first snap at prey was divided into a pre-approach
phase (hatched bars) and an approach phase (open bars) in three
different experimental treatments. Standard errors of the mean are
shown on top of the bars. Source: From Uiblein et at. (1992). Used with
permission.
several males reach the attractive female at the same
time. They can transmit information about their high
reproductive fitness by sending tactile signals through
the air by their elongated antennae, as is done by both
of the cricket species (Hubbell and Norton, 1978). In
M. polymorpha the male emits rhythmic water waves
with chelipeds (Fig. 2). They must repeat these signals
several times. The female eventually decides which male
is good, accepts the sperm transfer (or not), and escapes.
FIGURE 2 The male Munidopsis polymorpha (right) displays with
cheliped shaking in front of a female. Source: From Parzefall and
Wilkens (1975). Used with permission.
A comparable situation has been found in the live-
bearing poeciliid fish Poecilia mexicana, which lives in a
high-population-density cave habitat. The males check
conspecific females by nipping at the genital region.
The females ready for reproduction produce a species-
specific chemical signal and attractant for about 3 days
during a cycle of about 28 days. A female accepting a
male stops swimming and allows the copulation.
Normally, bigger males are preferred on the basis of
visual signals. Only the cavefish female is able to per¬
form this behavior in darkness (Fig. 3); she does so by
switching from the visual system to a lateral line system
(the fish have one lateral line system only) (Parzefall,
2001). The females of the cave population exhibited also
a preference for well-nourished males even under
completely dark conditions (Plath et al., 2005).
In the above-mentioned crab and cricket species,
data about sexual behavior in their epigean relatives
are lacking, so we cannot determine whether the
reproductive behavior has changed in adaptation to
the dark habitat. For Poecilia mexicana, comparative
data on epigean conspecifics reveal that in the epigean
habitat visually orientated sexual displays are lacking,
in contrast to other species of the genus such as
P. velifera, P. latipinna, and P. reticulata. So, P. mexicana
seems to be preadapted to cave life and has improved
its reproductive fitness in the dark by means of a
special female choice behavior based on a lateral line
system (the fish have one lateral line system only).
Species with lower population densities in cave habitats,
such as the characid fish Astyanax fasciatus or the salaman¬
der Proteus anguinus, attract conspecifics from chemical
signals transmitted in the water. Comparative studies with
the epigean proteid Necturus maculosus have demonstrated
that this information is species specific (Parzefall et al.,
1980). The animals also constantly deposit a substance
while in contact with the substrate and at communal
resting places. This substance is individual specific but
does not provide any detailed information about sex or
reproductive state; it merely brings members of the species
ENCYCLOPEDIA OF CAVES
BEHAVIORAL ADAPTATIONS
65
(A) Poecilia mexicana
(B)
1000 -
aoo -
0
CD
C/)
600 -
to
CD
CO
E
400-
u_
CD
CD
C
200 -
0)
E
1-
o-L
Female choice in darkness
n=12n.s.
2222222a
f
-,- 1 -
larged 1 small d"
EpigeanPop. Rio Oxolatun
n=20,p<0.02
- T -,--1-
large cf small c? large o’ smaller'
EpigeanPop Milky river Cave Pop. PS XUI
FIGURE 3 Female choice behavior in different populations of Poecilia mexicana in (A) light and (B) darkness. The female had the choice to
swim to the compartment of a big or a small mature male, and the time spent in a male compartment was measured. A clear partition pre¬
vented direct contact. The middle line in the box plot represents the median; the upper end of the box, the 75% value; the lower end of the
box, the 25% value. The whiskers represent the 90% and 10% values, respectively. Source: From Parzefall (2001). Used with permission.
together. For recognition of sex and reproductive state,
Proteus anguinus requires direct contact. When sexually
motivated, a male establishes a territory that a female may
enter only after direct body contact. The male sends a
chemical signal by fanning his tail against the female and
from time to time begins to walk away. The female follows
and nips at the genital region of the male. After a short
walk, they stop and the male deposits a spermatophore,
which the female retrieves and places in her cloacal region.
Unfortunately, comparative data in the epigean salaman¬
der Necturus maculosus are lacking.
Aggressive Behavior
Aggressive behavior consists of different patterns of
threatening postures and attacks followed by fights.
This behavior has various functional aspects and is
absolutely necessary in darkness; therefore, it has to be
adapted to cave conditions.
Food competition results in food territories for
groups, pairs, or solitary animals. Within a group, a
limited food supply can lead to aggressive encounters.
In general, defending food resources is only adaptive
when the costs are not higher than the potential incom¬
ing energy of the food. The majority of data available
regarding aggression among cave and epigean popula¬
tions are for the characid fish Astyanax fasciatus. The
epigean form is widely distributed in Mexico. When
undisturbed, the epigean fish defends small territories
of 10 to 20 cm, depending on body size, by fin spread¬
ing, snake swimming, and ramming (Fig. 4). In the lab¬
oratory, epigean fish of both sexes display the entire
ENCYCLOPEDIA OF CAVES
66
B
FIGURE 4 Aggressive patterns in the epigean Astyanax fasciatus. (A) Aggressive fine erection; the head-down position of the fish on the
right expresses a higher aggressive motivation. (B) Snake swimming is shown by the fish on the right and aggressive fin erection by the fish
on the left. (C) The fish on the left is ramming against the one on the right. (D) Both fish show circling and tail beating. Source: From ParzefaU
and Htiusberg (2001). Used with permission.
aggressive pattern. The subdominant fish demonstrates
submission by a head-up position and trying to hide or
escape. The fights can be very strong; in smaller
aquaria that offer no place to hide, the death of the sub¬
dominant fish can result.
Aggressive behavior depends on optical releasers.
Using dummies of different types, it has been found
that natural shape and locomotion are important visual
signals. Tests with infrared video have shown that the
epigean fish is not able to perform aggressive patterns
in complete darkness (Hausberg, 1995). They do not
establish territories at all. From such data we can con¬
clude that epigean fish, when colonizing caves, were no
longer able to perform aggressive encounters. In the
blind populations of the Pachon, Piedras, and Yerbaniz
caves, Hausberg (1995) noted a high percentage of fish
with injuries on fins and scales. In an experiment with
the Pachon population, the number of injured fish
increased in the absence of food. The aggressive behav¬
ior observed within this fish includes defending small
territories of a few centimeters by biting, circling, and
tail beating (Fig. 5). Also in these experiments, a striking
difference in swimming behavior was observed; the fish
that were regularly fed glided slowly through the water
of the entire aquarium without initiating aggressive
encounters against conspecifics. When food was lack¬
ing, the locomotor activity decreased. The fish mostly
hovered at the bottom and rhythmically flicked their
fins; fish entering the small area of a few centimeters
were attacked. The territory size was correlated with
the aggressiveness of the fish, and the aggressive pat¬
terns differed from those shown in epigean fish. The
cavefish has developed an aggressive behavior with sig¬
nals that are only effective in close body contact.
Among cave-living invertebrates, the galatheid crab
Mimidopsis polymorpha of the marine cave Jameos del
Aqua on Lanzarote in the Canary Islands feeds mainly
on diatoms on lava rocks. The animals keep a minimal
distance from one another according to the length of their
second antenna (Fig. 6). Any closer than this distance,
and Mimidopsis attacks with its extended chelipeds
and by snapping its pincers. This behavior does not
ENCYCLOPEDIA OF CAVES
BEHAVIORAL ADAPTATIONS
67
FIGURE 5 Ramming and circling in the Pachon cave population
of Astyanax fasciatus. Source: From Parzefall and Hausberg (2001). Used
with permission.
FIGURE 6 Aggressive interactions among Munidopsis polymorpha.
Source: From Parzefall and Wilkens (1975). Used with permission.
depend on optical releasers but on water movements
(Parzefall and Wilkens, 1975). The aggressive patterns of
Munidopsis are very similar to the one described for the
deep-water, bottom-living, epigean galatheid Munida
sarsi. The author believes that, despite their naturally dim
environment, vision is still the primary sense involved in
the aggressive behavior of this species, so it appears that
the aggressive behavior of galatheids is effective in light
and in darkness, with no striking differences.
In the blind cave salamander Proteus anguinus, studies
in the laboratory have revealed that males show aggres¬
sive behavior and territoriality for only a very short
(B)
FIGURE 7 Aggressive behavior of a Proteus anguinus male against
an intruder in his territory: (A) body contact (chemical identifica¬
tion); (B) tail beating; (C) biting. Source: From Parzefall (1976). Used
with permission.
reproductive period. Normally, the animals rest under
stones in groups of both sex without any aggressive reac¬
tion. When a male became sexually active, it begins to
control its conspecifics by contacts with his snout and
allows only females in the reproductive state to remain
in the hiding place. Intruders will be attacked by tail
beating, ramming, and biting in close body contact
(Fig. 7). After being attacked in a particular territory,
Proteus avoids that territory for several days on the basis
of chemical cues on the substrate. In the Poeciliid fish
Poecilia mexicana, males use aggressive behavior to estab¬
lish a size-dependent rank order within a mixed school.
The females have a reproductive cycle of about 28 days
and are attractive to males within the first 3 days of the
cycle. The dominant male controls the females by nip¬
ping in response to an attractive female. In the field, the
pair separates from the shoal and become more or less
sedentary. The male nips and tries to copulate while also
defending the female. During aggressive encounters
with more or less equal-sized males, small males use a
female-like body coloration to try to sneak copulations
(Parzefall, 1979). The population of P. mexicana that
ENCYCLOPEDIA OF CAVES
68
B
colonized a limestone cave in Tabasco (Mexico) does
not school, and the males do not fight. In laboratory
studies with epigean fish and cavefish having functio¬
nal eyes, a quantitative, genetically based reduction of
aggressive patterns and schooling has been demon¬
strated (Parzefall, 1979). The reaction is highly variable
within the population. Some of the cavefishes tested
seemed unable to understand the attacks and answered
by nipping and copulation attempts. It seems that
aggression in these cavefishes is a disadvantage, because
fighting males risk losing the opportunity for contact
with an attractive female in darkness.
DISCUSSION
Studies of behavior in cave dwellers have revealed
complex systems of responses to visual, chemical, and
tactile stimuli. Many animals can survive in complete
darkness with no visual signals. The use of weak elec¬
tric signals among cave dwellers has not been detected.
In some cases, an existing behavior (such as the aggres¬
sion exhibited by Astyanax fasciatus ) has changed to a
more effective behavioral system. These changes are
always based on existing above ground behavior, and
no completely new behavioral character has been found
in cave animals.
Acknowledgments
The author would like to thank the DFG (German research board)
for financial support over many years, the Mexican government for
providing research and collecting permits over many years, and M.
Hanel (Zoological Institute Hamburg) for drawing the figures.
See Also the Following Articles
Adaptation to Low Food
Morphological Adaptations
Natural Selection in Caves
Bibliography
Hausberg, C. (1995). Das aggressionsverhalten von Astyanax fasciatus
(Characidae, Teleostei): Zur ontogenie, genetik, und evolution der
epigaischen und hypogaischen form. Ph.D. dissertation.
University of Hamburg, Germany.
Hubbell, T. H., & Norton, R. M. (1978). The systematics and biology
of the cave-crickets of the North American tribe Hadenicini
(Orthoptera, Saltatoria, Ensifera, Rhaphidophoridae, Dolichopodinae).
Miscellaneous Publications of the Museum of Zoology of the University
of Michigan, 156, 1—124.
Hiippop, K. (1987). Food finding ability in cave fish ( Astyanax fascia¬
tus). International Journal of Speleology, 16, 59—66.
Parzefall, J. Z. (1976). Die Rolle der chemischen Information im
Verhalten des Grotten-olms Proteus anguinus LAUR. (Proteidae,
Urodela). Tierpsychol, 42, 29—49.
Parzefall, J. (1979). Zur genetik und biologischen bedeutung des
aggressionsverhaltens von Poecilia sphenops (Pisces, Poeciliidae):
Untersuchungen an Bastarden ober- und unterirdisch lebender
Populationen. Zeitschrift fur Tierpsycliologie, 50, 399—422.
Parzefall, J. (1989). Sexual and aggressive behaviour in species
hybrids of Poecilia mexicana and Poecilia velifera (Pisces,
Poeciliidae). Ethology, 82, 101—115.
Parzefall, J. (2001). A review of morphological and behavioural
changes in the cave molly Poecilia mexicana, from Tabasco,
Mexico. Environmental Biology of Fishes, 62, 263—275.
Parzefall, J., & Hausberg, C. (2001). Ontogeny of the aggressive
behaviour in epigean and hypogean populations of Astyanax fas¬
ciatus (Characidae, Teleostei) and their hybrids. Me'moires de
Biospeologie, 28, 153—157.
Parzefall, J., & Trajano, E. (2010). Behavioural patterns in subterra¬
nean fishes. In E. Trajano, M. E. Bichuette & B. G. Kapoor (Eds.),
Biology of subterranean fishes (pp. 81 — 114). Enfield, NH: Science
Publishers.
Parzefall, J., & Wilkens, H. (1975). Zur ethologie augenreduzierter
tiere: Untersuchungen an Munidopsis polymorpha koelbel
(Anomura, Galatheidae). Annales de Speleologie, 30, 325—335.
Parzefall, J., Durand, J. P., & Richard, B. (1980). Chemical communica¬
tion in Necturus maculosus and his cave-living relative Proteus angu¬
inus (Proteidae, Urodela). Zeitschrift fur Tierpsychologie, 53, 133—138.
Plath, M., Heubel, K. U., de Leon, F. G., & Schlupp, I. (2005). Cave
molly females ( Poecilia mexicana, Poeciliidae, Teleostei) like well-
fed males. Behavioral Ecology and Sociobiology, 58, 144—151.
Poulson, T. L. (1963). Cave adaptation in amblyopsid fishes.
American Midland Naturalist, 70, 257—290.
Uiblein, F., Durand, J. P., Juberthie, C., & Parzefall, J. (1992).
Predation in caves: effects of prey immobility and darkness on
the foraging behaviour of two salamanders, Euproctus asper and
Proteus anguinus. Behavioural Processes, 28, 33—40.
Vandel, A. (1964). Biospeologie la biologie des animaux cavernicoles.
Paris: Gauthier-Villars.
BREAKDOWN
Elizabeth L. White
Hydrologic Investigations
Caves as seen by cave explorers are remarkably stable.
Cavers very rarely encounter rockfalls or evidence of
recent rockfalls. Yet caves are frequently littered with
the piles of rock fragments known as breakdown. The
scale of breakdown ranges from small isolated loose
rocks to ceiling collapses that may completely block the
passage. To understand the presence of breakdown it is
necessary to consider both breakdown mechanics and
also the geological processes that could set the stage for
breakdown.
BREAKDOWN MORPHOLOGY
Piles of breakdown are unsorted and highly per¬
meable. Layering is undistinguishable or nonexistent.
Breakdown is an essential part of the cave landscape.
One can distinguish small-scale breakdown features
that are the various types of breakdown blocks
Encyclopedia of Caves.
© 2012 Elsevier Inc. All rights reserved.
BREAKDOWN
69
themselves and large-scale features that are cavern
features consisting of (or generated by) breakdown
processes.
Small-Scale Features
Breakdown can be classified by the relationship
of individual blocks to the bedding of the parent
bedrock:
Block breakdown consists of masses of rock with
more than one bed remaining as a coherent unit.
Slab breakdown consists of fragments of single beds.
Chip breakdown consists of small rock chips
and shards derived from the fragmentation of
individual beds.
This scheme has the advantage that breakdown
observed in the field can be properly classified without
speculation as to its origin; however, it has the disad¬
vantage of also being a function of the limestone lithol¬
ogy. Thus, limestone fragments of a given intermediate
size might be blocks if derived from a thin-bedded
limestone or slabs if derived from a massively bedded
limestone. In general, however, this classification has
been found useful for the areas studied.
The classification scheme given above is very broad
and general. For any specific investigation, it is possi¬
ble to derive much more detailed classification
schemes based on the characteristics of the fragment
surface, whether the fragments are dissolutional or
fracture, the fragment shape, the predominant mode of
origin, the morphological class of trench block, fault
wedge, pendant cluster or other form, and the position
of the breakdown features in relationship to the cave
passage (Jameson, 1991).
Block breakdown can be massive; blocks measure
up to tens of meters on a side and are usually bounded
by bedding planes along the bedding and by joint
planes across the bedding. Slab breakdown has a plate
shape, with slab thickness being controlled by the
thickness of the beds; the width of individual slabs
varies from tens of centimeters to many meters. Chip
breakdown ranges in size from centimeters to tens of
centimeters; the shape of chip breakdown is variable
and dependent on the process that created the break¬
down. Crystal wedging, frost pry, and closely spaced
joints produce very angular chunks, whereas pressure-
induced spalling and mineral replacement produce
flatter, more irregular shards.
In Mammoth Cave, slab breakdown is the most com¬
mon and is distributed through all levels of the cave.
Block breakdown occurs where major roof collapse has
taken place and where dividing walls have fallen
between coalescing vertical shafts. An extensive
breakdown has occurred in the upper gallery of the Great
Salts Cave Section. This passage is floored with block and
slab breakdown to a depth of 12 or more meters for a dis¬
tance of more than a kilometer. The largest breakdown
block so far observed in this passage is a single block
19 m long, 4.5 m wide, and 1 m thick.
Large-Scale Features
Although breakdown blocks form a variety of fea¬
tures in caves ranging from a few scattered blocks to
major collapsed passages, it is useful to describe two
types of features: terminal breakdown and breakout
domes. Terminal breakdown occurs at the end of col¬
lapsed major cave passages. Eroding valleys on the
surface have the effect of causing collapse (break¬
down) in the caves below. Massive breakdown that
completely occludes a cave passage is referred to as
terminal breakdown. In a number of caves, artificial
entrances have been created where the cave passage
would have intersected the surface topographic
valley. In the Central Kentucky Karst, a terminal
breakdown is the most common terminator of
passages. Often the terminal breakdown contains
sandstone as well as limestone fragments where the
collapse has extended upward to the overlying
caprock. Major trunk passages beneath the sand¬
stone-capped plateau were once continuous feeder
conduits carrying groundwater from the Sinkhole
Plain to the south and east of the plateau to Green
River in the north. These formerly continuous pas¬
sages have been truncated by ceiling collapse. Some
are actual intersections of the passages with the sur¬
face; others have collapsed at depth. The present-day
configuration of the cave system is due in large part
to these random features of collapse. Similar termi¬
nal breakdown occurs in many other caves with and
without caprock.
Breakout domes, among the most remarkable of cavern
features, are the huge rooms that form as a result of
major ceiling collapse. Some of these, such as Chief City
in Mammoth Cave, have floor dimensions of more than
100 m and ceiling heights of 30 m. The Rumble Room in
the Rumbling Falls Cave System under Fall Creek Falls
State Park in Tennessee has a ceiling about 20 stories
tall. This breakout dome is the largest in the eastern
United States and the second-largest breakout dome
known in the United States. Other breakout domes
include the large rooms in Camps Gulf Cave
(Tennessee) (Fig. 1), Rothrock's Cathedral in Wyandotte
Cave (Indiana), the entrance room in Hellhole Cave
(West Virginia), the entrance room in Marvel Cave
(Missouri), Devil's Sinkhole (Texas), and Salle de la
Verna in Pierre Saint-Martin (France). The details of the
ENCYCLOPEDIA OF CAVES
70
B
FIGURE 1 Breakout dome in Camps Gulf Cave, Tennessee. Note
figures for scale. Photo courtesy of Peter and Ann Bosted. Used with
permission.
enlargement mechanism are less clear in Devil's
Sinkhole than in the others mentioned, although it also
has the beehive-shaped room and the gigantic debris
cone typical of all breakout domes.
Careful examination of many breakdown areas
reveals a continuum of sizes, from very large break¬
down rooms to small, roughly circular or elliptical
breakdown areas in cave ceilings. The features at the
small end of the scale are sometimes only 3 m in diam¬
eter and involve only one or two beds. The morpho¬
logical term breakout dome describes all such features,
regardless of their size.
Debris piles vary in size from dome to dome, but in
those domes that are accessible the volume of debris is
much smaller than the enclosing volume of the dome.
Because the bulk density of the debris cone is consider¬
ably less than that of the original bedrock, it is appar¬
ent that large quantities of material must have been
removed. Large breakout domes must therefore have
formed at a time when water was actively circulating
near their base. The dome could then enlarge by disso¬
lution action of fallen blocks with concurrent stoping
of the sides. The dome itself is usually circular or ellip¬
tical in contour. The top is often capped by a single
massive bed.
BREAKD OWN MECHANICS
Most mechanisms advanced for cave breakdown
have been drawn from the rock mechanics of mining.
However, caves are not mines. Mines are fresh excava¬
tions in the bedrock and it is the stress pattern around
the opening that determines such hazards as rock
bursts, spalling, and roof collapse (Brady and Brown,
1993). Caves are thousands of years old and were exca¬
vated by slow dissolution of the bedrock rather than
by blasting. The horizontal strains in the cave would
be expected to have annealed out and the only residual
strain is that due to gravitational loading of the cave
ceiling by the overlying beds.
It is not known for certain that all horizontal strains
caused by tectonic deformation of the strata have been
annealed. Massive ceiling failures in limestone mines
are thought to be due to compressive stresses parallel
to the bedding (Esterhuizen et al., 2007). The circular
or elliptical pattern and dome-shaped residual cavity
seen in mine collapses have some similarity to break¬
out domes in caves (Fig. 1) and it is possible that simi¬
lar mechanisms are responsible.
The simplest model is that of brittle fracture of fixed
ceiling beams under gravitational loading (White and
White, 1969). The concept dates back to the work of
Davies (1951), who based his model on the mining lit¬
erature. The model assumes a rectangular passage
formed in near-horizontal limestone. A small amount
of elastic sag of the unsupported roof beds causes
these beds to separate slightly. Each bed offers some
support to the beds above it so that the stress pattern
closes to an arch or stress dome in the rock above the
cave passage. The gravitational load due to the rock
above the top of the stress dome is distributed onto
the rock that makes up the wall of the cave passage
and does not contribute to passage instability. Cave
passages at great depths are not less stable than more
shallow passages. It is only if surface erosion or other
geological processes breach the stress dome that the
passage stability is reduced.
Figure 2 shows the parameters of the fixed-beam
model and the dome shape of the stress distribution.
The beds act as fixed beams across the cave passage.
For the ceiling to be stable, the bending strength of the
beams must be greater than the gravitational force act¬
ing on the weight of the unsupported span. Thicker
beds are stronger than thin beds. There will be a criti¬
cal thickness (fcrit) for any given passage width at
which the strength of the bed is just sufficient to sup¬
port its weight. When the mechanics are worked out,
only the length of the beam (L) and the critical bed
thickness (fcrit) remain. The width of the beam cancels
out. The beam length (L) is set equal to the passage
width, while the extent of the ceiling bed along the
axis of the cave passage does not enter the calculation.
Figure 3 shows the roof stability according to the
fixed-beam model. For a fixed beam, the critical thick¬
ness (m) is:
ENCYCLOPEDIA OF CAVES
BREAKDOWN
71
FIGURE 2 Sketch showing the parameters for the fixed beam
model of cave breakdown. Note that the stress arch is a sketch, not
the result of a calculation.
FIGURE 3 Plot of cave ceiling stability based on the fixed beam
model and a range of bed shear strengths as indicated.
tcrit 2 S cos e
( 1 )
where p is the density of the bedrock (in kg/m 3 ), 0 is the
bedding dip (in degrees), and S is the flexural strength
(in MPa). If the ceiling beds are not supported at both sides
of the cave passage, they are treated as cantilever beams for
which the critical thickness is:
tcrit
3 p L 2
2 S cos 9
( 2 )
The fixed-beam model implies a completely elastic
response of the ceiling beds. It does not allow for plastic
deformation and long-term creep that could lead to bed
failure in the absence of any geologic triggers. It also has
no time dependence; a stable ceiling would remain
stable until some geologic process destabilized it.
If the fixed beam, brittle fracture model is complete,
explanations must be found for any recent breakdown.
Numerous occurrences of recent breakdown have been
reported. Some have occurred near entrances where
freeze—thaw cycles may be responsible. Some have
occurred deep in the cave but in areas of active vadose
water. However, a fraction of recent breakdown occur¬
rences are in dry passages with no obvious triggering
mechanism. There have been four documented roof
failures in Mammoth Cave in the past century. Three
were massive rockfalls, but the fourth involved a plas¬
tically deformed ceiling slab that had been mapped in
detail. The slab displayed extensive plastic deforma¬
tion in the 1960s. Sometime in the early 1970s it fell.
A more comprehensive model allows for inelastic
creep (Tharp and Holdrege, 1994; Tharp, 1995).
Materials break through a mechanism of crack propa¬
gation. Tharp's model is based on the propagation of
microcracks, which allow deformation and creep. The
crack propagation velocity is given by:
v = c
(3)
where K : is the stress intensity (MPa m 0 ' 5 ), K Ic is the frac¬
ture toughness of the bedrock, and c is a constant related to
the activation energy for crack movement. The parameter c
is given by:
c = V 0 e
_ H
TTT
(4)
where H is the enthalpy of activation = 67—147 kf/mole; R
is the gas constant = 8.3145 J/K; T is temperature in kelvins
and V 0 and n are fitting constants.
Propagation of microcracks allows inelastic defor¬
mation and also a time-to-failure. Using the Tharp
model, the time-to-failure is the time scale of crack
propagation through beds of nominal thickness. The
timeframe can range from thousands of years to as
much as 1 million years. Thus, the time-to-failure can
be in the same range as the age of the cave passage,
implying that breakdown can occur at any time, even
in the absence of geologic triggering processes.
The Tharp model introduces fracture toughness as
another parameter in addition to the flexural strength
for determining whether particular beds will collapse.
The Paleozoic limestones of the eastern United States
(where many of the breakdown investigations have
been made) are dense, fine-grained rocks. Coarse frac¬
turing occurs along joints and bedding plane partings.
But, within the rock mass there is little to inhibit crack
propagation, and these rocks break mainly by brittle
fracture. It is for this reason that the fixed-beam model
has worked so well.
Porous and vuggy rocks, such as the Tertiary lime¬
stone beds of the Caribbean, may have a lower
ENCYCLOPEDIA OF CAVES
72
B
flexural strength but they have higher fracture tough¬
ness, because pores and vugs inhibit crack propaga¬
tion. The caves on Mona Island in Puerto Rico have
large, relatively flat chambers with little breakdown
because of the toughness of the porous, young lime¬
stone beds. Because of the inability of the limestone
beds to propagate cracks, nearly-flat roof spans of
30 m or more are found throughout these caves.
GEOLOGIC INFLUENCES ON
BREAKD OWN PROCESSES
Geologic Processes That Initiate Breakdown
Within the context of either the fixed-beam or Tharp
model, any geologic process that lengthen the beams or
convert fixed beams into cantilever beams can move the
ceiling beds from stable to unstable configurations.
Beam thickness, flexural strength, and fracture tough¬
ness are properties of the bedrock and do not change
during evolution of the cave passage. A list of triggering
processes, not necessarily complete, is as follows:
Passage enlargement below the water table. Phreatic
passages continue to enlarge as water flows
through them. If the hydrologic conditions are such
that the passage is not drained, it may continue to
enlarge until it becomes mechanically unstable.
Removal of buoyant support. By Archimedes'
principle, the ceiling beds of a water-filled cave
passage are buoyed upward by a force
proportional to the ratio of the density of water to
the density of the bedrock. For limestones, with
typical densities in the range of 2.65 g/cm 3 , 35% to
42% of the buoyant support of the ceiling is lost
when the cave is drained.
Effects of base-level back flooding. During the time
when the emergent cave passage is in the
floodwater zone, rises and falls in base level
alternately fill and drain the cave passage.
Additional dissolution at this time, particularly
dissolution along ceiling joints, can turn fixed
beams into cantilever beams and destabilize the
ceiling.
Action of vadose water. Formation of vertical shafts,
solution chimneys, and solutionally enlarged
fractures by the action of undersaturated vadose
water often has the effect of cutting ceiling beds,
thus changing fixed beams into cantilever beams
(Fig. 4).
Ice wedging. Caves that draw in cold winter air
can have freezing conditions some distance
inside. When water moving through joints and
bedding plane partings freezes, the expansion
creates enough force to fracture the bedrock.
FIGURE 4 Collapsed ceiling slabs due to a ceiling channel cutting
the bedding perpendicular to the cave passage. Pohl Avenue,
Mammoth Cave, Kentucky. Photo tnj the author.
Crystal wedging. Replacement of calcite in the
bedrock by other minerals can exert a wedging
effect. Because gypsum has a greater volume
than the calcite it replaces, enough force is
generated to fracture the bedrock.
Crystal Wedging or Limestone Replacement to
Initiate Breakdown
Many breakdown areas in caves with extensive sulfate
minerals (primarily gypsum) suggest that crystal wedg¬
ing and replacement of limestone by gypsum and possi¬
bly other sulfate minerals are important factors in this
type of cavern collapse. Features that are characteristic of
mineral-activated breakdown are (1) walls and ceilings
fractured in irregular patterns, often with visible veins of
gypsum following the fractures; (2) breakdown consisting
of thin, irregular splinters and shards of bedrock (Fig. 5);
and (3) curved plates of bedrock ranging in size from a
few centimeters to more than a meter hanging from the
ceiling at steep angles cemented only by a thin layer of
gypsum. Microscope examination of thin sections of the
bent beds shows that the sagging and bending are due to
the direct replacement of limestone by gypsum. Another
characteristic feature is the collapses that take the form of
symmetrical mounds with coarse, irregular blocks at the
base grading upward into a rock floor at the top.
Crystal wedging produces a subset of chip break¬
down. Chip breakdown consists of rock fragments that
are smaller than individual bedding plane slabs and can
result from many processes, including purely mechanical
ones. Crystal wedging breakdown appears to be of two
types. Type I consists of angular rock fragments broken
on sharp planes that cut the bedding planes. Type I
breakdown, with fractures filled with gypsum, results
from mechanical wedging due to crystallization of the
gypsum. Similar rock fragments are found near cave
ENCYCLOPEDIA OF CAVES
BREAKDOWN
73
FIGURE 5 Shards of breakdown created by replacement of lime¬
stone by gypsum. Turner Avenue, Mammoth Cave, Kentucky. Photo
by the author.
entrances, where they result from frost action. Type II
breakdown is more complex. The fragments and plates
are angular and sharp and are fractured across the usual
zones of weakness—bedding planes and joints (Fig. 5).
Many of the fragments, only a few centimeters on a side
and less than a millimeter thick, crush like broken glass
when walked upon. These irregular plates are the signa¬
ture of the crystal replacement process. The limestone
bedrock is shattered and intermixed with gypsum so that
the passage walls become piles of rubble. Unlike most
breakdown, which results from purely mechanical pro¬
cesses, Type II breakdown involves a complex chemistry
of mineral replacement, the details of which are still
under investigation (White and White, 2003).
ROLE OF BREAKDOWN IN
SPELEOLOGICAL PROCESSES
Both geologic triggering and slow creep of beds
under load assure that breakdown can occur at any time
during the evolutionary history of a cave passage; how¬
ever, breakdown processes are most active during the
enlargement phase of cave development and during the
decay phase of cave development. The role of break¬
down in the enlargement phase includes the following:
Breakdown during the enlargement phase exposes
more limestone surfaces and thus increases the rate
of dissolution.
Upward stoping by breakdown processes can create
large chambers, if actively circulating water removes
the breakdown blocks at floor level.
Upward stoping along fracture zones with removal
of fallen blocks results in the formation of stoping
shafts; Sotano de las Golondrinas in Mexico is an
outstanding example, as are the tiankengs of China.
Breakdown in master conduits, particularly during
the floodwater stage, can provide a support
structure for groundwater dams. Silt and clay that
deposit behind the blockage seal the dam, raise
hydraulic heads upstream, and thereby generate a
hydraulic gradient for the formation of new tap-off
passages.
Breakdown continues to play a role during the stag¬
nation and decay phases of cave development as
follows:
Breakdown processes can stope upward to interconnect
previously isolated cave levels into an integrated
system of passages.
Truncation of cave passages by the formation of
terminal breakdown is a dominant process in the
breakup of continuous conduits into the fragments
characteristic of the decay stage of cave development.
The lowering land surface intersects the stress
dome, thus destabilizing the underlying cave
passage.
The final phase in the decay of caves is the passage
collapse that takes place when the eroding land
surface intersects the stress dome in the rocks of the
cave ceiling.
The final residue of a cave is a rubble zone consist¬
ing largely of breakdown.
See Also the Following Article
Clastic Sediments in Caves
Bibliography
Brady, B. H. G., & Brown, E. T. (1993). Rock mechanics for underground
mining. London: Chapman and Hall.
Davies, W. E. (1951). Mechanics of cavern breakdown. National
Speleological Society Bulletin, 13, 36—43.
Esterhuizen, G. S., Dolinar, D. R., Ellenberger, J. L., Prosser, L. ]., &
Iannacchione, A. T. (2007). Roof stability issues in underground
limestone mines in the United States. In S. S. Peng, C. Mark, G.
Finfinger, S. Tadolini, A. W. Khair, K. Heasley & Y. Luo (Eds.),
Proceedings of the 26th international conference of ground control in min¬
ing (pp. 336—343). Morgantown, WV: West Virginia University.
Jameson, R. A. (1991). Concept and classification of cave breakdown:
An analysis of patterns of collapse in Friars Hole Cave System,
West Virginia. In E. H. Kastning & K. M. Kastning (Eds.),
Proceedings of the Appalachian karst symposium (pp. 35—44).
Huntsville, AL: National Speleological Society.
Tharp, T. M. (1995). Design against collapse of karst caverns. In B. F.
Beck (Ed.), Karst geohazards (pp. 397—406). Rotterdam: A. A.
Balkema.
Tharp, T. M., & Holdrege, T. J. (1994). Fracture mechanics analysis of
limestone cantilevers subject to very long term tensile stress in
natural caves. In P. P. Nelson & S. E. Laubbach (Eds.), First North
American rock mechanics symposium proceedings (pp. 817—824).
Rotterdam: A. A. Balkema.
ENCYCLOPEDIA OF CAVES
74
B
White, E. L., & White, W. B. (1969). Processes of cavern breakdown.
National Speleological Society Bulletin, 31, 83—96.
White, W. B., & White, E. L. (2003). Gypsum wedging and cavern
breakdown: Studies in the Mammoth Cave System, Kentucky.
Journal of Cave and Karst Studies, 65, 43—52.
BURNSVILLE COVE, VIRGINIA
Qregg S. Clemmer
Butler Cave Conservation Society, Inc.
INTRODUCTION
For more than half a century, cavers have pushed,
explored, and mapped the caves of Virginia's Burnsville
Cove. Located near the Bath/Highland County border
approximately 80 km west of Staunton, Virginia, this
sparsely populated area is known for its rural, scenic
character. The systematic study and mapping of
Breathing Cave by Nittany Grotto of the Pennsylvania
State University, beginning in 1954, marked the first
organized effort to chart what was then the largest and
best known cave in Burnsville Cove. Initial work pro¬
duced an overland survey, passage cross sections and
longitudinal profiles, and an analysis of the cave's for¬
mation relative to stratigraphic folds and faults.
Bevin Hewitt's dramatic aqualung dive into the Mill
Run spring in 1956 and his discovery of Aqua Cave
beyond fueled interest in finding more caves in the
Cove. Ike Nicholson's discovery of Butler Cave two
years later (Clemmer, 2001) and the rapid discovery of
more than 16 km of large cave passages there, includ¬
ing the Sinking Creek trunk, confirmed that Burnsville
Cove possessed vast underground secrets.
The 1982 Burnsville Cove Symposium summarized
the geology and cave description of the Burnsville Cove
reported to that date. Nicholson and Wefer (1982)
described five caves—Boundless, Breathing, Butler,
Better-Forgotten, and Aqua—as being part of an under¬
ground integrated karst drainage. These caves, although
not connected by human transit, comprised the Butler
Cave—Sinking Creek Cave System, described at that
time as containing approximately 35 km of mapped
passages.
In the past three decades, additional discoveries and
mappings have greatly expanded the extent and
understanding of this cave system. With more than
102 km of mapped cave passages in Burnsville Cove,
this isolated corner of the Old Dominion ranks as one
of the primary karst regions in the United States
(Fig. 1).
PHYSICAL AND GEOLOGIC SETTING
Burnsville Cove sits astride the Highland/Bath
County line, 20 km southwest of the village of McDowell,
Virginia. Bordered by Jack Mountain on the west and
Tower Hill Mountain to the east, the Cove is a broad.
(a) (b)
FIGURE 1 (A) Map of Burnsville Cove showing the drainage system, location of caves, and large surface depressions. (B) Outline map
showing principal caves in Burnsville Cove in relation to each other. Maps courtesy of the Butler Cave Conservation Society.
Encyclopedia of Caves.
© 2012 Elsevier Inc. All rights reserved.
BURNSVILLE COVE, VIRGINIA
75
dual-synclinal valley, split longitudinally by a folded anti¬
clinal ridge known as Chestnut Ridge. The region mea¬
sures approximately 8 km long and 5 km wide, the
geologic structure plunging to the northeast.
The Helderberg limestones of Silurian-Devonian age
make up the exposed karst of the Cove. The Helderberg
rocks are not a uniform calcareous sequence, a fact that
significantly impacts the speleogenesis of the area
(White and Hess, 1982). The stratigraphic labeling of the
carbonate rocks has varied over time and with different
investigators. The sequence, bottom to top, consists of
the Tonoloway, Keyser, New Creek, Corriganville, and
Licking Creek limestones. Below the carbonate sequence
is the Wills Creek Shale and above it the Oriskany
Sandstone. Thicknesses of individual beds vary signifi¬
cantly over short distances. Of these members, the
Tonoloway and Keyser each contain three thin layers of
insoluble sandstone that significantly impact cave devel¬
opment in Burnsville Cove. With water rarely able to
breach these layers, cave passages are often floored
and/or roofed by these sandstones for long distances,
lending a characteristic flat ceiling to many of the
galleries.
The geologic structure of Burnsville Cove is not
straightforward. The two synclinal valleys are sepa¬
rated by the folded anticline of Chestnut Ridge, which
extends northeast from the little village of Burnsville.
Sinkholes dot the pastures and woods, particularly
the upper (southwest) portion of Burnsville Cove,
where Burnsville Sink harbors the entrances to
Butler Cave. At the Water Sinks 5.5 km to the north¬
east, a dramatic stream insurgence flows into
Water Sink Cave (the Subway). Water drainage in
almost all other circumstances in Burnsville Cove is
subsurface (Davis and Hess, 1982), discharging at sev¬
eral springs on the Bullpasture River 8 km northeast
of Burnsville.
Aqua Cave is the primary resurgence for Sinking
Creek Valley, a western synclinal valley containing
Boundless, Butler, Breathing, Better Forgotten, Barberry,
Basswood, Buckwheat, Blind Faith, Battered Bar, Water
Sink, Owl, Helictite, and other smaller caves, as well as
the Burnsville Turnpike/Black Canyon portion of
the Chestnut Ridge Cave System. Emory Spring, a road-
covered karst spring 1.5 km upriver from the Aqua
Cave resurgence, contains no known cave but may be
the resurgence for waters in Helictite Cave as well as
the recently opened Wishing Well Cave. Cathedral
Spring, located on the Bullpasture River 800 m down¬
stream of the Aqua Cave resurgence, is the primary out¬
flow for the caves of the eastern synclinal valley: Robins
Rift, By-the-Road, and the Burns-Blarney Stone portion
of the Chestnut Ridge Cave System. Of key note here is
the drainage divide in the Chestnut Ridge Cave System
made possible by the North/South Trunk passage cut¬
ting transversely through the Chestnut Ridge anticline.
Waters in the Burnsville Turnpike (possibly the largest
contiguous underground chamber known in Virginia,
averaging 20 m wide by 15 m high and extending more
than 1200 m, with the most massive segment measuring
45 m wide and 425 m long) continue into the narrow
confines of Black Canyon to the 622 Sump only to reap¬
pear in Aqua Cave and flow into the Bullpasture River.
Waters in the Cyanide Canyon section and all of the
Blarney Stone and Burns sections of the system emerge
at Cathedral Spring.
HISTORY OF EXPLORATION
Well-known Breathing Cave was mined for saltpe¬
ter during the Civil War. It gained prominence after
World War II as a popular sport cave with members
of the newly formed National Speleological Society
(NSS). Interest spread among NSS members and
other cavers to investigate Mill Run Spring on the
Bullpasture River. Bevin Hewitt's dive into Aqua
Cave in 1956 amply demonstrated that more cave
waited to be found, and Ike Nicholson's discovery of
Butler in 1958 attracted a considerable number of
experienced cavers to Burnsville Cove. The August
1958 Sand Canyon camp expedition (Nicholson and
Wefer, 1982) "recon mapped" 4500 m of new cave.
Dozens of leads abounded. Nittany Grotto joined the
survey in November 1958 and quickly mapped more
than 8 km of intricate passages off the upstream and
downstream trunk. Upstream discoveries in 1959 took
cavers beyond Penn State Lake to the distant, joint-
controlled passages of what later became known as
Mbagintao Land. Far downstream, two parallel
sumps—Last Hope and Rats' Doom—blocked Butler's
advance to the northeast, but four years later, an
obscure lead in Kutz Pit Junction revealed the
muddy, remote galleries of Marlboro Country.
Overland surveys from Butler to Breathing plus the
added data from both cave maps pointed to a possible
connection. By 1967, a host of cavers were pushing from
both caves to connect (Nicholson and Wefer, 1982).
Despite new discoveries in Breathing and repeated
efforts at digging in both caves, no connection has been
found and the caves remain more than 150 m apart.
On Chestnut Ridge, a small pit discovered in 1959
was pushed in hopes of extending Butler downstream,
but it quickly degenerated into a vertical crawl of
daunting challenges. Ten years later, cavers returned
to the Better Forgotten Cave and hammered through
the Vertical Crawl to find more than 300 m of a large
ENCYCLOPEDIA OF CAVES
76
B
trunk passage ending in a terminal sump. Upstream
portions of the cave remained choked in breakdown.
As the 1960s ended, a number of cavers, hoping to
protect the pristine nature of Butler Cave, formed the
Butler Cave Conservation Society (BCCS). In 1975, the
BCCS purchased the 65-acre tract of land containing
the Nicholson entrance to Butler Cave and today con¬
tinues as the nation's oldest, private cave conservation
preserve.
Exploration in Burnsville Cove slowed in the 1970s.
The Robins Rift dig was a dynamic project in the rela¬
tively caveless eastern synclinal valley and eventually
led to approximately 600 m of discovery, but surface
digs in the Cove yielded little significant cave.
Remapping in Butler took priority and despite a kilo¬
meter of virgin cave discovered in Mbagintao Land,
some began to feel that Burnsville Cove had yielded
most of its secrets.
In 1979, members of the Shenandoah Valley Grotto
visited an obscure blowing cave on Chestnut Ridge
first located by David Nicholson in 1957. In a series of
gritty, exhausting trips commencing in early 1982,
they dropped deep into the heart of the ridge, follow¬
ing good air and a small, contorted stream passage in
this cave they called Bobcat. Their discovery of the
North/South Trunk in 1983 led to new, exciting dis¬
coveries of extensive, large caves. Finding day trips
too short, highly fatiguing, and vastly inefficient for
mapping in such remote passages, they began camp¬
ing in the cave. From 1984 to 1990, 27 separate under¬
ground camps explored and surveyed more than
14.5 km of virgin cave, establishing Bobcat at the time
as Virginia's deepest cave.
In 1989, Ron Simmons conducted a series of cave
dives into the constricted fissures of Cathedral Spring.
After widening the cherty, underwater conduit, Simmons
mapped a larger, descending passage to a depth of
46 m. With 290 m surveyed and the cave continuing its
plunge, Simmons discontinued the exploration in the
face of serious hypothermia, extended decompression
times, and special gas requirements.
In March of 1991, ridge cavers dug into a small sink
on the eastern flank of Chestnut Ridge and found
Blarney Stone Cave. A twisting, wet passage similar to
that encountered in Bobcat led to an extensive, multi¬
level, decorated gallery of cave passages exceeding
6 km in length. Blarney Stone and Bobcat were con¬
nected in a dual team effort in August 1994, thus form¬
ing the 22.5-km Chestnut Ridge Cave System.
Protracted, in-cave excavation of an air-blowing lead
in Upper Ghost Hall led to the 2003 discovery of more
than 4 km of passages extending under the eastern val¬
ley and into the flank of Tower Hill Mountain. Aid
climbs up air-blowing Pigeon Tooth Dome and at the
far reaches of infeeding Falling Water hint at much
more discovery. Off the main trunk passage, descend¬
ing Duane's Drop led to a huge, treacherously floored-
in-breakdown trunk passage dubbed the Boulder Dash
that soon cleared into a sandy, potential camp room
named Phantom Ranch. A classic dry sump eventually
intersects the rarely seen Cathedral River which sumps
both up and downstream. An air-blowing lead on a
high wall overlooking the Boulder Dash—Opportunity
Knocks—was pushed in a series of ever brutal trips
until the team of Mike Ficco, Ed Kehs, and Jon
Lillestolen dug and mapped a physical connection
with Burns Chestnut Ridge Cave in December 2005,
establishing the Chestnut Ridge System as Virginia's
third longest cave system.
Digging for new caves continued, and in 1993
Barberry Cave (Schwartz, 1999) was discovered in an
obscure pasture sinkhole 1 km east of Burnsville. A
second entrance dug the following summer provided a
more comfortable, safer entry to a cave that had now
been mapped to a length of nearly 2 km. In November
1994, Ben Schwartz and Mike Ficco pushed an
extremely low water crawl more than 30 m to find a
large trunk passage headed to the northeast. Survey
here added 3 km of cave to Barberry, but the Barberric
Crawl soon had everyone considering a second series
of cave camps. The first campers into Barberry, how¬
ever, became temporarily trapped for several days
when high water completely sealed the Crawl. Despite
a media frenzy, the group engineered an intrepid self¬
rescue only to have the entrance later placed off-limits
by the landowner.
Undeterred, cavers resumed digging once again, this
time on the land of caver Nevin W. Davis. After drilling
a 25-cm hole 21 m down into the ceiling of the large
Barberry trunk passage beneath the Davis farm, cavers
began excavating and shoring to install a large metal
tank endwise. Further drilling and blasting, accompa¬
nied with a full complement of spectacular setbacks,
eventually opened the Big Bucks Pit entrance to
Barberry Cave in 1996. Meanwhile, an ongoing dig on
the slopes above the Water Sinks broke into virgin cave
in March 1996. Named Helictite Cave for its pretty dis¬
play of such speleothems in the entrance area, the cave
led to a rabbit warren of joint-controlled passages that
currently exceeds 11 km.
Strong air also lured cavers into Burns Chestnut
Ridge Cave. Extremely arduous trips to "bottom" this
tight cave had exhausted and frustrated a succession of
caving teams for more than three decades, but a con¬
certed effort led by Gregg Clemmer, Nevin W. Davis,
and Tom Shifflett finally revealed a sizable stream pas¬
sage 200 m below the entrance. Subsequent trips pushed
the depth to 240 m, surpassing the Chestnut Ridge Cave
System as Virginia's deepest. As previously stated, con¬
nection of Burns to the Chestnut Ridge Cave System in
ENCYCLOPEDIA OF CAVES
BURNSVILLE COVE, VIRGINIA
77
December 2005 established the system as Virginia's third
longest (32+ km) and second deepest (245 m).
Given this success, digging for caves in Burnsville
Cove accelerated. Buckwheat Cave was opened after
an hour's effort in March 1998. Pushing low crawls
and following air brought mappers to within 4 m of
connecting with the far downstream end of Barberry.
Cave diggers found Blind Faith the following March
after digging down 4 m into a blind sinkhole. The cave
has been mapped to more than 1 km of passage and
drains toward Woodzell Sink.
A dig in April 2000 yielded an improbable entrance
to the surprisingly extensive Battered Bar Cave. Atop a
narrow karst saddle on the west flank of Chestnut
Ridge, the cave currently exceeds 2100 m in mapped
length. Two months after this discovery, a utility line¬
man climbed up a roadside bank 1.5 km east of
Burnsville to check a telephone repeater and discov¬
ered a blowing hole. This By-the-Road Cave is now
gated and managed by the BCCS in a unique agree¬
ment with the Virginia Department of Transportation,
which owns the entrance.
Recent digging has opened Basswood (named for
the Virginia state champion basswood tree shadowing
the entrance) in 2002, Fuhl's Paradise Pit in 2005,
Buckets of Smoke and Water Sink (the Subway) in
2007, and Wishing Well in 2010.
DESCRIPTI ONS O F CAVES
Breathing Cave
Breathing Cave is one of the best known and most
visited noncommercial caves in Virginia. Until the dis¬
covery of Butler Cave, it was the largest known cave in
the state (Douglas, 1964). Partially mapped several
times after 1945, the cave was more completely sur¬
veyed by Nitttany Grotto of the Pennsylvania State
University in the late 1950s who charted it to a length
of 7.3 km (Holsinger, 1975). A recent survey by the
Gangsta Cavers now charts the cave to 9.9 km.
Breathing was mined for saltpeter during the Civil
War. A challenging, heavily joint-controlled, parallel
maze cave, it divides 30 m inside the large sinkhole
entrance into the Historic Section or saltpeter mine on
the left and the Main Section on the right. Here, to the
right, several kilometers of parallel, interconnected pas¬
sages trend southeast, ending in a series of very low,
narrow, wet crawls one surveyor termed pseudopsy-
phons. At the extreme terminus, the cave approaches
the Good News Passage in Butler Cave (Nicholson and
Wefer, 1982). The cave is a registered National Natural
Landmark.
Aqua Cave
Aqua Cave was discovered by Bevin Hewitt in July
1956 by diving 8 m horizontally into Mill Run Spring.
A low airway was subsequently blasted from the left
ceiling of the spring, giving cavers without dive gear a
sporting access to nearly 2 km of bracing, fairly large
river cave. The 1984 discovery of an extensive upper
level, the Big Brother section, pushed the cave to a
length of almost 3 km. Ron Simmons' dive of the
stream's in-cave resurgence charted flooded passages
headed into Chestnut Ridge, but did not discover any
air-filled continuations.
Butler Cave
With almost 27 km of mapped passages, Butler Cave
ranks as the second longest cave in Burnsville Cove.
Formed mostly in the Tonoloway limestone, extensive
portions of the cave are sandwiched between two of the
sandstones, giving some passages a distinctive flat ceil¬
ing over great expanses (White and Hess, 1982). The
cave underlies the western synclinal valley of the Cove,
a fact that in large part determines its passage layout
(Fig. 2). Entering the spacious, central Trunk Channel at
Sand Canyon, visitors follow the axis of the synchne as
they walk up- or down stream. Infeeder branch passages
intersect primarily from the west. A number of wet-
weather streams course through these side passages to
the Main Trunk. All waters in Butler Cave have been
dye-traced to their resurgence at Aqua Cave (Davis and
Hess, 1982). Although not as spectacularly decorated as
other caves in the Cove, Butler awes with its sheer vol¬
ume. It remains a cave of great challenges, be they a
novice's first ghmpse of the Moon Room or a veteran's
long trek to the remoteness of such exotic destinations as
Djibouti or the Doom Room. Like Breathing Cave, Butler
Cave is a registered National Natural Landmark. The
cave is owned and managed by the Butler Cave
Conservation Society. A second entrance to the system,
accessed by an angled 29° culvert pipe, was dug in 1998.
Better Forgotten Cave
Better Forgotten is an aptly named, tight, muddy,
multidrop cave on the west flank of Chestnut Ridge
near the Bath/Highland County line. Its 12-m pit
entrance leads to a series of narrow vertical drops as
the cave develops down-dip. This eventually intersects
a 580-m-long section of stream trunk passage that ends
downstream in a sump (Holsinger, 1975). The cave is
1200 m long and reaches a depth of 130 m. The stream
has been dye-traced to its resurgence at Aqua Cave on
the Bullpasture River. Better Forgotten Cave is owned
by the BCCS.
ENCYCLOPEDIA OF CAVES
78
FIGURE 2 The Glop Slot is a very narrow squeeze at the bottom
of the entrance pit at the Nicholson entrance to Butler Cave. Until
the opening of the SOFA entrance, all who entered the cave had to
pass through it. Photo from BCCS Oscar Estes collection. Used with
permission.
Boundless Cave
Boundless Cave opens in a small sink 800 m south¬
west of the entrance of Butler Cave. Trending north¬
east, the passage is characteristically very low and
filled with sand and cobbles. A small stream has been
traced to its resurgence in Aqua Cave more than 6 km
to the northeast (Holsinger, 1975).
Robins Rift
Opened by digging a large, cold air sink located at
the western base of Tower Hill Mountain, Robins Rift
quickly developed a notorious reputation for entrance
instability. At least four separate cave digs over the past
35 years have attempted to keep this cave open to visita¬
tion. With more than 600 m of passage surveyed, it
stood as the largest cave in the eastern synclinal valley
of the Burnsville Cove until the discovery of Blarney
Stone Cave in 1991. Access is currently blocked by yet
another entrance collapse. Water in Robins Rift has been
traced to the Cathedral Spring resurgence (Davis and
Hess, 1982).
Bobcat Cave (Chestnut Ridge Cave System)
Bobcat Cave was previously known as Chestnut
Ridge Blowing Cave (Douglas, 1964). The two
entrances to Bobcat are situated about 10 m apart near
the top of Chestnut Ridge just south of the Bath/
Highland County line. The 550-m-long entrance series
is a muddy, contorted, plunging slot following a small
stream that had been dye-traced to Cathedral Spring.
This stream is finally intersected by Tombstone Alley,
a dry, paleo-overflow segment that leads to the North/
South Trunk, a passage ranging from 6 to 21 m wide
and up to 12 m high (Rosenfeld and Shifflett, 1995).
As one of three large trunk passages in the cave, the
North/South Trunk winds through large breakdown
and exceptional displays of aragonite trees (Figs. 3
and 4). To the north, pits interrupt the trunk passage,
which appears to terminate at voluminous SVG Hall.
A 30-m lead climb here, however, leads to the blowing
Porpoise Passage, which crosses over the Chestnut
Ridge anticline. Beyond, the 6-m Mud Piton Climb
leads to the 24-m Damart drop and a second pitch of
11m, Polypro Pit. This drops into the second trunk
passage consisting of the Sixth of July Room to the
south and the highly decorated Jewel Cave/Big Bend
area to the north. Maret's Lead out of Sixth of July
rambles through big rooms floored in slippery mud
and challenging down-climbs for 600 m, ending in a
sump 190 m beneath the entrance. This has been dye-
traced to the Aqua Cave resurgence. A small, blowing
infeeder here leads to Black Canyon, 800 m of washed,
scrambling stream cave in very dark limestone. A
sporting up-climb through a cascade intersects the
beginning of the Burnsville Turnpike, the third trunk
and by far the largest. Extending more than 1200 m
with widths approaching 44 m and heights up to 30 m,
the Turnpike is one of the biggest and most remote
underground passages in Virginia. The Turnpike ends
in up-trending breakdown, with a stream entering
from a lower level of breakdown.
The North/South Trunk south of its intersection
with Tombstone Alley is a pleasant walk of 300 m
toward the Camp Room. Off the southeast side of the
Camp Room, a steeply plunging down-dip lead drops
into the Shamrock Dome area of the cave. A series of
muddy slots and down-climbs leads to Satisfaction
Junction and the 722 Sump. Southwest from the Camp
Room, but on the same relative level, the North/South
Trunk continues as a series of rambling, up- and
down-climbs. A tight slot with air can be followed
through collapsed breakdown to the South Lead
ENCYCLOPEDIA OF CAVES
BURNSVILLE COVE, VIRGINIA
79
FIGURE 3 An anthodite speleothem called the Elk Horn in
Bobcat Cave; the white vertical piece at the bottom is about 25 cm
long. Photo by Ron Simmons.
FIGURE 4 Crystal speleothems found in a section of Bobcat Cave
called the North/South Trunk. Photo by Ron Simmons.
Terminus, a comfortable room ending in more break¬
down. A small hole, dug under a ledge, drops into the
Blarney Stone continuation of the system. This cave is
considered one of the most demanding in the Old
Dominion.
Blarney Stone Cave (Chestnut Ridge
Cave System)
Blarney Stone Cave was discovered by digging open
an obscure sink on the eastern slope of Chestnut Ridge
in March 1991. Its muddy fissure entrance series—four
short rope drops and a series of challenging down-
climbs—is a shorter version of the Bobcat entrance
series. After 300-m, a small overflow tube leads to
larger cave. On an upper level, a 5-m aid climb across a
14-m-deep shaft accesses an impressive paleotrunk
named Ghost Hall. Sporting highly decorated stalag¬
mites, stalactites, totem poles, and columns. Ghost
Hall leads south to expansive Upper Ghost Hall.
Black Diamond crawl, a small stream crawl with black
gravel, also exits the south end of Ghost Hall. Black
Diamond intersects a pit that descends to the Pearly
Gates, named for a beautiful and prolific display of
cave pearls. Moon River extends upstream and down¬
stream below this point for nearly 600 m. Numerous
side passages abound here, the most notable being
Stairway to Heaven, an extensive series of challenging
up-climbs that rise more than 150 m. North of Ghost
Hall, the cave winds through the totem poles of
Leprechaun Forest. Extraordinary crystalline white
chandeliers decorate a delicate section of the wall. A
large walking passage rambles to the north, eventually
finding the obscure lead beyond the Earthworks to the
cave's connection with Bobcat Cave. Very promising
leads remain in the southeastern section of the cave
that connects with Burns, particularly climbs at the end
of Falling Water and Pigeon Tooth Dome.
Burns Chestnut Ridge Cave (Chestnut Ridge
Cave System)
Before its 2005 connection with the Chestnut Ridge
Cave System, Burns was the deepest cave in Burnsville
ENCYCLOPEDIA OF CAVES
80
B
Cove at 240 m. The entrance series of low, sinuous,
muddy crawls; tight, body-sized cracks; and plunging
slot canyons to the bottom of the cave is one of the most
arduous 500 m of cave in the United States. An impres¬
sive stream canyon below the 198-m level soon sumps
upstream but flows north to a series of cascades at the
213-m level. Here, a high lead some 12 m above the
stream leads to nearly 2-km of walking passage, eventu¬
ally giving access to the rarely visited Cathedral River,
and ultimately its connection with the Blarney Stone sec¬
tion of the Chestnut Ridge Cave System. This passage
was mapped downstream for 335 m to a point where the
water came to within 10 cm of the ceiling. The down¬
stream cave approaches Robins Rift to within 300 m.
Barberry Cave
Barberry Cave has three excavated entrances, but
only one open for access. The entrance to Big Bucks Pit
is an excavated 21-m shaft dropping into an impres¬
sive 23-m-high trunk passage. At the bottom of the pit,
this decorated, spacious stream trunk extends 670 m to
the north and 550 m to the south and is aligned with
the south end of the Burnsville Turnpike 1.6 km to the
north (Fig. 5). The stream ends in a sump, but the
trunk passage continues another 150 m to massive
breakdown and in very close proximity to Buckwheat
Cave. The WOWay, a sizable side lead entering the
main trunk passage from the west, extends via watery
passages to within 120 m of Butler Cave. An air-blow¬
ing lead at the end of this very tortuous passage holds
promise of a connection with Butler.
FIGURE 5 Morphine Waterfalls. The main stream in Barberry
Cave, Bath County, Virginia, is flowing down the bedding before
cascading over a 2-m waterfall. The main passage in this area is 12 m
wide, 20 m high, and has massive flowstone decorating the walls.
Photo by Philip C. Lucas. Used with permission.
Buckwheat Cave
Buckwheat Cave, dug open in 1998, plunges as a
walking stream passage into a series of low water
crawls, blocked by massive breakdown. Coming to
within 4 m of portions of the north extensions of
Barberry Cave, Buckwheat drains a small part of the
western flank of Chestnut Ridge. To date, 670 m of cave
have been mapped in Buckwheat to a depth of 42 m.
A connection to Barberry Cave is proven via air and
barometric studies but remains elusive.
Blind Faith Cave
Blind Faith was discovered the year after Buckwheat
Cave by digging a 4-m shaft in a small sink in the next
wooded valley 600 m north of Buckwheat. A series of
crawls and challenging down-climbs eventually drop
into a going stream passages. This degenerates down¬
stream in an extremely low, down-trending passage.
Upstream, the cave winds for several hundred meters
along the western flank of Chestnut Ridge but stops
well short of connecting to nearby Buckwheat. More
than 1000 m of cave has been charted in Blind Faith to a
depth of 48 m.
Battered Bar Cave
Battered Bar Cave is located about 450 m north of
the Blind Faith entrance on the edge of a deep sink
corresponding to the terminal end of the Burnsville
Turnpike in the Chestnut Ridge Cave System. A nar¬
row, 18-m pit leads to an even tighter slot that slopes
down 8 m to the top of a slippery 12-m-deep shaft. A
steep, muddy up-climb leads to a third drop of 4 m.
Beyond, through massive breakdown blocks, the cave
opens up dramatically. In the first big room, a fissure
leads down to the Ramp, a steeply inclined 30-m-long
chute-way floored with breakdown lingering at the
angle of repose. Two walking passages extend to the
south from the bottom of the Ramp, the left-handed
ENCYCLOPEDIA OF CAVES
BURNSVILLE COVE, VIRGINIA
81
passage dividing after some 125 m into left and right
branches. The left branch approaches the downstream
end of Blind Faith Cave, hinting at a possible connec¬
tion and ending near an unusual folded limestone fea¬
ture called the Stone Rainbow. The right branch
continues south 200 m, plunging dramatically to a
sump in a passage covered with pure white sand. The
right trunk at the bottom of the Ramp snakes through
breakdown, then extends for 300 m into a maze of
small passages ending in breakdown with some leads
blowing air. Underneath the Ramp, a stream can be
followed for 150 m into massive breakdown, including
a pool that may be diveable. Any connection with the
Burnsville Turnpike is approximately 250 m beyond.
Battered Bar has been surveyed for more than 2080 m
with a measured depth of 119 m.
Helictite Cave
Helictite Cave is formed near the top of the
Helderberg limestone sequence, dissolved mainly out
of the New Creek and Licking Creek limestones. A
vast maze cave of tubes and canyons with one major
paleostream passage, Helictite possesses dramatic
examples of dogtooth spar, cave pearls, lrelictites, and
slickenslides (Figs. 6—8). With more than 11 km of
cave mapped since its dug-open discovery in 1996,
Helictite is not a typical Burnsville Cove Cave and
drains to Emory Spring instead of Aqua.
By-the-Road Cave
Found in June 2000, the entrance is a recent sink col¬
lapse located in the eastern synclinal valley 800 m
southwest of Robins Rift. Strategically placed to offer
access to the sumped, upstream portion of the Burns
section of the Chestnut Ridge Cave System, By-the-
Road is currently mapped to an ongoing in-cave dig.
Considerable air blows from the cave in hot weather.
The cave is gated and managed by the Butler Cave
Conservation Society at the request of the Virginia
Department of Transportation.
Basswood Cave
Dug open in July 2002 after several months of exca¬
vation, Basswood Cave is accessed by a 7-m ladder and
4-m climb that lead to a series of low, narrow crawls
with additional, short pitches that end in a seasonally
wet, very low crawl. In 2008—2009, cavers spent consid¬
erable time widening an air-blowing crawl midway
through the cave to discover an additional 140 m of
cave passages, only to have the cave, still blowing
strong air, shrink to less than passable dimensions.
FIGURE 6 A cluster of helictites, the signature of Helictite Cave,
Highland County, Virginia. When this cave was first entered some of
the first speleothems seen were helictites. As exploration and surveys
continued it was realized that this cave was far richer in this type of
speleothem than other Virginia caves. This photo is part of a large
cluster of helictites located just 30 meters inside the entrance. Photo
by Arthur N. Palmer. Used with permission.
FIGURE 7 Pool spar, Helictite Cave, Highland County, Virginia:
one of the crystal forms of calcite sometimes called dogtoothed spar. In
Helictite Cave, the dogtoothed spar in this photo is located in a pool
that is about 6 m long and 2 m wide. Photo by Philip C. Lucas. Used
with permission.
ENCYCLOPEDIA OF CAVES
82
B
FIGURE 8 The Slickenslides Room. Helictite Cave has areas
where faults are intersected by passage development. Here the slick¬
enslides of a fault have dropped into the cave passage below. This
makes for a dramatic passage where the scrapings or slickenslides
are easily recognized on the ceiling and floor of this passage. Photo
by Philip C. Lucas. Used with permission.
Although physically near Barberry and Buckwheat,
Basswood Cave remains unconnected.
Fuhl’s Paradise Pit
Discovered in the spring of 2005, Fuhl's was found by
excavating a stone-filled sink that always melted snow
in winter. Its 7-m entrance pit leads to 160 m of mostly
dry, walking passage with no obvious continuations,
hinting that Fuhl's is an isolated segment of ancient
fossil trunk passage.
Water Sinks Cave (The Subway)
The discovery of this voluminous cave in 2007, now
mapped to 4.5 km, revealed virgin cave beneath long-
known Owl Cave. Although the two have no known
connection, both caves are prominent Cove insur-
gences that drain to Aqua.
Buckets of Smoke Cave
This 5-m pit sits astride the near crest of Chestnut
Ridge and is one of the highest entrances in Burnsville
Cove. Discovered in 2007, Buckets of Smoke's marginal
drainage over its 240-m mapped length has not yet been
traced to either the Cove's eastern or western synclines.
Wishing Well Cave
Wishing Well Cave was opened in the Summer of
2010 by a caver team led by Phil Lucas, digging over a
3-year period at the bottom of a 16-m shaft at the
northern end of Burnsville Cove. Initial exploration
and mapping follow a strong current of air channeled
through fractured layers of sandstone atop deeper
layers of limestone. This latest discovery in Burnsville
Cove possesses considerable potential not only as a
new cave but for its possible connection to nearby
Helictite as well as its potential connection into the
Emory Spring drainage.
SPECIAL ATTRIBUTES
A preliminary report on the cave fauna of
Burnsville Cove was published in the Burnsville Cove
Symposium (Holsinger, 1982) and identified 11 inver¬
tebrates and 8 vertebrates. Two species, the amphipod
Stygobromus conradi and a beetle ( Pseudanophtalmus ),
are endemic to Burnsville Cove. A more recent report
(Hershler et al., 1990) noted the discovery of a new
species of aquatic snail, Fontigens morrisoni, citing
Butler Cave as one of only two locations. An updated
report describes the discovery of a new species of
springtail ( Arrhopalites ) from Butler Cave, since found
in another Virginia cave outside the Cove. The
appearance of unknown animal tracks deep in the
North/South Trunk of Bobcat Cave and the finding of
a "raccoon-like" skeleton beneath the Camp Room
proved a startling discovery. Photographs and cast¬
ings were subsequently identified as belonging to
Martes pennanti, which is a fisher unreported in
Virginia for almost two centuries.
A preliminary report on cave mineralogy of the
Burnsville Cove was also published in the Burnsville
Cove Symposium (White, 1982) and characterized the
secondary mineral deposits as sparse though widely
dispersed. Recent discoveries in selected areas of the
Chestnut Ridge System, Barberry, and Helictite caves
reveal an astounding array of helictites, aragonite
trees, anthodites, cave pearls, and moon milk.
Sediment studies in Butler Cave have found evidence
of magnetic reversal as well as iron fixing filamentous
bacteria in a brownish-yellow layer of goethite.
After more than a half-century of systematic investi¬
gation, it is now understood that the caves of the
Burnsville Cove are hydrologically connected. Despite
the physical barrier of the Chestnut Ridge anticline, the
improbable presence of the North/South Trunk in the
Chestnut Ridge System provides a key underground
connection between the Butler Cave—Sinking Creek
drainage of the western syncline and the caves of the
less understood eastern syncline of the Burns—eastern
Bobcat—Blarney Stone—Cathedral Spring drainage.
Despite the connection of Bobcat and Blarney Stone
caves in 1994, and their subsequent connection to Burns
in 2005, connections between other caves of Burnsville
ENCYCLOPEDIA OF CAVES
BURNSVILLE COVE, VIRGINIA
83
remain elusive. With dedicated cavers mapping an
average of 1.6 km of new cave a year over the past half
century, Burnsville Cove continues to be recognized as
one of the world's preeminent cave/karst frontiers.
Bibliography
Clemmer, G. S. (2001). That cave just had to be there. Ike Nicholson
and the discovery of Butler Cave. In J. A. Campbell (Ed.), Virginia
Cavalcade, 50(1), 52—63. Richmond, VA: The Library of Virginia.
Davis, N. W., & Hess, J. W. (1982). Hydrogeology of the drainage
system, Burnsville Cove, Virginia. National Speleological Society
Bulletin, 44, 78-83.
Douglas, H. H. (1964). Caves of Virginia. Falls Church, VA: Virginia
Cave Survey.
Hershler, R., Holsinger, J. R., & Hubricht, L. (1990). A revision of the
North American freshwater snail genus Fontigens (Prosobranchia:
Hydrobiidae). Smithsonian Contributions to Zoology, 509, 49.
Holsinger, J. R. (1975). Descriptions of Virginia caves. Bulletin No. 85.
Charlottesville, VA: Virginia Division of Mineral Resources.
Holsinger, J. R. (1982). A preliminary report on the cave fauna of
Burnsville Cove, Virginia. National Speleological Society Bulletin, 44,
98-101.
Nicholson, I. K., & Wefer, F. W. (1982). Exploration and mapping of
the Sinking Creek System. National Speleological Society Bulletin,
44, 48-63.
Rosenfeld, J. R., & Shifflett, T. E. (1995). The caves of Burnsville
Cove, Virginia. In C. Zokaites (Ed.), Underground in the
Appalachians: A guidebook for the 1995 NSS convention (pp. 9—19).
Huntsville, AL: National Speleological Society.
Schwartz, B. (1999). Exploring Barberry Cave. NSS Neivs, 57,
268-275.
White, W. B. (1982). Mineralogy of the Butler Cave—Sinking Creek
System. National Speleological Society Bulletin, 44, 90—97.
White, W. B., & Hess, J. W. (1982). Geomorphology of Burnsville
Cove and the geology of the Butler Cave—Sinking Creek System.
National Speleological Society Bulletin, 44, 67—77.
ENCYCLOPEDIA OF CAVES
CAMPS
Qregg S. Clemmer
Butler Cave Conservation Society, Inc.
The thorough exploration and survey of an extensive
cave system demand that all participants "push the
limits" of the cave to its "bitter end." Such idealistic,
oft-used phrases employed by cavers reflect a deeply
held philosophy—an ethic some would say—that goal-
oriented, expedition-style caving requires careful prep¬
aration, long-term dedication, and extensive stamina.
"Push trips" to the bottom or to the far reaches of vast,
complex cave systems challenge all three requirements.
Planning entails cooperation, competency, specific
goals, a mountain of gear, and the occasional kilometer
of rope. Participation impacts everything from bank
accounts and vacation time to jobs and marriages.
It is a given that everyone enters the cave in superb
physical and mental condition. But what happens when
the mountain of gear is consumed, when the kilometer
of rope is rappelled, when the strongest caver is
exhausted, and the cave still goes down and down, on
and on? In the years since the founding of the National
Speleological Society, Inc. in 1941, U.S. cavers have con¬
tinued to push the limits. In the first half-century of the
Society's existence, for a variety of reasons—personal
comfort, novelty and intrigue, or simply because the cave
went on and on—cavers in a few instances resorted to
underground camps to pursue their respective goals.
Cavers in Europe confronted the same challenges.
Exploration in Switzerland's Holloch expanded to an
underground camp in 1949. After four cavers were
trapped by high water for 9 days in 1952, all explora¬
tion camps in Holloch became wintertime endeavors
(Courbon et al., 1989). The widely acclaimed 1952 descent
into Pierre Saint-Martin employed a 5-day underground
camp but ended tragically with the death of Marcel
Loubens (Tazieff, 1953). Gouffre Berger became the first
caver to break 1000 m in depth, a feat realized in 1956
which owed much to the staged underground camps of
1954-1955 at 494 m and 860 m (Cadoux, 1957). A 1955
expedition into the Cigalere also employed staged cave
camps, but, instead of pushing the cave ever deeper,
explorers confronted a daunting series of waterfalls as
they ascended into the mountain (Casteret, 1962).
The evolving European model for underground
expedition camping utilized advance supply teams to
rig pits, lay phone lines, transport mountains of gear,
and establish camps. Single-rope techniques being
unknown, pits were negotiated by cable and rope lad¬
ders backed up by belay. The early exploration of
Utah's Neff Canyon Cave in October 1953 followed
this European model and utilized a support crew to
aid four cavers on a 33-hour trip to the bottom of the
cave. Engaged in a rivalry with a local climbing club,
the group carried in a large amount of gear—sleeping
bags, cable ladders, 150 m of rope, field phones, and
coils of wire—to support their effort. But, after endur¬
ing a "fitful sleep in cold, cramped quarters" in their
unsuccessful attempt to find the cave's deepest
point, the explorers emerged "completely exhausted."
One member of the support crew spent a week in the
hospital suffering "utter fatigue." They chalked up
their failure to "bulky packs and unmanageable gear
in the narrow, jagged passageways" (Green and
Halliday, 1958).
A few months later, Floyd Collins' Crystal Cave
Expedition (C-3 Expedition) electrified the caving com¬
munity with a sensational attempt in Flint Ridge to
push the far reaches of Kentucky's most extensive cave.
With movie cameras rolling and backed by a full com¬
plement of sponsors, fawning reporters, and radio
broadcasters, the cavers entered the cave with ambi¬
tious goals for a week underground. Metal "Gurnee
cans" protected gear through tight, rocky crawl ways.
Field phones connected remote sections of the cave
with the surface support crew. Experts in cave biology,
geology, hydrology, medicine, and meteorology accom¬
panied the hard-charging explorers, all eager to mea¬
sure the cave, the cavers, and the phenomena therein.
Encyclopedia of Caves.
85
© 2012 Elsevier Inc. All rights reserved.
86
c
The expedition ended with numerous official reports
detailing everything from sleeplessness and mild shocks
from ring voltages in the phone system to the morale¬
boosting effects of candy and tobacco being delivered
from the surface. The festive, self-congratulatory tone at
the expedition's end ignored the large amounts of trash
buried or burned in the cave. Several kilometers of
abandoned telephone wire would litter the passages for
decades (Brucker and Lawrence, 1955). Thankfully, the
"success" of the C-3 Expedition was never repeated, but
in a subsequent report 2 years later one participant
recommended "simplicity in all phases of trip organiza¬
tion" as a future goal, warning that too often "success
is judged by size instead of actual accomplishment"
(Smith, 1956).
In contrast, the Butler Cave Camp of August 1958
commenced as a closely held secret. Not eager to get
"scooped" and exploring mostly in blue jeans and
wool shirts, the seven-man crew eagerly pushed deep
into this Virginia discovery without field phones, sur¬
face support crew, or scientific agenda. At the end of
an exciting, tiring week, they exited with 5 km "recon-
mapped" and a collection of superb color slides, but
they buried their trash and spent carbide in the cave.
Youthful exuberance and naivete characterized the
August 1962 cave camp in Indiana's Sullivan cave.
With lofty goals to map the cave, sample the soil for
microbes, conduct psychological surveys on partici¬
pants, and clean up extensive vandalism, the teenagers
elected to spend 2 weeks camping in the cave despite
its relatively close proximity to the entrance. Field
phones connected them to the surface. Equipped with
sleeping bags on cots and supplied with a double¬
burner Coleman stove, canned goods, rye bread, and
even fresh vegetables (celery, carrots, and lettuce), the
explorers endured a miserable, cold existence in wet,
muddy clothes despite five complete changes of
underground wardrobe. Although they mapped 2 km
of passage, the young "Sullivaneers" discovered only
half of it to be virgin cave. One participant character¬
ized part of their mapping as "a comedy of errors."
Vastly more significant discoveries rewarded a two-
person, week-long camp in Ellison's Cave, Georgia, in
1969. Despite the dramatic failure of their only stove at
base camp more than 250 m below the surface, this
man-and-woman team stomached cold food and dank
conditions to survey almost 4 km of cave without field
phones and surface crew, all with minimal impact to
the cave (Smith, 1977).
A bizarre example of cave camping occurred in
1972 in Midnight Cave, Texas, when one man spent
6 months underground. Dismissed as nothing more
than a publicity stunt by some, the venture did garner
enormous attention, including a feature article in National
Geographic. The subject entered the cave ostensibly to
investigate the long-term psychological and physiological
effects of solitary confinement and sensory deprivation.
Amenities included a canopied sleeping area on a
wooden platform, extensive incandescent lighting, field
phone, books, and record player. At the end of his time
underground, the relieved cave dweller declared his trial
a success as the "longest beyond time experiment in his¬
tory" (Siffre, 1975).
Ongoing exploration of Wind Cave in South Dakota
employed an underground camp in 1972. Situated
near the Master Room, the relatively comfortable camp
was supported by a surface crew, stocked by supply
teams, and connected to the surface by field phones.
To thwart hypothermia, participants toyed with the
novel notion of running heat lamps in the camp on
480 V piped down the telephone cable. Despite the dis¬
covery of major extensions to the cave, no one favored
a second underground camp the following year. "The
logistics of running a base camp, although successful,
were very difficult and time-consuming," wrote one
organizer. Henceforth, the survey reverted to "long,
single-day trips from the surface" (Scheltens, 1988).
All of this experience gained was lost on organizers
of Project SIMMER when 118 cavers descended on
Simmons-Mingo Cave in West Virginia in October
1973. Ambitiously planned much like a military opera¬
tion with a chain of command, mess tent, and adminis¬
tration tent, the expedition ultimately consumed 10
hours of preparation to every hour actually spent
underground. Planners managed to lay more than
15 km of wire for field phones, then touted their work
as the "world's largest in-cave communication net¬
work." The Gurnee can of the C-3 Expedition morphed
into a "Carts can," a stovepipe and plywood contrap¬
tion used to haul gear into the cave. Plagued by poor
sleeping bags, wet clothes, ringing phones, and a mis¬
erable camp spot, the crews mapped less than a kilo¬
meter of cave. Project SIMMER had profited nothing
from the C-3 experience of 20 years earlier and never
issued a final expedition report. Participants even
abandoned the phone wire in the cave. Yet, beyond
these disappointments, the overall underground man¬
ager opined at the end of the experience that for deep
or remote cave exploration, "the small camping party
[would be] more efficient than the larger, more for¬
mally organized group."
American deep-caving expeditions to Mexico also
began camping underground. European participants with
extensive expedition experience contributed a wealth of
knowledge toward maintaining a comfortable, efficient
subterranean camp. Prolonged underground stays were
begun in the mid-1960s and by the late 1970s had pushed
the reaches of vast, deep, technically difficult caves (such
as Sotano de San Agustin, La Grieta, Sumidero Yochib,
Sotano del Rio Iglesia, and Sistema Purificacion) far
ENCYCLOPEDIA OF CAVES
CAMPS
87
FIGURE 1 A meal being prepared at camp 3 in Sotano de San
Agustin.
beyond the reaches of conventional day-trip caving
(Stone, 1978). (See Figs. 1 and 2.)
Outside the warmer climes of Mexico, occasional
cave camping in the temperate latitudes of the United
States generated little appeal for second attempts.
Despite some glowing declarations of expedition suc¬
cess, the cave campers of Neff Canyon, C-3, Butler
Cave, Sullivan Cave, Ellison Cave, Wind Cave, Project
SIMMER, and even a successful, comfortable camp
deep in Fern Cave in Alabama brokered little enthusi¬
asm to repeat their adventures. Pushing caves to their
limits in the chillier continental 48 states went back to
being brutal day-trip endeavors. Remote underground
camps were best left for the warmer caves south of the
border. The 1983 discovery of a large cave system
under Chestnut Ridge, near Burnsville, Virginia, pro¬
vided the impetus for yet another group to try an
extended underground camp in the United States.
More than 20 km of challenging, decorated, virgin gal¬
leries rewarded those who endured the cold, sloppy,
tortuous entrance series of Bobcat Cave (see Figs. 3
and 4). Yet exhaustion and the real threat of hypother¬
mia limited all efforts to safely extend exploration via
increasingly longer day trips. With no other choices,
cavers with decades of grueling experience grudgingly
confronted the possibility that camping underground
was the only feasible way to continue the survey.
Given the Spartan experiences a generation earlier at
nearby Butler Cave, few relished the idea. No cave in
the United States had ever been continuously pushed
and mapped in such a manner. Nevertheless, over the
next 10 years, more than 15 km of passageways were
explored and surveyed via 27 underground camps in
Bobcat Cave, culminating in the 1994 connection with
7-km-long Blarney Stone Cave.
To be fair, few caves offer the isolation and daunting
physical challenges that justify camping underground,
but Bobcat Cave did, and once the decision to camp had
been made the question became one of how to thwart
FIGURE 2 Cooking area of the White Lead Room in La Nita,
which is part of Sistema Huautla in Oaxaca in southern Mexico.
FIGURE 3 Bobcat Cave; a view of the cooking area after several
days of use.
hypothermia, obtain adequate nourishment, maintain
endurance, and still safely and efficiently get the cave
competently explored and mapped. Custom nylon cover¬
alls were the first big difference from previous camps.
Until the early 1980s, experienced American cavers—
with rare exceptions—went underground clothed pri¬
marily in cotton and wool. "Farmer John" coveralls ruled
the day. Wet suits were tight and uncomfortable but bat¬
tled hypothermia and provided protection far better than
ENCYCLOPEDIA OF CAVES
88
c
FIGURE 4 Bobcat Cave; one caver's area in the main camp cham¬
ber. He is using a hammock instead of a sleeping pad on the ground.
blue jeans and corduroy jackets or the wool sweaters and
flannel shirts of earlier times. Comfort counted, and the
eternal cold of soggy cotton and smelly wool when still
kilometers from the entrance begged for garments prom¬
ising warmth, agility, and the ability to stay dry despite
the wearer's body heat.
Nylon did just that. Participants in the first Bobcat
camp purchased yards of the fabric, adapted a borrowed
coverall pattern, and sewed their own. Cave packs
evolved the same way. The bulky, battered metal towing
cans of the C-3 Expedition and Project SIMMER never
even came under consideration. Instead, long, cylindri¬
cal, flexible nylon duffel bags, also self-sewn, performed
admirably. With a tether on one end for upright attach¬
ment to seat harnesses when ascending or descending
drops, a handle in the middle for grasping in crawls and
crevices, and back straps for carrying over long distances,
the "camp duff" proved invaluable for getting gear and
food into camp. Double or triple thicknesses of trash
bags protected food, clothes, and sleeping bags from dev¬
astating leaks. Sucking out the air from such packed bags
before tying them off provided additional space. Two
decades later, nylon packs and coveralls enjoy almost
universal use in caving and even exude their own fash¬
ion statements, having spawned a cottage industry in
custom cave gear, vertical rigs, and personalized repair
using a variety of incredibly durable fabrics. Changing
into dry, warm camp clothes on reaching camp boosted
morale, especially if one's body heat aided the process;
thus, wickers and polypropylene replaced cotton and
wool undergarments.
Sleeping underground, though, had always been a
prolonged struggle against chill and dampness. Cotton
or down sleeping bags were dismal failures, but light¬
weight fiberfill or synthetic bags worked nicely when
laid on a foam pad (or, for example, a Therm-a-Rest
pad) atop a reflective ground cloth. A stocking cap kept
head and ears warm all night. Some campers even wore
gloves. A dry change of socks, bound up in small plastic
bags, assured dry feet even when moving about camp
in wet, muddy cave boots. An extra polypropylene top
and bottom, properly bagged, provided the luxury of a
pillow. Hammocks, although favored by some on
Mexican expeditions, were quickly abandoned after a fit¬
ful night tossing in the damp, 48° F chill of Bobcat Cave.
The camp site itself needed to be relatively level, spa¬
cious enough for sleeping quarters and a community
kitchen and eating area, and fairly close to reliable
water. A drop of iodine per gallon of water accom¬
plished water purification. The latrine was located in
respectable proximity to camp, dug into a clay bank.
Eating revolved around breakfast and dinner, sup¬
plemented during the day by personal preferences
(energy bars, gorp, cheese, candy, premade sand¬
wiches, even a baked potato). Freeze-dried food cov¬
ered most menus, being far tastier than the wretched
examples of the past and significantly lighter than
canned goods. (Note: In desert caves or where water is
scarce, canned goods could be a significant supply for
both food and water.) Tea, coffee, sugar, salt, oatmeal,
dried fruit, pepper and other spices, and even luxury
condiments were easily stuffed in zip-lock bags and
buried in the depths of packed sleeping bags.
Aside from ropes, climbing gear, bolt kits, and sur¬
vey gear, community camp gear on the initial trip
included a small white gas stove with repair kit; sev¬
eral full, secured fuel bottles; a cooking pot for hot
water; three or four collapsible plastic gallon jugs;
first-aid kit; and trowels and toilet paper, all divided
among the participants. Outside of replenishment
items, these were secured in the cave from camp to
camp. Luxury items ranged from washcloths and per¬
sonal journals to cards and a harmonica. Carbide pro¬
vided 90% of the lighting, with candles around camp
adding an intimate touch and saving acetylene.
Cavers know that adaptability and incentive remain
a vital part of pushing the limits. Future camps may
very well embrace caving LED lamps, for which
rumored 50-hour burn times on one set of four D cells
would surely lighten camp duffs of pounds of bulky
calcium carbide on the way in and spent carbide on the
way out. Every Bobcat expedition entered the cave as a
small, self-contained team. Never did more than nine
cavers (three teams of three) participate; six proved the
average. Surface crews lounging in administration tents
fielding phones attached to kilometers of wire strung
through near-virgin cave were never an option and
would probably violate the conservation ethic of today's
cavers. Instead, with a safety contact just a few kilo¬
meters from the entrance, the expedition entered the
cave with competent associates on the surface aware
and available. In the years since, cave camping has
remained a seldom-used tool of American cavers.
ENCYCLOPEDIA OF CAVES
CASTLEGUARD CAVE, CANADA
89
The recent multisump isolation of camping on a tarp
suspended above water deep in Mexico's Sistema
Huautla is surely the extreme (Stone et al., 2002), but the
continued success of camps in making new discoveries
such as in Kentucky's Fisher Ridge, New Mexico's
Lechuguilla, and Virginia's Omega System is a tribute to
cavers' adaptability to the challenging extremes of the
caves they continue to push.
See Also the Following Article
Exploration of Caves—General
Bibliography
Brucker, R. W., & Lawrence, J., Jr. (1955). The caves bei/ond: The story
of Floyd Collins' Crystal Cave exploration. New York: Funk &
Wagnall's.
Cadoux, J. (1957). One thousand meters down. London: George Allen &
Unwin.
Casteret, N. (1962). More years under the Earth. London: Neville Spearman.
Courbon, P., Chabert, P., Bosted, P., & Lindsley, K. (1989). Atlas of
great caves of the world. St. Louis, MO: Cave Books.
Green, D. J., & Halliday, W. R. (1958). America's deepest cave.
National Speleological Society Bulletin, 20, 31—37.
Scheltens, J. (1988). 50 miles beneath the wind. NSS News, 46, 5—14;
28-37.
Siffre, M. (1975). Six months alone in a cave. National Geographic, 147(3),
426-435.
Smith, M. O. (1977). The exploration and survey of Ellison's Cave,
Georgia. Birmingham, AL: Smith Print and Copy Center.
Smith, P. (1956). Seven principles of effective expedition organiza¬
tion. National Speleological Society Bulletin, 18, 46—49.
Stone, B. (1978). Underground camps for deep caves. AMCS
Activities Newsletter, 8, 37—45.
Stone, B., Am Ende, B., & Monte, P. (2002). Bei/ond the deep: The deadly
descent into the world's most treacherous cave. New York: Warner Books.
Tazieff, H. (1953). Caves of adventure. New York: Harper & Brothers.
CASTLEGUARD CAVE, CANADA
Derek Ford
McMaster University
Castleguard Cave is the longest cave system currently
known in Canada (21 km) and the foremost example any¬
where of a cavern extending underneath a modern glacier
(Fig. 1). It displays many striking features of interactions
between glaciers and karst aquifers, a complex modern
climate, rich mineralization, and a troglobitic fauna that
has possibly survived one or more ice ages beneath deep
ice cover in the heart of the Rocky Mountains (Ford
(1983); Ford et al., (2000); Muir and Ford (1985)).
GEOGRAPHICALSETTING
The cave is located in the northwest corner of Banff
National Park, Alberta, very close to the Continental
Divide. The region is one of rugged alpine mountains
with many horn peaks, cirques, and U-shaped valleys
typical of intensive glacial erosion, plus a few small
but high plateaus. The range of elevation is from
1500 m asl in the floors of trunk valleys to summits at
3500 m. Mean annual temperatures are 0 to — 14°C
across this height range. Natural boreal forests extend
up to ~2100 m, passing into grass and low shrub tun¬
dra and then alpine desert generally above 2400 m.
The Columbia Icefield is a plateau ice cap 320 km 2 in
area and 200 to 300 m thick, the largest remaining ice
mass in the Rocky Mountains. Valley glaciers radiate
up to 10 km out from it today. Ice thickness and extent
were much greater during the major glaciations, when
the glaciers extended 100 km or more from the icecap,
with only the mountain peaks protruding as nunataks.
The karst rocks are resistant carbonates of Cambrian
age. The Cathedral Formation (>560 m thick) is mas¬
sively bedded, very resistant crystalline limestone that
contains the cave. Above it, the Stephen Formation
(80 m) is a limestone shale that can block much des¬
cending groundwater but leaks readily through some
major fractures ( i.e ., it is an aquitard). It is overlain by
further thick-bedded limestones and dolostones, the
Eldon and Pika Formations. The summit strata are
mechanically weaker shales, sandstones, and dolo¬
stones. Beneath the Icefield, around the cave and north
of it, these rocks dip regularly south-southeast at 4 to
6°. South of the cave and parallel to it there is a sharp
downfold in the Cathedral rocks that caused some slip¬
page of bedding planes (thrusting) to the north. A val¬
ley is excavated along the downfold, with a glacier
from the Icefield at its head and the Castleguard River
starting at the glacier snout. Castleguard karst ground-
water drainage reaches the river via 60 or more springs.
On the surface, the Cathedral limestones host a suite
of small but typical alpine karst landforms such as kar-
ren, solution and suffosion dolines, and vertical shafts.
They are particularly well seen in the Meadows, a
broad, shallow valley north of the cave mouth (Fig. 1).
Many of these features were overridden and lightly
eroded by glaciers during a minor re-advance—the
"Little Ice Age," which occurred during the past 500
years. The glaciers are now receding. Meltwater
streams sink underground around their edges or in the
Meadows. At places, streams can be heard cascading
down shafts still concealed beneath the flowing ice.
MORPHOLOGY AND GENESIS
Castleguard Cave is a textbook example of a meteoric
water dissolutional cave in limestone. Cavers can enter it
only at its downstream end at 2010 m asl in the north
wall of Castleguard River valley, more than 300 m above
Encyclopedia of Caves.
© 2012 Elsevier Inc. All rights reserved.
90
FIGURE 1 Schematic section through Castleguard Mountain, Alberta, Canada, showing the geological formations, the location of
Castleguard Cave, Castleguard II, and the Meadows.
the valley floor. From there, the cave ascends 386 m to
terminations underneath the Icefield, where explorers
are farther from their only entrance (and exit) than in any
other known cave. There are three distinct morphologic
sections. The Headward Complex is comprised of inlet pas¬
sages beneath the modern Icefield that were created by
repeated glacial blockage and rerouting of sinking waters
in the past. The passages are plugged by glacier ice or
debris today. Younger vadose shafts pass down through
them and become blocked by constrictions or debris
below. The Downstream or Entrance Complex includes low
tunnels in two major bedding planes created by flooding
and obstructions by glacier ice in Castleguard Valley in
the past. Finally, the Central Cave is a sequence of remark¬
ably long, straight conduits created where one master
bedding plane is intersected by a pair of vertical joints
that are linked by a sedimentary dike (Grottoes Dike)
crossing them (Fig. 2 A); in the bedding plane there
is some evidence of crushing and shearing, indicating
that differential slip opened it up a little, permitting
groundwater to penetrate at its juncture with the joints
and dike.
The cave possibly originated as a single phreatic loop
beneath the Stephen impermeable cover rocks that de¬
scended more than 370 m below a paleowatertable and
then re-ascended to ancient springs just below the
Meadows. More certainly, as Castleguard Valley was
entrenched below the Stephen Formation the cave
became enlarged to nearly its modern dimensions, as
shown in Figure 2B. It was then comprised of two shal¬
low, principal loops with vadose canyon entrenchments
up to 20 m deep at their upstream ends, grading down¬
stream into phreatic tubes 4 to 5 m in diameter and of
beautiful circularity (Fig. 3). The downstream loop dis¬
charged into the Helictite Passage in the Entrance
Complex by a vertical lift (phreatic shaft) of 24 m.
Following further entrenchment of the Valley, the main
cave headwaters were diverted into a lower cave
(Castleguard II) and residual waters drained through
constricted undercapture passages in the bottoms of the
loops (Fig. 2C). The undercaptures channel local invasion
waters passing through the leaky Stephen rocks today,
but the Central Cave and Headward Complex are essen¬
tially abandoned hydrologic relicts. The Downstream
Complex, however, can be flooded with terrifying rapid¬
ity when waters pour out of another, quite independent
lifting shaft within it (Boon's Blunder), which fills the first
kilometers of low passages entirely and discharges the
waters through the explorers' entrance (Fig. 4).
Modern Hydrology
Modern waters drain underground from the glacier
soles, alpine karst, and Meadows to a set of springs
extending 3 km downstream from the Big Springs,
which are a trio of dramatic overflows 15 to 40 m above
the valley floor (Fig. 1). The waters flow through the
putative series of inaccessible caves (Castleguard II), as
illustrated in Figure 5. The Artesian Spring, the lowest
in elevation, is perennial. As the annual melt season
progresses, upstream springs such as Gravel and
Watchman springs begin to flow. The Big Springs,
ENCYCLOPEDIA OF CAVES
CASTLEGUARD CAVE, CANADA
91
P24
“ phreatic passage
\ drained phreatic passage
0,5 5ii vadose entrenchment
GROl
HOLES-IN-
THE FLOOR
HELICTITE
PASSAGE
/
SUBWAY
I
300 m.
(C)
Summer
flood water
zone
underfit trenches
7.6 km
BLUNDER
FIGURE 2 (A) The initial phreatic passages in the Central Cave and Downstream Complex showing the master bedding plane and inter¬
secting vertical fractures that guided them. They discharged into the Entrance Complex bedding planes via P24, a vertical shaft up a small
fault. (B) The cave at the close of the principal enlargement phase. Drawdown vadose canyons supplied water to a succession of shallow phre¬
atic loops. (C) The modern cave; small invasion vadose streams have cut shafts and underfit trenches in the drained galleries and are lost into
impenetrably small, undercapture passages continuing on down into Castleguard II.
100 m higher than the Artesian Spring, have a maxi¬
mum discharge >7 m 3 s -1 and handle average summer
floods. Their capacity is exceeded when there is very
strong melting at the head of the Meadows and/or on
the Icefield. Groundwater then backs up in the aquifer
until it floods Boon's Blunder and the Downstream
Complex of the cave, 270 m above the Big Springs.
Cave discharges can be >5 m 3 s _1 . Tracer dye placed in
glacier edge sinkholes can reach the Big Springs 4+ km
distant and 750 m lower in as little as 3.5 hours. This is
again a textbook example—here, of a dynamic alpine
karst aquifer. The large number and great height range
of its springs are attributed to repeated disruptions
such as debris plugs during the glaciations.
The Boon's Blunder flooded shaft was dived for the
first time in 1987, when it was found to be a phreatic
ENCYCLOPEDIA OF CAVES
92
c
FIGURE 3 The architecture of phreatic looping in
the Central Cave: (A) In the Central Grotto, a cross-
section on the master bedding plane (Fig. 2A) and
aligned on a Neptunian dike (on left). Water flowed
gently upslope here to spill down into (B) the First
Fissure, a vadose entrenchment below the master
plane (seen in roof), and then into (C) the Subway, a
textbook example of the ideal phreatic passage. At the
downstream end of the Subway, the water then flowed
upward again under hydrostatic pressure, creating (D)
the 24-m Shaft, a phreatic lifting element.
FIGURE 4 (A) The Ice Crawls; the residue of sum¬
mer floodwaters is frozen solid in the exploring sea¬
son, April. (B) The Ice Crawls end at this floodwater
phreatic lift, a 7-m shaft where the violent flow creates
a shingle beach at the foot. (C) Scene 100 meters from
the entrance in June, with the 7-m shaft in far rear.
The winter airflow in has reversed and moist flow out
of the cave has coated its frozen walls with hoarfrost.
The first flood will remove this. (D) The cave entrance
at the end of winter, overlooked by the Watchman
peak in the background.
lifting shaft like the 24-m shaft into the Central Cave.
It reaches the same master bedding plane (Fig. 2A),
where a bedding passage trends northwest up the stra-
tal dip and subparallel to the First Fissure. In a series
of dives in 2009 and 2010 the British cave diver.
Martin Groves, swam this upstream for 845 m before
reaching the air-filled subway-type passage that
remains to be explored. From his estimates of the pas¬
sage dimensions the volume of water in perched stor¬
age here is ~ 6000 m 3 , which compares very well with
ENCYCLOPEDIA OF CAVES
CASTLEGUARD CAVE, CANADA
93
Frost
Pot
Meadows
Karst Underflow
Waterflow
FIGURE 5 The proportional flow model of the inaccessible Castleguard II cave system, its feeder invasion vadose shafts, and discharge
springs that Smart (1983) derived from quantitative hydrological studies. Cross-sectional areas of individual passages are proportional to their
share of the total groundwater flow measured in the system, and their complicated pattern of interconnections is deduced from dye tracing
and flood overflow behavior.
an estimate of 5400 m 3 that Smart (1983) derived from
change of D/H ratio in the waters over the course of
the first summer flood event, that is, that the water is
being expelled by simple piston flow.
Cave Sediments, Speleothems, and Dating
Subglacial boulders, gravel, and sand were swept or
bulldozed into the head of the cave. Many have been par¬
tially cemented by calcite and later eroded, indicating a
long history of filling and removal. Throughout the
Central Cave are remnants of three partial fillings with
varved silts and clays, separated by phases of erosion or
calcite deposition. They are deposits of glacier flour, set¬
tled out of suspension on occasions when the cave
became backflooded with subglacial waters because the
Castleguard Valley was filled with flowing ice.
Despite its location beneath glaciers or alpine desert,
the cave beyond the modern entrance flood zone is well
decorated with speleothems, chiefly very pure, white
calcite. There has been much speculation about mechan¬
isms for its deposition in the absence of any sources of
soil CO 2 overhead (see Atkinson, 1983). There is one
double layer of cave pearls that are all edge-rounded
cubes 5 to 7 mm in diameter, a unique deposit. In the
warmer central sector (temperature of +2 to +3°C, rela¬
tive humidity of >95%), there are small evaporative
aureoles of aragonite, huntite, hydromagnesite, gypsum,
mirabilite, and epsomite; evaporation in these extreme
conditions is due to strong drafts blowing through the
cave when it is not in flood.
Although most speleothems appear to postdate the
latest phase of varve deposition, there are many
remains of older stalactites and stalagmites that suf¬
fered erosion during one or more varve floods.
Uranium series dating and paleomagnetic studies of
some of the oldest examples show that the cave
became relict (i.e., Castleguard II was already well
developed) more than 780,000 years ago. The varved
clays are younger. This antiquity is typical of multi¬
level alpine caves.
THE CLIMATE OF THE CAVE
Castleguard Cave is a chilly place; however, because
it passes through a big mountain (Fig. 1) the geother¬
mal heat flux is able to warm the central, most
ENCYCLOPEDIA OF CAVES
94
c
insulated parts around the Grottoes to approximately
3°C. In winter, this is much warmer than outside tem¬
peratures. The cave then functions as a chimney, with
the warmer central cave air flowing upstream into the
sole of the Icefield and dense, cold air pouring in
through the explorers' entrance to replace it. The cold,
dry air freezes residual pools of summer floodwater in
the Entrance Complex, giving cavers a 300-m belly
crawl over dusty ice. The first liquid water is encoun¬
tered only 1 km inside the cave. There is a subsidiary
daily effect because the draft is strongest and coldest
immediately before dawn, not a good time to be com¬
ing out through the ice crawls. In summer, the
dynamic situation is reversed as cold, moist air from
the interior flows down and out through the explorers'
entrance into the warm exterior. Thick hoarfrost is
deposited onto the chilled entrance walls until the first
floods of summer arrive to remove it.
An unexpected feature of the 2009 and 2010 dives
was that the temperature of the perched water body in
Boon's Blunder was gauged at 5°C, whereas air and
seepage water temperatures in the Central Cave a few
hundred meters away at the same elevation do not rise
above 3.0 to 3.5°C; the heart of the mountain is chilled
below its geothermal balance by the winter airflow here.
THE FAUNA OF CASTFEGUARD
Harsh as conditions are, the cave has a small popula¬
tion of animals that prefer it or are entirely dependent on
its protection. The first, troglophiles, are packrats nesting
above the floodwater limits in the Entrance zone. They
have their own private entrances from the Meadows
overhead that are too small for human cavers. Two spe¬
cies of blind and unpigmented (troglobitic or fully
adapted) crustaceans live in pools in the Central Cave,
where the water apparently never freezes. An isopod,
Salmasellus steganothrix, is known elsewhere in the
Canadian Rockies. An amphipod, Stygobromus canaden¬
sis, is known only in Castleguard. It is possible that the
cave served as a subglacial refuge for these species dur¬
ing the last glaciation or longer. A packrat nest and other
organic material were found in the Headward Complex
in 1983 and carbon-dated to ~9000 years BP, indicating
that in early postglacial times the Icefield had receded
from at least part of it, thus providing some food for the
troglobites downstream.
CONCFUSIONS
Castleguard Cave has proved to be a fine laboratory
for the study of dissolutional cavern genesis and of
groundwater flow under glacier cover. It appears that
most passages in the Cave that are large enough for
humans have now been found and mapped, except
perhaps in the Headward area and the new discover¬
ies beyond Boon's Blunder. The underlying, very
dynamic complex of Castleguard II that drains much
of the Columbia Icefield today continues to defy all
attempts at physical exploration. Under the Meadows
and around the summit massif of the mountain more
than 100 mapped sinkholes plus unmeasured subgla¬
cial streams feed water into Castleguard III, another
almost entirely unknown system that may contain
both extensive fossil and active galleries. There is great
potential here for future generations of speleologists.
Bibliography
Atkinson, T. C. (1983). Growth mechanisms of speleothems in
Castleguard Cave, Columbia Icefields, Alberta, Canada. Arctic
and Alpine Research, 15(4), 523—536.
Ford, D. C. (Ed.), (1983). Castleguard Cave and Karst, Columbia
Icefields Area, Rocky Mountains of Canada: A symposium. Arctic
and Alpine Research, 15(4), 425—554.
Ford, D. C., Lauritzen, S.-E., & Worthington, S. (2000). Speleogenesis
of Castleguard Cave, Rocky Mountains, Alberta, Canada. In A.
Klimchouk, D. C. Ford, A. N. Palmer, & W. Dreybrodt (Eds.),
Speleogenesis: Evolution of karst aquifers (pp. 332—337). Huntsville,
AL: National Speleological Society.
Muir, R. D., & Ford, D. C. (1985). Castleguard. Gatineau, Quebec:
Parks Canada Centennial Publication, Department of Supply and
Services.
Smart, C. C. (1983). The hydrology of the Castleguard Karst,
Columbia Icefields, Alberta, Canada. Arctic and Alpine Research,
15(4), 471-486.
CAVE DWELLERS IN THE
MIDDLE EAST
Paul Qoldberg and Ofer Bar-Yosef
Boston University and Harvard University
This article offers information about prehistoric caves
in the Middle East. Caves are a source of fascination for
the public as well as many scientists. In this short entry
we will attempt to briefly describe and summarize
some of the discoveries in Middle Eastern caves and
what those discoveries reveal about human behavior.
INTRODUCTION
Caves figure strongly in the archaeological record,
through most archaeological time periods and across
most continents. Caves not only served as loci of habi¬
tation where daily activities such as food processing
and sleeping took place but also functioned as gather¬
ing places for spiritual and religious activities.
Prehistoric caves—particularly those in the Middle
East—are special sedimentary environments (Fig. 1). In
Encyclopedia of Caves.
© 2012 Elsevier Inc. All rights reserved.
CAVE DWELLERS IN THE MIDDLE EAST
95
FIGURE 1 Map of Middle East and the location of major prehistoric cave sites in the area. Only a few caves are discussed here.
ENCYCLOPEDIA OF CAVES
96
c
essence, because of the overall low energies of the depo-
sitional processes, caves serve as excellent sedimentary
traps: whatever is brought into the cave tends to stay
there. As a result, caves can preserve faithful records of
past environments, as well as past human activities.
Environmental information is conveyed by the presence
of macro- and microfaunal remains—particularly the
latter. Plant remains, an additional environmental indi¬
cator, are scarce, although microbotanical remains such
as pollen and phytoliths can be found. Phytoliths tend
to be better preserved and abundant and provide
insights not only about past local and regional condi¬
tions but also about human activities, such as the gath¬
ering of plants for fuel, bedding, and food.
In addition, anthropogenic deposits also reveal the
nature of past human activities at a cave by incorporat¬
ing animal bones and stone artifacts. These objects
inform us about hunting and butchering practices, the
manufacture and use of curated stone tools, and
the use and function of space ( e.g ., working areas, fire¬
places, dumping, and sleeping zones).
The above statements are pertinent to many cave sites
throughout the world. Middle Eastern caves, however,
benefit from their location at the crossroads to human
past and present migrations and thus offer the opportu¬
nity to monitor and document important phases in
human evolution, particularly with regard to the origin
of modern humans and the demise of Neanderthals.
LOCATION AND PRESENT
ENVIRONMENT
Most of the caves in the Middle East are situated
either in the Mediterranean climatic zone (Fig. 1), par¬
ticularly in a belt close to the present Mediterranean
coastline, or in the drier, steppic and desertic region.
The Mediterranean climate, which includes distinct cli¬
matic gradients from west to east and from north to
south, is characterized by warm, dry summers and
cool, wet winters in which 500 to more than 1000 mm
of precipitation can fall. Similarly, the transition from
steppe to desert is marked by reductions in rainfall
from 500 to less than 100 mm.
Most of the caves outlined in this section developed
in limestone and formed during the late Tertiary/
Quaternary period under phreatic conditions and com¬
monly as enlargements along joints. Vadose expansion
is more recent, resulting in a vaulted or domed mor¬
phology with vertical chimneys that extend to the sur¬
face above several of the caves. Many of these same
caves have depressions or sinkholes that are situated
beneath the chimneys. In any case, modern-day karstic
activity is generally negligible, and the remains of for¬
merly more extensive and well-developed dripstone
features signify karstic conditions that were much
different from what occurs today.
CAVE DEPOSITS AND PROCESSES
Depositional and post-depositional processes oper¬
ating in karstic caves are relatively well known but
less so in prehistoric caves; however, recent research,
particularly in Middle Eastern prehistoric caves, has
begun to reveal a high degree of marked complexity.
Material of geological origin may accumulate in the
interior in a number of ways, including (1) gravity-
derived rockfall from the walls and roof of the cave;
(2) aqueous processes associated with fluvial and phre¬
atic deposition, or runoff; (3) colluviation of soils
derived from the surfaces above and outside the cave;
and (4) Aeolian deposition of sand and silt.
Middle Eastern caves contain distinct biogenic contri¬
butions, such as bird and bat guano. In addition, copro-
lites and organic matter remains produced by
carnivores—particularly hyenas—are quite common. In
many caves, the presence of millimeter-sized rounded
stones indicates distinct gastrolith input by pigeons. Plant
remains, such as grass and wood, can be washed and
blown into the cave, or alternatively brought into the cave
by past human occupants for fuel, bedding, or shelter.
Anthropogenic contributions to cave sedimentation
in this region are noteworthy and tend to be unnoticed
in many prehistoric cave settings throughout the world.
These accumulations consist of primarily bone and shell
remains, as well as the buildup of ashes, organic matter,
and charcoal that are associated with fireplaces and
burning activities. Additional traces of fine-grained soil
and sediment that were tracked into the cave by
humans and other animals are subtle but important.
Once within the cave itself, deposited material is
commonly modified by a number of processes that are
often penecontemporaneous with deposition. Deposits
are sometimes modified by wind, runoff, and dripping
water or subjected to burrowing by animals or to tram¬
pling by humans and other occupants. Mineral and
organic residues are sometimes moved by human or
animal activity from their original location within the
cave. Bone and lithic discard evidence of activities and
indications of removal, dumping, and trampling of
material associated with the cleaning or maintenance
of hearths are not uncommon.
Prehistoric caves in this region are damp and act as
sinks for water, organic matter, and guano and accord¬
ingly tend to act as chemical engines whereby numer¬
ous mineral transformations take place (Karkanas et al.,
2000; Schiegl et al., 1994, and Schiegl et al., 1996).
These secondary alterations commonly include the pre¬
cipitation (and dissolution) of carbonates that form
ENCYCLOPEDIA OF CAVES
CAVE DWELLERS IN THE MIDDLE EAST
97
FIGURE 2A Tabun Cave, Mount Carmel. Shown in this photo¬
graph (taken during the 1969 season) are bedded sandy and silty
sediments that were blown into the cave from the adjacent coastal
plain. These contain Lower and Middle Paleolithic industries, includ¬
ing the Upper Acheulian (UA), Acheulo-Yabrudian (AY), and
Mousterian (M). Note the strong dip of the lowermost sediments that
plunge into a subsurface swallow hole. This subsidence took place
during the earlier phases of deposition, as the bulk of the sediments
from the middle part of the photograph upward are roughly horizon¬
tal. The uppermost part of the photograph (see Fig. 2B) is composed
of interbedded clay and ashy deposits, punctuated with numerous
blocks of roof fall. The dating of the deposits here is a subject of
debate, but estimates using thermoluminescence (TL) dating on
burned flints suggest that the deposits down to the top of the large
hole on the left are close to 350,000 years old (Mercier et al., 1995).
speleothems, layered travertines, or the so-called cave
breccia, which represent calcite-cemented clastic sedi¬
ments typically consisting of inwashed soil material.
Another characteristic type of mineral alteration
involves the formation of several types of phosphate
minerals, along with the formation of opal and the
breakdown of clays.
Paleoenvironmental and paleoclimatic changes ex¬
pressed in the faunal records, as well as by depositional
and post-depositional accumulations and removals, are
registered in a rather coarse chronology. Speleothems in
FIGURE 2B Tabun Cave, Mount Carmel. This photograph is from
the upper part of the cave, above the uppermost ledge in Fig. 2A. In
contrast to the lower deposits, these are largely anthropogenic in
nature and consist of lighter and darker bands of ash and red clay,
respectively. These anthropogenic layers are overlain by massive red¬
dish clay that contains decimeter-sized blocks of roof fall. The latter
attests to an enlargement and eventual opening of the chimney that
leads up to the surface of Mount Carmel. It is through this chimney
that the clayey terra rosa soils were washed into the cave.
Nahal Soreq Cave and a cave near Jerusalem provided
detailed records of paleoclimatic fluctuations of the past
170,000 years. These karstic caves were rarely penetrated
by humans, and these oscillations follow the pattern
known from other localities of the Northern Hemisphere.
THE PREHISTORIC AND HISTORICAL
SEQUENCE OF CAVE OCCUPATIONS
IN THE MIDDLE EAST
Prehistoric caves in the region produced the basic
archaeological sequence of at least the past 400,000 years
(Fig. 1); several examples are illustrated in the following
text, with comments related to the themes discussed
above. The oldest known cave deposits were exposed in
Umm Qatafa, Tabun (Fig. 2A,B), and the Yabrud IV rock-
shelter, where the lowermost layers—more than 400,000
years old—contained a core-and-flake industry, a phe¬
nomenon that intersects the Acheulian sequence. The
Late (or Upper) Acheulian is known from several caves.
This industry is characterized by hand axes or bifaces,
but these tools are not quantitatively the most dominant
tool type. The Late (or Upper) Acheulian is followed by
the Acheulo-Yabrudian, where the combination of
bifaces and scrapers, often shaped on thick flakes,
appear in Tabun, Hayonim, Qesem, Zuttiyeh, and the
Yabrud I rockshelter, as well as in open-air sites in the
El-Kowm basin (northeast Syria). This entity occupies
the Levant between the Acheulian sequence and the
ENCYCLOPEDIA OF CAVES
98
c
FIGURE 3 Qafzeh Cave, Lower Galilee. This cave is famous for
its series of Middle Paleolithic human burials in layers XVII to XXI.
The morphological analyses identified the remains as archaic modern
humans and are therefore considered as coming out of Africa. These
layers, dated by thermoluminescence (TL) from about 95,000 to
114,000 calendar years BP (Valladas et a]., 1988), consist of fine¬
grained angular rock fall that has been reworked by slopewash and
colluvial processes. The darker band in layer XXI results from man¬
ganese enrichment associated with a subsurface spring that was
operational during that time.
Mousterian or Middle Paleolithic. Its geographic distri¬
bution indicates an origin in the northern Levant.
Cave sites with Middle Paleolithic (250,000—270,000
to 48,000—46,000 calendar years BP) remains have pro¬
duced a wealth of evidence, as well as a large number
of human burials and isolated human bones. The earlier
deposits contain skeletal remains identified as archaic
modern humans (also known as the Skhul-Qfazeh
group; Fig. 3). Occurring stratigraphically above these
human remains are those of southwestern Asia
Neanderthals that were found in Kebara (Fig. 4),
Amud, Dederiyeh, and Shanidar. Based on the evi¬
dence, both human populations demonstrated good
hunting skills, the use of fire, and the ability to procure
raw materials for making stone tools from a radius of 5
to 20 km around the sites. Mobility between lowland
and highland areas has been easier to trace in Lebanon,
south Jordan, and the Zagros because of the greater
topographic relief in these areas.
Among Upper Paleolithic and Epi-Paleolithic (48,000—
46,000 to 11,500 calendar years BP) cave occupations,
the best known are (1) Ksar 'Akil, with an unusual 18-m
sequence of Upper Paleolithic layers; (2) the few
Levantine Aurignacian sites (e.g., Yabrud II, El-Wad,
Kebara, and Ufagizili); and (3) those attributed to the
Natufian culture that immediately preceded the earliest
farming villages. The Natufians camped in caves, built
rooms inside the main chambers, and buried their
dead. Their use of caves was probably more intense
than that of their predecessors. During the prehistoric
periods, caves were often used for camping by entire
FIGURE 4 Kebara Cave, Mount Carmel. Situated about 12 km
south of Tabun, Kebara is considerably younger, dating to about
60,000 to 70,000 calendar years BP. This view of the southeast corner
of the cave shows mostly Upper Paleolithic (UP) deposits in the
walls on the right and left and Middle Paleolithic (MP) sediments on
the floor of the excavation. The Upper Paleolithic deposits are com¬
prised of finely bedded and laminated silt and sand-sized material
(S) that has washed back into the cave via runoff. It commonly
includes sand-sized aggregates that have been reworked from older
sediments near the entrance. A large piece of limestone roof fall (LS)
in the wall to the left shows that some gravity deposition also takes
place. Interestingly, the other portions of this block have been sub¬
jected to diagenesis in which the original dolomite block has been
transformed into a number of phosphate minerals, including franco-
lite, crandallite, montgomeryite, and leucophosphite (Weiner et ah,
1993). Punctuating these geogenic deposits are isolated hearths (H)
representing some anthropogenic material within the mostly geo¬
genic sediments. The Middle Paleolithic deposits, on the other hand,
are mostly anthropogenic and consist of hearths and hearth products
(charcoal, organic matter, and ash). Some of these hearths are
revealed by a white area (see the base of the photograph).
groups and rarely used as stations for performing spe¬
cific tasks. One of the best examples is the use of caves
at higher altitudes for short-term camps by hunters.
Since the Neolithic period (11,500 to 7500 calendar
years BP) the use of caves changed dramatically.
Several caves served as camps for special artisans, but
in more than one case the caves became sacred and
were utilized, as in Nahal Hemar and Nahal Qana
caves, for storage of paraphernalia or other sacred
objects. Larger caves could have served as locations for
ceremonies, as indicated by a few caves in Turkey.
During the Chalcolithic and Bronze ages (7500 to 3200
calendar years BP), caves were employed for various
purposes. Certain karstic caves, such as Peqi'in, served
as burial grounds, while others functioned as storage
facilities, animal pens, and even refugia. During the
Bronze Age, caves continued to be used in similar ways,
and a unique example is the warrior burial in a cave near
the Jordan Valley. Finally, a common use of most caves
during the last millennium BC and the first two millen¬
nia AD was by shepherds who often spent the late fall
and winter months in these protected shelters.
ENCYCLOPEDIA OF CAVES
CAVE ECOSYSTEMS
99
See Also the Following Articles
Ancient Cavers in Eastern North America
Maya Caves
Bibliography
Ayalon, A., Bar-Matthews, M., & Kaufman, A. (2002). Climatic condi¬
tions during marine oxygen isotope stage in the eastern
Mediterranean region from the isotopic composition of spe-
leothems of Soreq Cave, Israel. Geology, 30(4), 303—306.
Bar-Matthews, M., Ayalon, A., & Kaufman, A. (1997). Late
Quaternary paleoclimate in the eastern Mediterranean region
from stable isotope analysis of speleothems at Soreq Cave, Israel.
Quaternary Research, 47, 155—168.
Bar-Matthews, M., Ayalon, A., Kaufman, A., & Wasserburg, G. J.
(1999). The eastern Mediterranean paleoclimate as a reflection of
regional events: Soreq Cave, Israel. Earth and Planetary Science
Letters, 166, 85—95.
Frumkin, A., Ford, D. C., & Schwarcz, H. P. (1999). Continental oxy¬
gen isotopic record of the last 170,000 years in Jerusalem.
Quaternary Research, 51, 317—327.
Karkanas, P., Bar-Yosef, O., Goldberg, P., & Weiner, S. (2000).
Diagenesis in prehistoric caves: The use of minerals that form in
situ to assess the completeness of the archaeological record.
Journal of Archaeological Science, 27, 915—929.
Karkanas, P., Goldberg, P., 2010. Phosphatic Features. In: G. Stoops,
V. Marcelino, F. Mees (Eds.), Interpretation of Micromorphological
Features of Soils and Regoliths. Elesevier, Amsterdam, pp. 522—542.
Mercier, N., Valladas., H., Valladas, G., & Reyss, J.-L. (1995). TL dates
of burnt flints from Jelinek's excavations at Tabun and their
implications. Journal of Archaeological Science, 22, 495—509.
Schiegl, S., Lev-Yadun, S., Bar-Yosef, O., Goresy, A. E., & Weiner, S.
(1994). Siliceous aggregates from prehistoric wood ash: A major
component of sediments in Kebara and Hayonim caves (Israel).
Israel Journal of Earth Science, 43, 267—278.
Schiegl, S., Goldberg, P., Bar-Yosef, O., & Weiner, S. (1996). Ash
deposits in Hayonim and Kebara Caves, Israel: Macroscopic,
microscopic and mineralogical observations, and their archaeo¬
logical implications. Journal of Archaeological Science, 23, 763—781.
Valladas, H., Reyss, J. L., Joron, J. L., Valladas, G., Bar-Yosef, O., &
Vandermeersch, B. (1988). Thermoluminescence dating of
Mousterian "Proto-Cro-Magnon" remains from Israel and the ori¬
gin of modern man. Nature, 331(6157), 614—616.
Weiner, S., Goldberg, P., & Bar-Yosef, O. (1993). Bone preservation in
Kebara Cave, Israel, using on-site Fourier transform infrared
spectrometry. Journal of Archaeological Science, 20, 613—627.
CAVE ECOSYSTEMS
Kevin S. Simon
The University of Auckland, New Zealand
DEFINITION AND BOUNDARIES
Caves are unique among ecosystem types in their
physical structure and biological function. Embedded
in rock and lacking light, the geological setting defines
the bounds of cave ecosystems and dictates how
energy and matter move through them. As humans,
we typically think of the bounds of cave ecosystems as
the walls of the passages that we enter and explore,
ranging from dimly lit cave entrances to deep, dark
passages. As it turns out, the cave environment that
humans experience is really just one habitat within a
broader ecosystem; we simply see only the portion of
the ecosystem we fit into. Beyond those passages are
voids, some quite small and some filled with water,
which are interconnected and through which move
organisms, energy, and matter. If the cave ecosystem
extends beyond the cave walls, where then do we
draw the line that bounds the ecosystem? The first real
attempt at defining the bounds of a cave ecosystem
was made by French biologist, Raymond Rouch. In his
research at the Baget Basin (summarized in Rouch,
1986) he realized that when he collected animals in
caves for study, he was really only sampling a small
portion of a bigger system. By sampling at sinking
streams and at springs that drained different portions
of the aquifer, he found a much greater abundance
and diversity of animals than he observed from sam¬
pling only in caves. He realized that the true bound¬
aries of the ecosystem extended well beyond caves, to
where water first entered the subsurface to where it
returns to the surface at springs. Rouch argued that
the true ecosystem was not really a "cave" ecosystem
at all; rather it was a "karst" ecosystem, that is, the
entire drainage basin and the highly interconnected
and diverse habitats within it (Rouch, 1977). The edges
of the ecosystem are not cave walls; rather they are the
hydrologic boundaries that determine where water
flows on and under the landscape. As it turns out, the
same physical and chemical force, flowing water, that
creates caves also defines the boundaries of the ecosys¬
tem, the various habitats within the ecosystem, and
how energy and matter move through it.
PHYSICAL ENVIRONMENT AND
HABITAT ZONES
Rainwater that falls on the surface of a karst land¬
scape follows gravity, percolating through soils and
into a network of small fractures just below the soil
surface or sinking rapidly through relatively large
openings as sinking streams (Fig. 1). That water flows
through cave streams to the water table and ultimately
it returns to the surface at springs. The entire land area
and aquifer volume that drain water to the point of
resurgence are the drainage basin that comprises the
karst ecosystem. Within that basin are definable habi¬
tats that differ in physical characteristics and biological
communities. This zone of contact between soils and
bedrock, the epikarst, is the interface zone between the
Encyclopedia of Caves.
© 2012 Elsevier Inc. All rights reserved.
100
FIGURE 1 Diagram of energy flux among karst ecosystem zones.
DOM and POM are dissolved and particulate organic matter,
respectively.
surface and caves. It is a unique environment com¬
prised of a heterogeneous network of small fissures
that are intermittently wetted, with some spaces per¬
manently wetted while others are sometimes or always
dry. Water that percolates through the epikarst some¬
times drips into larger cave passages and into small
pools and cave streams. These streams coalesce with
other streams that are fed by larger conduits which are
directly connected to the surface through relatively
large openings, visible as sinking streams at cave
entrances. While epikarst and conduits are both inter¬
faces, or ecotones, with the surface, they differ dramat¬
ically in physical structure, biological communities,
and capacity to deliver energy and matter to the
subsurface. The small, heterogeneous fractures in the
epikarst result in relatively slow water flow rates and
permit only very small particles or dissolved material
to pass from surface to subsurface. Further, the epi¬
karst represents a large area of interface within the
basin. In contrast, the much larger conduit openings,
such as cave entrances, can support much higher and
more variable flow and pass larger particles and
organisms from the surface to the subsurface. While
the conduits are larger, they are generally restricted to
a few specific focal points within the basin.
Water derived from epikarst and conduit flow paths
pass through streams which are characterized by the
same types of physical structures (riffles and pools)
seen in surface streams. Depending on the nature of
connection to the surface, cave streams can differ dra¬
matically in temperature, flow, and chemistry, with
conduit-fed streams generally much more variable in
all of these aspects. These streams and passages below
the epikarst, but above the water table, comprise the
unsaturated (or vadose) zone of the ecosystem. The
unsaturated zone contains both aquatic and terrestrial
habitats, and it is the component of ecosystems that
most humans are familiar with as caves. Eventually,
cave passages and fractures in the rock intersect the
water table, they become completely filled, and the
system is comprised exclusively of aquatic habitat.
This habitat is generally only accessible via diving or
through wells driven below the water table. By nature
of the amount of time the water has spent under¬
ground, this zone of the aquifer is more stable in terms
of temperature and chemistry than the epikarst or
unsaturated zone. Ultimately, water from the saturated
zone resurges at springs, which are the downstream
interface with the adjacent surface ecosystem.
The terrestrial component of cave ecosystems fol¬
lows the same fracture system as water, and indeed
represents conduits that are currently or historically
channels for water flow. It is, however, restricted to
the unsaturated portion of the aquifer and has only
been studied in areas directly accessible by humans.
Howarth (1983) categorized the terrestrial environment
of caves into five zones: (1) entrance, (2) twilight,
(3) transition, (4) deep, and (5) stagnant air. These
zones are in fact a gradient from a surface in which
light declines and temperature and humidity become
less variable. This framework applies only to the larger
conduit openings, but there is likely a terrestrial habi¬
tat within the epikarst, but it is virtually impossible to
access and characterize. The terrestrial and aquatic
habitats within karst ecosystems should not be viewed
as separate and they are likely extremely connected, as
is the case for surface ecosystems. For example,
organic matter carried into caves by streams is depos¬
ited into the terrestrial habitat by dynamic overflood¬
ing and drying cycles. Animals and organic matter
from terrestrial zones can in turn be deposited into
streams. However, the nature of this connection has
not been well studied.
ENERGY FLUX AND LIMITATION
Darkness is probably the most defining feature of
cave ecosystems, and it has perhaps had the greatest
influence on organisms in caves. The effect of no light is
manifest through energy flux in food webs. In nearly all
ecosystems, light-driven photosynthetic production by
plants is the key source of energy that fuels all other liv¬
ing organisms in the ecosystem. How can food webs
persist in caves where there is no light? There are essen¬
tially two options. The first is to fix energy derived
from something other than light into organic molecules.
When the chemical conditions are right, microbes can
use an array of metabolic strategies that extract energy
from chemical sources to create microbial biomass
ENCYCLOPEDIA OF CAVES
CAVE ECOSYSTEMS
101
(Engel, 2007). This chemical-driven microbial produc¬
tion, chemoautotrophy, can support diverse and fasci¬
nating food webs in caves (Sarbu et al., 1996). It turns
out though, that most caves that humans can access
lack the unique chemistry needed to support substan¬
tial chemoautotrophic production. There are, however,
some large aquifers such as the Edward Aquifer in
Texas, United States, that appear to have very large,
deep systems driven by chemoautotrophy.
In the absence of light and chemoautotrophy, the
only other option for sustaining a food web is dead
organic matter from plants, or detritus, which is
imported from the surface. This detritus ranges from
dissolved compounds to large particles such as leaves
and wood (Simon et al., 2007). That detritus is deliv¬
ered to the subsurface through two key vectors: flow¬
ing water and organisms. As water flows from the
surface, it carries with it detritus that is either dis¬
solved or entrained in the water. In the case of water
percolating through soils and into the epikarst, only
dissolved organic matter and very fine particles that
pass through soils and fractures in the rock are trans¬
mitted. That material can be used by microbes which
are in turn consumed by small animals, notably cope-
pods, which can achieve very diverse communities in
the epikarst. In the case of larger conduits connected to
the surface, water can carry not only dissolved detri¬
tus, but also much larger particles ranging from leaves
to logs. That material is carried through caves, and
deposited in streams or on stream banks, especially
during floods. Thus, the nature of the physical struc¬
ture of the aquifer dictates what type and how much
energy is made available to various habitats in caves.
Animals and plants can also deliver organic matter to
caves. Well-known examples include bats and crickets,
which forage on the surface and return to the subsur¬
face where their feces, eggs, and bodies are consumed
by microbes and animals in caves. Indeed, entire spe¬
cialized communities (guanobionts) rely on bat and
bird feces. This is well demonstrated for terrestrial
habitats, but less clear in aquatic habitats in caves. In
other cases, those animals that exit caves and forage
on the surface are themselves energy within caves. For
example, cave cricket eggs are preyed on by special¬
ized cave beetles. In a few cases, plant roots can pene¬
trate down into cave passages where material from
those roots can be consumed by cave animals.
If one French scientist, Rouch, was the pioneer of
defining the ecosystem, it was another French ecolo¬
gist, Janine Gibert, who pioneered the study of how
energy and matter move through karst ecosystems.
Gibert did this by characterizing the amount and type
of organic matter moving from overlying soils, out of
the epikarst and through the vadose zone in cave
streams, and ultimately out of the saturated zone via
a spring at the Dorvan-Cleyzieu basin (Gibert, 1986).
Among Gibert's most important findings were that
organic matter dissolved in water was a much larger
flux than was the flux of organic particles such as
pieces of leaves and wood and the bodies of surface
animals. Those fluxes were extremely variable over
time and tightly linked to patterns of high and low
flow as driven by rainfall. Finally, she noted that
microbes were likely to be key players in mediating
the transfer of energy from detritus to animals in
karst.
More recent research at Organ Cave in the United
States has confirmed that bacterial productivity fuels
cave stream food webs (Simon et al., 2003). In Organ
Cave, some streams are fed only by epikarst water
while others receive water from sinking streams and
shafts. Leaves placed in the streams did not move far
before being retained and eventually decomposed.
Leaf decomposition was rapid only in those streams
that contained a particular stygophilic amphipod
which shreds and consumes leaves readily. In the
other streams with only stygobitic animals, leaves
were not readily consumed and were only slowly
degraded by microbes in streams where particulate
organic matter was scarce. A series of experiments
subsequently showed that carbon from bacterial pro¬
duction was used by all animals as a food source in
both types of streams. Interestingly, it appears that dis¬
solved organic matter, especially that derived from
soils, rather than large particles such as leaves is the
main energy source for food webs even when leaves
and wood are available. While some cave animals
can use the larger particles directly as food, many
seem to rely almost exclusively on microbes that
use dissolved organic compounds. The relationship
between microbes and animals is not unidirectional,
as the animals also strongly influence the productivity
of the microbes that they consume (Cooney and
Simon, 2009). These issues of patterns of energy trans¬
port and use are important as energy scarcity has been
considered the key driving force in the evolution of
cave organisms. For example, there is a great deal of
indirect evidence such as reduced metabolic rate,
larger but fewer eggs, and increased longevity of sub¬
terranean animals that is consistent with the hypothe¬
sis of energy limitation of karst ecosystems.
NUTRIENTS
Availability of detritus, or carbon, may not be the
only factor that limits energy availability in karst eco¬
systems. Karst ecosystems based on chemoautotrophy
are clearly limited by the availability of inorganic ele¬
ments (especially forms of iron, nitrogen, and sulfur;
ENCYCLOPEDIA OF CAVES
102
c
see Engel, 2007) other than carbon. The physical factors
that dictate the dynamics of these elements have
strong influence on the magnitude and location of
the energy base of food webs in those systems, with
interface zones such as the water—air interfaces in
Movile Cave (Sarbu et ah, 1996) acting as hotspots for
productivity in karst ecosystems. The importance of
inorganic nutrients in limiting productivity in other
karst ecosystems is possible as microbes that use detri¬
tus can be limited by nutrients such as nitrogen and
phosphorus. This issue is almost completely unstudied
in karst ecosystems and the few data available are
contradictory.
CONNECTIVITY AND DISTURBANCE IN
KARST ECOSYSTEMS
Caves present an interesting dichotomy in terms of
stability and disturbance that is determined in large
part by the nature of hydrologic connectivity with the
surface. On one hand, caves seem like rather stable,
benign places. They are always dark and generally
have stable temperature and humidity. On the other
hand, caves can be spectacularly harsh places to live.
Food is scarce and, during floods, some cave streams
become raging torrents. The reality is that, depending
on place and time, caves can vary dramatically in sta¬
bility and harshness. This is dictated in large part by
the nature of connectivity to the surface. Streams and
pools fed by epikarst water are much more stable in
flow, temperature, and chemistry than are streams
connected to the surface by large openings. The geo¬
logical framework of the aquifer and the climate the
system is located in also strongly influence the nature
of disturbance and stability in caves. High flow during
storms, in particular, is a key disturbance in karst sys¬
tems as it refills voids, inundates terrestrial habitat,
increases delivery of energy and matter from the sur¬
face and scours microbial films. Ironically, while it was
recognized that floods are likely to be quite important
for cave animals (Hawes, 1939), few studies have sys¬
tematically examined how floods influence the animal
and microbial communities in caves.
Disturbance by humans is now a component of
most karst ecosystems. This disturbance can take the
form of pollution by chemicals and sediments that
have had detrimental effects on karst ecosystems. Of
particular importance is the role humans play in
changing energy availability in caves via pollution. In
particular, humans almost invariably increase the
input of organic matter and nutrients to caves. This
comes from a diverse array of sources ranging from
septic system waste to bits of cotton clothing left by
cavers. Often the organic matter humans introduce is
readily used by microbes and it can indirectly change
the physical environment by reducing oxygen avail¬
ability and smothering habitat. One common outcome
of this perturbation is a shift in abundance of cave ani¬
mals such that a few tolerant taxa proliferate and
become abundant while other more sensitive taxa are
reduced in abundance or extirpated. A second com¬
mon outcome is the increased abundance of surface
animals in caves at the expense of cave-adapted taxa.
It appears that the increased energy availability lessens
the competitive advantage that cave animals have in
low productivity environments. In some cases humans
can reduce energy input to caves by severing connec¬
tivity to the surface. Examples include human-induced
loss of bats and alteration of surface vegetation, soils,
and hydrology in ways that cut off organic matter
input to the subsurface. Finally, human activity can
change the magnitude and pattern of water flow, the
lifeblood of most karst ecosystems, by altering surface
vegetation, soils, and streams or through extraction of
water from within aquifers.
See Also the Following Articles
Food Sources
Chemoautotrophy
Bibliography
Cooney, T., & Simon, K. S. (2009). Influence of dissolved organic
matter and invertebrates on the function of microbial films in
groundwater. Microbial Ecology, 58(3), 599—610.
Engel, A. S. (2007). Observations on the biodiversity of sulfidic karst
habitats. Journal of Cave and Karst Studies, 69(1), 187—206.
Gibert, J. (1986). Ecologie d'un systeme karstique jurassien.
Hydrogeologie, derive animale, transits de matieres, dynamique
de la population de Niphargus (Crustace Amphipode). Memoires
de Biospeologie, 13(40), 1—379 (in French).
Hawes, R. S. (1939). The flood factor in the ecology of caves. Journal
of Animal Ecology, 8, 1—5.
Howarth, F. G. (1983). Ecology of cave arthropods. Annual Review of
Entomology, 28, 365—389.
Rouch, R. (1977). Considerations sur l'ecosystem karstique. Compte
Rendus Academie des Sciences de Paris Serie D, 284, 1101 — 1103 (in
French).
Rouch, R. (1986). Sur l'ecologie des eaux souterraines dans la karst.
Stygologia, 2(4), 352-398 (in French).
Sarbu, S. M., Kane, T. C., & Kinkel, B. F. (1996). A chemoautotrophi-
cally based cave ecosystem. Science, 272(5270), 1953—1955.
Simon, K. S., Benfield, E. F., & Macko, S. A. (2003). Food web struc¬
ture and the role of epilithic films in cave streams. Ecology, 84(9),
2395-2406.
Simon, K. S., Pipan, T., & Culver, D. C. (2007). A conceptual model
of the flow and distribution of organic carbon in caves. Journal of
Cave and Karst Studies, 69, 279—284.
ENCYCLOPEDIA OF CAVES
CAVE, DEFINITION OF
103
CAVE, DEFINITION OF
William B. White and David C. Culver'
The Pennsylvania State University, t American University
DEFINITIONS AND POINTS OF VIEW
Humankind has been involved with caves for millennia.
Caves were places of shelter for early humans in many
parts of the world. They have served as places of wor¬
ship for many societies in many times and places.
Caves have been used as storehouses, as munitions fac¬
tories, and as resting places for the dead. Caves play a
prominent role in the myths and legends of many cul¬
tures throughout recorded history. Caves are secret
places. Small children make "caves" by draping blan¬
kets over furniture. In contemporary society, caves fre¬
quently appear in movies and in cartoons. Show caves
continue to draw thousands of visitors each year. Caves
are also of great interest to scientists and explorers.
Because caves are voids in rock, they are considered
geological features, and indeed many textbooks ( e.g .,
Ford and Williams, 1989) firmly defend this point of
view. However, caves are more than their geology
because of their interaction with people and with organ¬
isms. One textbook (White, 1988) recognizes their
human appeal by defining caves as "a natural opening
in the Earth, large enough to admit a human being, and
which some human beings choose to call a cave."
CAVES AS PLACES FOR EXPLORATION
Caves are enticing, awaking an interest in many to see
"where it goes." Although the first cave explorers are
lost in the mists of history, cave exploration as a special¬
ized human activity dates from the middle of the nine¬
teenth century (Shaw, 1992). Earlier authors include
discussions of caves in general travelogues and descrip¬
tions of regions. The most famous of these is Valvasor's
1656 book. The Glories of the Duchy of Carniola (an area
that is part of present-day Slovenia). In North America,
Horace C. Hovey, Luella Owen, and a few others wrote
popular accounts of their expeditions. The first modern
speleologist is said to be Adolf Schmidl, who explored
many Slovenian caves in the mid-nineteenth century.
Somewhat later came Edouard A. Martel in France who
is considered the father of cave exploration. Unlike his
American counterparts, Martel organized serious caving
expeditions into the large caves and deep pits of the
Pyrenees, the Alps, and many other places in Europe, an
activity that also required inventing the technology of
cave exploration as he went along.
Cave exploration by organized caving societies was
well under way in Europe in the early twentieth cen¬
tury. Organized exploration came later in the United
States when several caving groups formed the
National Speleological Society in 1941. During the past
50 years, cave exploration has blossomed into a recrea¬
tional activity for thousands of individuals throughout
the world and is organized into hundreds of caving
clubs, national societies, and specialized scientific orga¬
nizations. Most cavers take their explorations seriously
and spend substantial time in preparing maps and
writing detailed reports. Exploration and surveying of
caves are among the few remaining activities where
useful contributions to knowledge can be made by
nonprofessionals.
Much of the allure of caves resides in their remote¬
ness and wilderness character even beneath urban
sprawl. This is, in part, because the underground land¬
scape with its total darkness and unusual shapes of
rock and mineral deposits is so alien compared with
the familiar surface landscape. Caves are remote in the
sense of the time and effort required to explore them.
The farthest reaches of a large cave system may be
only a few kilometers from the entrance as the crow
flies and may be no more than 10 or 15 km as the caver
crawls. And, of course, the outside world is only a few
tens or hundreds of meters away, vertically, through
solid rock. However, reaching the farthest corners of a
large cave system, doing a bit of exploring and survey¬
ing, and returning to the entrance may require 24 to
36 hours. Or, it may require several days and an
underground camp. In the same time one could have
traveled comfortably across the continent on a jet
plane, attended a conference on the opposite coast,
returned home, and been clean and well rested to
boot. The remoteness of the cave arises not from dis¬
tance but from the time needed to traverse it, from the
obstacles that must be overcome, and from the sense
of the strange and the bizarre in the cave landscape, a
landscape duplicated nowhere on the Earth's surface.
Caving, from this point of view, is truly an esthetic or
wilderness experience. It requires solitude, a leisurely
pace, and a sense of absorption into the environment.
The emergence of the caver into the misty air under
the bright stars of a summer's night is indeed the
return from another world.
CAVES AS GEOLOGICAL REPOSITORIES
AND CAVES AS PARTS OF
GROUNDWATER FLOW PATHS
The cave environment may be described as dark,
wet, neutral to mildly alkaline, and oxidizing. The
variations of these parameters are much less than the
Encyclopedia of Caves.
© 2012 Elsevier Inc. All rights reserved.
104
c
variations of similar parameters on the Earth's surface.
Chemical reactions under these very precisely con¬
trolled conditions permit the growth of unusual miner¬
als and the growth of crystals of exceptional size.
Mineral deposits take on the form of stalactites, stalag¬
mites, flowstone, and other forms known collectively
as speleothems. Because these deposits are nourished by
water seeping down from the surface, changes in the
climate and vegetation on the surface leave their signa¬
tures in the growth bands of the speleothems. The
deposits of caves have become an important source of
paleoclimatic information.
Because caves are void spaces, they tend to fill up
with various materials collapsed or washed in from
the surface. Debris piles near cave entrances preserve
archaeological and paleontological deposits. Stream-
borne deposits of clays, silts, sands, and gravels record
past flood conditions. Because the filling of caves takes
place slowly over very long periods of time and
because still older deposits are preserved in the high
abandoned levels of cave systems, cave deposits are a
history book for the ice ages.
Caves, of course, do not form in isolation. Every
cave is related to a drainage system that now or at
some time in the past carried water on its way from an
inlet point to an outlet at a big spring. The under¬
ground pathways in limestone through which water
moves from sinkholes and sinking streams to the out¬
lets at springs are known as conduits. Conduit systems
are very long but most of them are invisible. Conduits
may be water-filled or may have no humanly accessi¬
ble entrances. Another definition of caves is that they
are those conduit fragments that are accessible to
human exploration. Careful inspection of the size,
shape, and patterns of caves as well as details of solu-
tional sculpturing on the cave walls provides much
information on the present or past behavior of the
groundwater flow system.
Much of the value of caves to the geological sciences
is that they preserve records over long periods of time.
Streams erode and deepen surface valleys and in so
doing destroy the stream channels, floodplains, and
valley shapes that were there before. In contrast, caves
deepen by forming new passages at lower levels, leav¬
ing the old levels high, dry, and well preserved.
CAVES AS HABITAT
We can also look for a definition of caves by consid¬
ering what animals consider caves to be dwelling
places. As habitats, caves have several distinct environ¬
mental properties. In temperate zones in the summer,
they tend to be cooler; conversely, they tend to be
warmer in the winter. This characteristic of
temperature buffering is shared not only by what we
would call true caves by any definition, but also by
rock overhangs and shelters as well. Some of the fauna
of true cave entrances such as phoebes and swallows
are found in rock overhangs as well. If we define caves
as an area of temperature buffering, then not only will
rock overhangs be included but so will many man¬
made structures such as cellars and the underside of
bridges.
A more appropriate environmental parameter to
consider is darkness. Species isolated in the darkness
of caves evolve a characteristic morphology, including
loss of eyes and pigment. The presence of eyeless,
depigmented species in a habitat would be one way to
define a cave from an organism's point of view. What
would be the characteristics of such a definition? First,
the cavity would have to be large enough to have a
zone of darkness. This would eliminate many short
cavities. That is, the length of the habitat must be great
enough relative to the diameter of the opening so that
sunlight does not penetrate to the far reaches of
the habitat. In practical terms, this length ranges
from a few meters to hundreds of meters. This defini¬
tion would exclude many open-air pits, which are
formed the same way caves are—by dissolution of car¬
bonate rock through the action of carbonic and sulfuric
acid.
The second characteristic of such a habitat would be
that it had been around long enough for animals either
to evolve in situ or for animals to migrate into the habi¬
tat. It seems likely that most caves and karst areas are
old enough to have an eyeless, depigmented fauna.
Some of the youngest caves known are on carbonate
islands such as San Salvador Island in the Bahamas.
These caves, less than 125,000 years old, have an eye¬
less, depigmented cave fauna. In glaciated regions, the
caves may be quite old but have only been ice free for
perhaps 10,000 years. In this case, there are few if any
terrestrial species that are either depigmented or eye¬
less. There are typically eyeless, depigmented aquatic
species. These may have survived underneath the ice in
unfrozen freshwater or have colonized from unglaciated
areas. In any case, 10,000 years is almost certainly not
long enough for species to evolve eyelessness in situ, a
process that probably takes between 100,000 and
1,000,000 years (Culver et al., 1995). Caves in evaporites
(gypsum) and lava form much more quickly (and dis¬
appear more quickly). There are few reports of eyeless,
depigmented species in gypsum caves, but lava tubes
often have a rich fauna. The reasons for this are com¬
plex, but one of the factors is that the lava fields them¬
selves are considerably older than a lava tube. Species
found in lava tubes only a few thousand years old
almost certainly migrated there from other lava tubes
and cavities in the lava.
ENCYCLOPEDIA OF CAVES
CAVE, DEFINITION OF
105
FIGURE 1 Types of caves.
A third characteristic of such a habitat is that, if dark¬
ness combined with the presence of eyeless species is
how it is delineated, then it will include a wide variety
of aquatic habitats in darkness that have no connection
with our intuitive idea of a cave. These include the
underflow of rivers, marine and freshwater beaches, and
any subsurface water. Many of these habitats, collec¬
tively termed interstitial or alternatively permeable and
small cavities (as opposed to permeable large cavities of
caves), have a rich subterranean fauna of eyeless, depig-
mented species. At least in principle, we can recognize
the differences between species in interstitial habitats
and species in caves. While they share a lack of pigment
and eyes, interstitial species tend to have shorter appen¬
dages, smaller body size, and a more worm-like appear¬
ance than do cave species (Coineau, 2000). In practice, it
is difficult to distinguish species from many taxonomic
groups. For example, snails occur in both habitats but
there is little to distinguish the two groups morphologi¬
cally. In areas with carbonate rocks, we can also distin¬
guish the two by the size of the cavity. The simple,
nonscientific definition of caves as natural cavities in a
rock that can be entered by people does not help in this
case. Cave animals can obviously thrive in cavities that
are too small to be entered by people. Cavity diameters
exist in a continuum, with a breakpoint between laminar
and turbulent flow cavities of approximately 1 cm. The
transition between laminar and turbulent flow is also an
important biological transition, and one that in a general
sense separates interstitial and cave habitats.
Thus, we can define a cave from a biological point
of view as a cavity, at least part of which is in constant
darkness, with turbulent water flow and with eyeless,
depigmented species present.
TYPES OF CAVES
When we define caves in terms of human access
rather than in terms of geologic setting, caves can be
found in many different geologic settings and have
been formed by many kinds of processes (Fig. 1).
Caves can be formed by purely mechanical processes.
Bulk masses of rock can be fractured and shifted by
tectonic processes such as faults. Likewise, rock masses
can break along fractures and then pull apart by the
ENCYCLOPEDIA OF CAVES
106
c
rock masses sliding under the influence of gravity. In
either case, void spaces formed between the rock
masses are called tectonic caves. Tectonic caves tend to
form in hard, massive, brittle rocks such as sandstones.
They tend to be small, typically a few tens of meters,
although some reach lengths of hundreds of meters.
Boulder piles, if the boulders are sufficiently large to
allow humans to explore the pore spaces between the
boulders, can form caves. These openings are called talus
caves. Because talus caves are simply the interstices
between a pile of boulders that may or may not be
completely roofed over, talus caves really do not have a
definable length. Some talus caves in the Adirondack
Mountains have been claimed to be several kilometers
in length, although this requires a fairly generous defini¬
tion of a cave. Rocks may be differentially eroded to pro¬
duce cave-sized openings. The categories of erosional
caves are defined by the erosion process.
Sea caves are formed by wave action on sea cliffs.
Fractures in the rock produce zones of weakness that
focus the attack of the waves. Sea caves are formed in
many kinds of resistant rocks in many parts of the
world. Fingal's Cave on the Isle of Staff a in the Scottish
Flebrides (made famous by Mendelssohn in his
Hebrides Overture ) is in columnar basalt. Sea caves form
in granite on the coast of Maine. Many sea caves are
found on Santa Cruz and Anacapa Islands off the coast
of Southern California (Bunnell, 1988, 1993). Sea caves
extend for distances from a few meters to a few hun¬
dred meters into the sea cliff. Frequently, the caves
consist of an inner chamber that is much larger than
the connection passage to the open sea. Access to a sea
cave may be possible at low tide, while the cave may
be flooded at high tide. Some sea caves have blow
holes in the back that spurt water when tides or storm
surges rush into the cave.
Aeolian caves are formed by wind action. These are
common in the arid regions of the American Southwest.
Wind-blown sand scours sandstone cliffs and sculpts
out chambers a few meters to a few tens of meters in
diameter. A typical aeolian cave is a bowl-shaped
chamber carved in solid rock. Ceiling heights vary from
1 to 2 meters. Often the entrance is a smoothly carved
hole in the cliff much smaller than the chamber inside.
Rock shelters are borderline caves. In places where a
resistant bed of rock, typically sandstone, overlies weak¬
er beds of rock, typically shales, the weaker rock can be
eroded away, leaving the resistant rock beds to form a
roof. Rock shelters are usually small, a few meters to a
few tens of meters in depth, although they can be tens
to hundreds of meters wide. As caves, rock shelters
usually do not extend to total darkness; however, such
shelters were habitats for early humans. They are often
rich repositories of archaeological material and are
sometimes referred to as caves by archaeologists.
Fine-grained, poorly consolidated sediment can be
swept away by stormwaters. In such locations as
Badlands National Monument in western South Dakota,
sediments flushed by stormwaters produce small caves
in the loosely consolidated silts and clays. These are
known as suffosional caves.
Of great interest to cave explorers are a variety of
caves in volcanic rocks. Lava conduits (also known as
tubes, tunnels, or pyroducts) form on the sides of volca¬
noes within streams of lava by a process called inflation
or by the crusting over of channels. Internal erosion by
the flowing lava often enlarges these conduits to result
in open passages with diameters of meters to tens of
meters and lengths up to many kilometers. The
processes that form these primary caves are entirely
different from the processes that form secondary solu¬
tion caves in limestone and other rock but they are
caves in the same sense that other natural openings of
human size are caves. Lava caves are also habitat for
organisms. Volcanic areas that have discharged fluid
basaltic lavas usually have systems of lava conduits
including volcanic islands such as Hawaii, Galapagos,
the Canaries and Azores, Jeju Island (South Korea),
Iceland, and intracontinental, hot-spot derived lava pla¬
teaus such as southern Oregon, Northern California,
Kenya, Syria, Jordan, and Saudi Arabia, as well as
southern Australia and Undara/Queensland. Lava
Beds National Monument in Northern California pro¬
vides easily accessed examples (Waters et al., 1990).
When glaciers melt, the meltwater drains down into
the glacier through fissures. This water moves along
the base of the glacier and, because it is slightly
warmer than the freezing point of water, it gradually
carves out long tunnels that open at the front of the
glacier, forming the rivers that drain the glaciers.
Glacier caves are ice tunnels with floors of rock and
walls and ceilings of ice. When the surface of the gla¬
cier is below freezing, the tunnels drain and become
open to exploration. When the glacier is melting, the
tunnels are often filled with water.
Most important are the solution caves. These form
by chemical dissolution of the bedrock by circulating
groundwater. They come in a great variety depending
on the type of rock and the source and chemistry of the
water that does the dissolving. Most solution caves are
formed in limestone, a rock consisting largely of cal¬
cium carbonate, or dolomite, a rock consisting largely
of calcium magnesium carbonate. These would be
called limestone caves or dolomite caves. In arid
regions, the more soluble rock, gypsum and calcium
sulfate, is exposed at the land surface, and one finds
gypsum caves, for example, in West Texas, western
Oklahoma, and New Mexico, where gypsum rock
occurs at the land surface. Caves easily form in salt and
a few other highly soluble materials, but such caves are
ENCYCLOPEDIA OF CAVES
CAVEFISH OF CHINA
107
rare because salt does not survive at the land surface
except in a few extremely arid regions.
Figure 1 shows the variety of sources for the water
responsible for the development of caves. Most of the
caves are dissolved by the movement of groundwater
in contemporary drainage basins. In coastal regions,
the mixing of fresh groundwater with saltwater pro¬
duces an aggressive solution that can dissolve out
caves. Some caves (for example, the large caves of the
Black Hills of South Dakota) are formed from hot
water rising up from deep within the rock. Carlsbad
Caverns in New Mexico and other caves of the
Guadalupe Mountains have been formed by sulfuric
acid derived from the oxidation of hydrogen sulfide
migrating upward from the oil fields to the east.
In summary, caves form in a great variety of rocks
by a great variety of geological and chemical processes.
Each has its importance to geology. However, the com¬
mon theme that binds this diverse collection of cavities
together is their interest to human explorers and their
use as habitat by cave-adapted organisms.
Bibliography
Bunnell, D. E. (1988). Sea caves of Santa Cruz Island. Santa Barbara,
CA: McNally and Loftin.
Bunnell, D. E. (1993). Sea caves of Anacapa Island. Santa Barbara, CA:
McNally and Loftin.
Coineau, N. (2000). Adaptations to interstitial groundwater life. In
H. Wilkens, D. C. Culver, & W. F. Humphreys (Eds.), Subterranean
ecosystems (pp. 189—210). Amsterdam: Elsevier.
Culver, D. C., Kane, T. C., & Fong, D. W. (1995). Adaptation and natu¬
ral selection in caves. Cambridge, MA: Harvard University Press.
Ford, D. C., & Williams, P. W. (1989). Karst geomorphology and hydrol¬
ogy. London: Unwin Hyman.
Shaw, T. R. (1992). History of cave science. Sydney, Australia: Sydney
Speleological Society.
Waters, A. C., Donnelly-Nolan, J. M., & Rogers, B. W. (1990). Selected
caves and lava-tube systems in and near Lava Beds National
Monument, California. U.S. Geological Survey Bulletin, 1673.
White, W. B. (1988). Geomorphology and hydrology of karst terrains.
New York: Oxford University Press.
CAVEFISH OF CHINA
Li Ma and Ya-hui Zhao
Department of Biology, University of Maryland, U.S. A.
Key Laboratory of the Zoological Systematics and Evolution,
Institute of Zoology, Chinese Academy of Sciences,
Beijing, China
THE DISTRIBUTION AND DIVERSITY
OF CHINESE CAVEFISH
Cavefishes, or hypogean fishes, are a distinctive group
of fishes. Their life history binds them to be located
in caves and other subterranean waters. In 1854,
Schiner divided all creatures that live in subterranean
and underground waters into three types: troglobites,
troglophiles, and trogloxenes. Troglobites have typical
adaptive characteristics allowing them to complete
their entire life cycles in caves; troglophiles lack special
adaptations to caves but nonetheless rely heavily
on subterranean waters during their life cycles; and tro¬
gloxenes enter the cave accidentally and their life cycles
and characteristics do not necessarily require caves.
Globally, most species of cavefish are distributed
in the tropics and subtropics, with 75% of all known
populations predominantly located in Southeast Asia
or Central and South America, where the landscape
is composed of limestone and other soluble rock
types (also known as a karst landscape). The southern
and western parts of China have more than 620,000 km 2
of karst (Huang et ah, 2008), which provides suitable
conditions for cavefish evolution. The earliest record
of cavefish in Chinese history was in 1436 when a local
doctor, Mao Lan, recorded cavefish in South Yunnan.
This was the well-known golden-line fish, which
lives mostly in Dian Lake near caves in Yunnan
Province. This golden-line fish, now recognized as
Sinocyclocheilus grahami, is a troglophilic, partially
cave-dwelling fish. The earliest paper report on a tro-
globitic fish is also in China. In 1540, Yingjing Xie, a
local governor of Guangxi, went to A'lu Cave and
recorded "there was a kind of transparent fish coming
out if the subterranean river rose very much" in his
travel notes on A'lu Cave. This transparent (also blind)
fish still lives in the same cave and was described as
Sinocyclocheilus hynlinus (Fig. 1) in 1994 (Chen et al.,
1994).
At the end of 2010, China had 95 described
species of hypogean (cave and phreatic) fishes which
belong to three families: Cyprinidae, Cobitidae, and
Balitoridae (Romero et al., 2009; Table 1). There are 55
species of Chinese hypogean cyprinids that include 4
genera: Onychostoma, Sinocrossocheilus, Sinocyclocheilus,
and Typhlobarbus. Onychostoma macrolepis is a cavefish
that was found to "hibernate" during winter. This is
the only hypogean species in China found north of the
Yangtze River. The family Cobitidae is a family charac¬
terized by a wormlike or fusiform body, whose mem¬
bers are mostly bottom dwellers in freshwater. Two
troglomorphic species of the genus Protocobitis have
been described and both are from China. The family
Balitoridae is another freshwater family of the order
Cypriniformes, which is characterized by having three or
more pair of barbels. Some species are scaleless.
Many tend to hide underneath rocks. The Chinese
hypogean fishes of this family include 7 genera:
Heminoemacheilus, Oreonectes, Paracobitis, Paranemacheilus,
Schistura, Triplophysa, and Yunnanilus.
Encyclopedia of Caves.
© 2012 Elsevier Inc. All rights reserved.
108
c
FIGURE 1 A comparison between the general
morphologies of (A) adult S. grahami, (B) S. hyalinus,
and (C) S. brevibarbatus. Scale bars = 1 cm.
The genus Sinocyclocheilus (Cyprinidormes, family:
Cyprinidae), established by Fang Bingwen in 1936,
is endemic to China and can be found in the karst cave
waters and surface rivers or lakes in Yunnan-Guizhou
Plateau and the surrounding region (including east of
Yunnan Province, south central of Guizhou Province,
and northwest of Guangxi Zhuang Autonomous Region)
(Fig. 2). Almost all Sinocyclocheilus species live in caves
for at least part of their life cycles. Yunnan has the most
Sinocyclocheilus species, which is mainly attributed to an
abundance of troglophiles, while Guangxi and Guizhou
contain about half of the troglobitic cavefish in China.
The genus Sinocyclocheilus is mainly distributed near
25°N. In river systems, most species of Sinocyclocheilus
are found in the Qianjiang River north tributary of
Xunjiang River, upstream of the largest Xijiang River
tributary of the Zhujiang River system, and more
upstream of the Hongshuihe Valley (including the asso¬
ciated underground river). Sinocyclocheilus may be the
biggest cyprinid genus in China with more than 60 nom¬
inal species currently described. At least 50 species are
valid among them, and 25 (50%) are troglobites (Zhao
and Zhang, 2009).
BIOLOGY OF CHINESE CAVEFISH
Chinese cavefish vary greatly in their size. For
example, the maximum length of adult cavefish is quite
variable. Some species are large, like Sinocyclocheilus
altishoulder and S. hugeibarbus, where the largest recorded
individual is 200 mm (standard length); whereas in
Oreonectes anophthalmus, the average adult length is
around 30 mm.
Due to the absence of photosynthesis in caves,
almost all Chinese cavefish are predators. Based on
stomach contents, their diet seems to consist of fish
scales, mollusks, and insects. The diet also can include
algae (mostly diatoms), insect larvae, and other less
defined organic materials.
Reproduction also varies considerably in the genus
Sinocyclocheilus. Some species show continuous repro¬
ductive activity in all four seasons, a phenomenon
that is seldom seen at other cyprinid fish. Normally
most fish in the subtropical zone of the Northern
Hemisphere multiply during the first half of each year.
Perhaps because these areas have four distinct seasons,
and the winter temperature is low, fish in these areas
ENCYCLOPEDIA OF CAVES
CAVEFISH OF CHINA
109
TABLE 1 All the Valid Species
of Chinese Hypogean Fishes
Species
Namer
Year
Distribution
Family: Cyprinidae
Genus Onychostoma
Gunther
1896
1. Onychostoma macrolepis
Bleeker
1871
Beijing, Hebei, Shandong, Shanxi, Henan
Genus Sinocrossocheilus
Wu
1977
2. Sinocrossocheilus bamaensis
Fang
1981
Guangxi, Guizhou
3. Sinocrossocheilus liuchengensis
Liang
1987
Guangxi
4. Sinocrossocheilus megalophthalmus Chen, Yang, and Cui
2006
Guangxi
Genus: Sinocyclocheilus
Fang
1936
5. Sinocyclocheilus altishoulderus
Li and Lan
1992
Guangxi
6. Sinocyclocheilus anatirostris
Lin and Luo
1986
Guangxi
7. Sinocyclocheilus angularis
Zheng and Wang
1990
Guizhou
8. Sinocyclocheilus angustiporus
Zheng and Xie
1985
Guizhou, Yunnan
9. Sinocyclocheilus anophthalmus
Chen and Chu
1988
Yunnan
10. Sinocyclocheilus aquihornes
Li and Yang
2007
Yunnan
11. Sinocyclocheilus bicornutus
Wang and Liao
1997
Guizhou
12. Sinocyclocheilus brevibarbatus
Zhao, Lan, and Zhang
2009
Guangxi
13. Sinocyclocheilus brevis
Lan and Chen
1992
Guangxi
14. Sinocyclocheilus broadihornes
Li and Mao
2007
Yunnan
15. Sinocyclocheilus cyphotergous
Dai
1988
Guizhou
16. Sinocyclocheilus donglanensis
Zhao, Watanabe, and Zhang
2006
Guangxi
17. Sinocyclocheilus furcodorsalis
Chen, Yang, and Lan
1997
Guangxi
18. Sinocyclocheilus grahami
Regan
1904
Yunnan
19. Sinocyclocheilus guilinensis
Ji
1982
Guangxi
20. Sinocyclocheilus guishanensis
Li
2003
Yunnan
21. Sinocyclocheilus huaningensis
Li
1998
Yunnan
22. Sinocyclocheilus hugeibarbus
Li and Ran
2003
Guizhou
23. Sinocyclocheilus hyalinus
Chen and Yang
1994
Yunnan
24. Sinocyclocheilus jii
Zhang and Dai
1992
Guangxi
25. Sinocyclocheilus jiuxuensis
Li and Lan
2003
Guangxi
26. Sinocyclocheilus lateristritus
Li
1992
Yunnan
27. Sinocyclocheilus lingyunensis
Li, Xiao, and Luo
2000
Guangxi
28. Sinocyclocheilus longibarbatus
Wang
1989
Guangxi
29. Sinocyclocheilus longifinus
Li and Chen
1994
Yunnan
30. Sinocyclocheilus luopingensis
Li and Tao
2003
Yunnan
31. Sinocyclocheilus macrocephalus
Li
1985
Yunnan
32. Sinocyclocheilus macrolepis
Wang
1989
Guizhou, Guangxi
33. Sinocyclocheilus macrophthalmus
Zhang and Zhao
2001
Guangxi
( Continued )
ENCYCLOPEDIA OF CAVES
110
TABLE 1 ( Continued )
C
Species
Namer
Year
Distribution
34. Sinocyclocheilus macrosccilus
Li
1992
Yunnan
35. Sinocyclocheilus maculatus
Li
2000
Yunnan
36. Sinocyclocheilus maitianheensis
Li
1992
Yunnan
37. Sinocyclocheilus malacopterus
Chu and Cui
1985
Yunnan
38. Sinocyclocheilus microphthalmus
Li
1989
Guangxi
39. Sinocyclocheilus multipunctatus
Pellegrin
1931
Guangxi, Guizhou
40. Sinocyclocheilus oxycephalus
Li
1985
Yunnan
41. Sinocyclocheilus purpureus
Li
1985
Yunnan
42. Sinocyclocheilus qiubeiensis
Li
2002
Yunnan
43. Sinocyclocheilus qujingensis
Li, Mao, and Lu
2002
Yunnan
44. Sinocyclocheilus rhinocerous
Li and Tao
1994
Yunnan
45. Sinocyclocheilus robustus
Chen and Zhao
1988
Guizhou
46. Sinocyclocheilus tianlinensis
Zhou, Zhang, and He
2003
Guangxi
47. Sinocyclocheilus tileihornes
Mao, Lu, and Li
2003
Yunnan
48. Sinocyclocheilus tingi
Fang
1936
Yunnan
49. Sinocyclocheilus wumengshanensis
Li, Mao, and Lu,
2003
Yunnan
50. Sinocyclocheilus xunlensis
Lan, Zhao, and Zhang
2004
Guangxi
51. Sinocyclocheilus yangzongensis
Tsii and Chen
1977
Yunnan
52. Sinocyclocheilus yaolanensis
Zhou, Li, and Hou
2009
Guizhou
53. Sinocyclocheilus yimenensis
Li and Xiao
2005
Yunnan
54. Sinocyclocheilus yishanensis
Li and Lan
1992
Guangxi
55. Typhlobarbus nudiventris
Chu and Chen
1982
Yunnan
Family: Cobitidae—loaches
Genus Protocobitis
Yang
1994
56. Protocobitis polylepis
Zhu, Lu, Yang, and Zhang
2008
Guangxi
57. Protocobitis typhlops
Yang, Chen, and Lan
1994
Guangxi
Family: Balitoridae
Genus Heminoemacheilus
Zhu and Cao
1987
58. Heminoemacheilus hyalinus
Lan, Yang, and Chen
1996
Guangxi
59. Heminoemacheilus zhengbaoshani
Zhu and Cao
1987
Guangxi
Genus Oreonectes
Gunther
1868
60. Oreonectes anophthalmus
Zheng
1981
Guangxi
61. Oreonectes furcocaudalis
Zhu and Cao
1987
Guangxi
62. Oreonectes macrolepis
Huang, Chen, and Yang
2009
Guangxi
63. Oreonectes microphthalmus
Du, Chen, and Yang
2008
Guangxi
64. Oreonectes retrodorsalis
Lan, Yang, and Chen
1995
Guangxi
65. Oreonectes translucens
Zhang, Zhao, and Zhang
2006
Guangxi
Genus Paracobitis
Bleeker
1863
( Continued )
ENCYCLOPEDIA OF CAVES
CAVEFISH OF CHINA
111
TABLE 1 ( Continued )
Species
Namer
Year
Distribution
66. Pamcobitis maolanensis
Li, Ran, and Chen
2006
Guizhou
67. Pamcobitis posterodarsalus
Ran, Li, and Chen
2006
Guangxi
Genus Paranemacheilus
Zhu
1983
68. Paranemacheilus genilepis
Zhu
1983
Guangxi
Genus Schistura
McClelland
1838
69. Schistura dabryi microphthalmus
Liao and Wang
1997
Guizhou
70. Schistura lingyunensis
Liao and Luo
1997
Guangxi
Genus Triplophysa
Rendahl
1933
71. Triplophysa aluensis
Li and Zhu
2000
Yunnan
72. Triplophysa gejiuensis
Chu and Chen
1979
Yunnan
73. Triplophysa longibarbatus
Chen, Yang, Sket, and Aljancic
1998
Guizhou
74. Triplophysa nandanensis
Lan, Yang, and Chen
1995
Guangxi
75. Triplophysa nasobarbatula
Wang and Li
2001
Guizhou
76. Triplopliysa qiubeiensis
Li and Yang
2008
Yunnan
77. Triplophysa rosa
Chen and Yang
2005
Chongqing
78. Triplophysa shilinensis
Chu and Yang
1992
Yunnan
79. Triplophysa tianeensis
Chen, Cui, and Yang
2004
Guangxi
80. Triplophysa xiangshuingensis
Li
2004
Yunnan
81. Triplophysa xiangxiensis
Yang, Yuan, and Liao
1986
Hunan
82. Triplophysa yunnanensis
Yang
1990
Yunnan
83. Triplophysa zhenfengensis
Wang and Li
2001
Guizhou
Genus Yunnanilus
Nichols
1925
84. Yunnanilus bajiangensis
Li
2004
Yunnan
85. Yunnanilus beipanjiangensis
Li, Mao, and Sun
1994
Yunnan
86. Yunnanilus discoloris
Zhou and He
1989
Yunnan
87. Yunnanilus longidorsalis
Li, Tao, and Lu
2000
Yunnan
88. Yunnanilus macrogaster
Kottelat and Chu
1988
Yunnan
89. Yunnanilus macrolepis
Li, Tao, and Mao
2000
Yunnan
90. Yunnanilus nanpanjiangensis
Li, Mao, and Lu
1994
Yunnan
91. Yunnanilus niger
Kottelat and Chu
1988
Yunnan
92. Yunnanilus obtusirostris
Yang
1995
Yunnan
93. Yunnanilus paludosus
Kottelat and Chu
1988
Yunnan
94. Yunnanilus parvus
Kottelat and Chu
1988
Yunnan
95. Yunnanilus pulcherrimus
Yang, Chen, and Lan
2004
Guangxi
finish their breeding activity as early as possible, so
that the young fry have time to grow before the next
cold season. Thus, it is likely that breeding of most
Sinocyclocheilus species occurs between the spring and
summer. Since the streams in all caves inhabited by
Sinocyclocheilus connect with surface rivers, their water
levels are affected by regional precipitation and can
change radically within a short time period. These
ENCYCLOPEDIA OF CAVES
112
(A) (C)
FIGURE 2 Distribution map of Chinese hypogean fishes. (A) Distribution of hypogean Cyprinidae (excluding the genus Sinocyclocheilus).
(B) Distribution of the species in the genus Sinocyclocheilus. (C) Distribution of hypogean Cobitidae and Balitoridae.
radical changes in water levels may trigger reproduc¬
tive activity. The sex ratio of S. tingi varies from 1.31 to
4.2, with females being more abundant than males. The
egg diameter of S. macrolepis is 1.5 ~ 2.0 mm, whereas
the diameter of mature S. bicormutus and S. angularis
eggs is 1.5 ~ 2.0 mm, and the mature egg of S. tingi
averages 2.1 mm. According to dissection and observa¬
tion, it was found that the absolute brood size of
S. macrolepis is about 2000 and the average brood size of
S. tingi is 2595 ~ 3087. In both species, embryos develop
synchronously, suggesting that ovulation occurs simul¬
taneously. The fecundity of 2-year-old S. yangzongensis
is 3585, and increases to 22,000 eggs after 5 years.
THE CHARACTERISTICS OF
CHINESE CAVEFISH
Chinese cavefish have five main characteristics that
make them unique among the cavefish of the world.
1. China hosts the greatest and most diverse number
of troglobitic fish species. So far, at least 46 species
of troglobites have been found in China with the
number still increasing with the discovery of
new species.
2. Although China has the richest variety of troglobite
fish species, they mainly belong to two genera
from Cypriniformes, Sinocyclocheilus and Triplophysa,
which together account for 71.7% of the Chinese
troglobitic cavefish.
3. The range of preferred habitat is relatively narrow.
Chinese troglobitic fish, excluding Triplophysa
xiangxiensis which lives in Hunan Province and
T. rosa that lives in Chongqing, are all distributed
in Yunnan Province, Guizhou Province, and the
Guangxi Zhuang Autonomous Region. In particular,
the Guangxi Zhuang Autonomous Region has the
largest number of troglobitic cavefish (22 species,
accounting for 48% of all Chinese troglobite fish).
Chinese troglobitic fish species are most concentrated
in the karst areas of the Yunnan-Guizhou Plateau.
Most of these species are located only in one
independent water body, such as one small river,
one brook, even one cave. For example, Oreonectes
anophthalmus is unique to Taiji Cave of Wuming,
Guangxi.
ENCYCLOPEDIA OF CAVES
CAVEFISH OF CHINA
113
4. Intense speciation is seen among Chinese cavefish,
despite the narrow area of habitat in which they
live. The genus Sinocyclocheilus is a good example.
Its range reaches from Yunnan Yimen in the west
to Guangxi Fuchuan in the east, and north from
Guizhou Huaxi to Yunnan Yanshan in the south.
The distance between east and west is 900 km,
but the distance between south and north is only
300 km. In this relatively small area, 50 species of
Sinocyclocheilus have been recorded, making
Sinocyclocheilus the largest genus of Chinese
Cyprinidae, as well as the genus with the greatest
amount of adaptive morphology. The large diversity
of Sinocyclocheilus species is probably the
result of strong selective pressures of isolated
cave environments, combined with the unique
environmental effects that might be present in the
Qinghai-Tibet Plateau uplift. It is interesting to note
that this intense speciation is seldom seen in the
other areas of the world where cavefish are
distributed.
5. Each species of Chinese cavefish tends to have a
small population size. Limited food resources are
probably responsible for this. A good example is
S. hyalinus, of which only very few specimens were
found in their only habitat, A'lu cave in Guangxi
Zhuang Autonomous Region.
MORPHOLOGY AND ADAPTATION
Cave habitats are unique environments characterized
by permanent darkness, the absence of green plants, and
seasonal scarcity. Chinese cavefish have evolved a series
of constructive and regressive morphological changes
to survive in these harsh conditions. Constructive fea¬
tures often include a protruding jaw, an increase in the
number of taste buds, over-developed barbels, and vari¬
ous specialized appendages, while regressive changes
mainly include eye degeneration, reduction or loss
of pigmentation, and the disappearance of scales. In
addition to these common adaptations, Chinese cavefish
have other unique features, such as the development
of a humped back, a horn, and a head drape. Below we
discuss the morphology and adaptation of Chinese cave¬
fish from the perspective of their constructive and
regressive structures.
Changes in Body Shape
Cavefish often have a very different outward appear¬
ance compared to their surface counterparts. There are
three different body shapes of cavefish, the fusiform
type, the humpback type, and the head-horn type.
In Sinocyclocheilus, the head shape of the surface species
is very similar to the normal shape of other fishes (e.g.,
S. grahami), in which the length of the head is longer
than the height of the body; the snout is slightly
pointed; the mouth is subinferior; and the upper jaw
is elongated (Fig. 1). In some cave-dwelling fish, the
head is duckbilled, with the anterior half-depressed and
the posterior half-raised, and the jaws are wider and
more protruding than their related surface species.
For instance, the jaws of S. hyalinus and S. anophthalmus
are wider and more protruding than S. grahami (Fig. 3).
The pleat on the head ridge that is found in S. hyalinus
FIGURE 3 Differences in head shape between (A) S. grahami, (B) S. hyalinus, and (C) S. anophthalmus. The jaws of S. hyalinus and
S. anophthalmus adults are wider and more protruding than the jaw of S. grahami.
ENCYCLOPEDIA OF CAVES
114
c
FIGURE 4 Examples of the different species of Sinocyclocheilus. (A) S. tianlinensis ; (B) S. microphthalmus; (C) S. furcodorsalis;
(D) S. broadihornes; (E) S. tileihornes; (F) S. hugeibarbus; (G) enlarged image of S. hugeibarbus head; (H) S. rhinocerous; (I) enlarged image of
S. rhinocerous head.
is also a very unique morphological feature (Fig. 3), and
may have the function of increasing sensitivity to fluc¬
tuations in water flow.
Humpback and Horn
One of the prominent constructive structures of
Sinocyclocheilus is the humpback and horn, which can
be very well developed in some of the troglobitic spe¬
cies. The humpback is free of bone and mainly consists
of adipose tissue (Wang et al., 1995), while the horn
consists of the frontal and parietal bones. The horn
structure is found in both sexes of the same species,
and the shape of the horn also varies between different
Sinocyclocheilus species: some are small (S. angularis),
some are forked (S. bicornutus), some show a thin pro¬
trusion (S. rhinocerous), and several show a tile shape
(S. tileihornes) (Fig. 4).
The histology of the S. hyalmus head horn has been
extensively studied (Fig. 5). Overall, the structure of
horn can be divided into three segments, from anterior
ENCYCLOPEDIA OF CAVES
CAVEFISH OF CHINA
115
FIGURE 5 Horn structure of S. hyalmus. (A) Lateral view of head; (B) magnified image of horn; (C) dorsal view of horn. The horn is
divided into three segments; (D), (E), and (F) show the histological structure of horn, ke, keratinized epithelium; dfl, dense fibrous layer; at,
adipose tissue; mf, muscle fibers; cf, collagen fibrils; ca, cartilage.
to posterior: the first is the apical section, the second is
the middle section, and the third is the basal section.
The apical section is characterized by a thick layer of
keratinized epithelium containing thick collagen fibrils
(Fig. 5D). The middle section is composed of kerati¬
nized epithelium, lying over a compact fibrous layer,
adipose tissue, and muscle fibers. A thick dermal retic¬
ular layer forms the compact fibrous layer under the
thick keratinized epithelium. There are adipose tissue
and muscle fibers under the fibrous layer (Fig. 5E),
and the basal section has a much thinner layer of
collagen fibrils and a compact fibrous layer under
the keratinized epithelium, which is characterized by
the presence of very thick adipose tissue and muscle
fibers under the fibrous layer (Fig. 5F). Neither nerve
ending nor glands are present in the horn. These char¬
acteristics suggest that the function of the horn may be
to store fat for nourishment of the adjacent brain.
The horn may also be used for protecting the brain
from bumping into rock walls. Li and Tao (1994) stud¬
ied horn structure in S. rhinocerous, and found that its
bony part is composed of three pieces, each shaped
like a right-angled triangle (Fig. 6A). As shown in the
schematic diagram (Fig. 6B), "A" stands for the tip
of the horn and "BCD" stands for the base of horn.
Thus when "A" hits a rocky surface, the pressure will
transmit downward to line "BE," "CE," and "DE" and
then disperse and decrease greatly, thereby protecting
FIGURE 6 Bone structure of the horn of S. rhinocerous (A) with a
schematic diagram (B).
the brain during swimming in rocky caves (Li et al.,
1997). Although the bony horn of S. rhinocerous can't
completely protect the body from injuries produced by
bumping into rock walls, it could disperse pressure
equally throughout the head surface to reduce the
magnitude of the force. There are also some other
hypotheses on functions of horn-like character, but real
answers still need further studies.
ENCYCLOPEDIA OF CAVES
116
c
FIGURE 7 The barbels, nostril, and caudal fin of Triplophysa.
The humpback is found in different clades of
Sinocyclocheilus, demonstrating that it is the result
of convergent evolution, and likely to be adaptive.
Although the precise function of the horn-like struc¬
ture is still unknown, it is found only among some
hypogean species of the genus Sinocyclocheilus.
Sensory Apparatus
Sensory structures such as barbels, taste buds, nostrils,
and the lateral line system have been augmented in cave-
fish. Most of the Chinese cave-dwelling species have a
more sensitive sense of smell, with the nostrils more
anterior than their surface relatives, and in some species
the nasal valve is hyper-developed and takes on a beard
shape (Fig. 7). At the same time, cave species often have
an increased number of taste buds: for example, S. hyali-
nus has more taste buds than S. grahami.
The barbel has a tactile function, and also has taste
buds on it, implying both mechanosensory and gusta¬
tory functions. Fish can use barbels to detect the
surrounding environment and amplify the search
for food. There are abundant neurons and sinuses in
the barbels of Sinocyclocheilus anatirostris, showing
ENCYCLOPEDIA OF CAVES
CAVEFISH OF CHINA
117
that the barbels have a prominent sensory function.
All Sinocyclocheilus have two pairs of well-developed
barbels; however, the extent of barbel development
differs according to the level of exposure to light.
Among those Chinese cave-dwelling species that
live in both cave and non-cave environments with
abundant light, most have barbels that are moderately
developed and extend from the anterior edge of the
eyes to the preopercular bone. Those partially cave¬
dwelling species that live in shaded bodies of water
have more developed barbels that can extend to
the trailing edge of the opercle, and may even extend
to the starting point of the pectoral fin, such as is
the case in S. longibarbatus. There are several cave-
restricted species of Chinese cavefish in which barbel
augmentation is often less than that found in partially
cave-dwelling species. The least developed barbels
do not reach the leading edge of the eye, as in
S. cyphotergous, while the most developed barbels
among the cave-restricted species just touch the preo¬
percular bone, as in S. microphthalmus. This shows that
the development of the barbels is inversely correlated
with eye development: the barbels are long when the
eye is smaller, a possible tradeoff for loss of eyesight.
This is a very interesting phenomenon if one takes into
account that cave-restricted species also have a very
developed sensory tube and a special projection on the
head. Therefore, it is reasonable to conclude that
enhanced barbels is a primitive feature compensating
for the loss of eyesight, whereas the sensory tube and
projection are more advanced compensatory forms.
The lateral line system of Sinocyclocheilus is very
developed and specialized. In addition to possessing the
tubal system of Cyprinidae, both sides of the lateral line
canal also have many short branches named sensory
tubes. These are especially abundant in the head
(Fig. 3C). In species such as S. cyphotergous, which shows
the most extensive sensory tube network, sensory tubes
can even be found under the surface of the skin on both
sides of lateral line at the anterior part of the trunk.
These special sensory tubes are found in all species of
Sinocyclocheilus, which suggests that this feature was
present in their common ancestor. It is estimated that
the initial evolution of the sensory tube system occurred
in a common ancestor living in an environment that
may have had intimate connections with a karst cave
environment. During the course of evolution and adap¬
tation, this common ancestor acquired the trait first, and
then passed it to its descendants.
Specialized Appendages
Troglobitic species often have more developed
appendages, such as modified pectoral and pelvic fins,
than troglophilic species (Fig. 1). The overdeveloped
appendages could diminish the energy consumption of
the animal and improve the efficiency of movement, as
seen in Sinocyclocheilus. In Sinocyclocheilus, the degener¬
ation of the opsin system directly weakens the motor
skills of the fish and results in a decreased ability to
respond quickly to environmental changes. At the
same time, the enhanced pectoral fin increases the
fish's ability to balance itself. The morphological
changes in the pectoral girdle and pectoral fin are
detailed as follows. Overall, the pectoral girdle of
Sinocyclocheilus cave species tends to be more elongated
and narrower than surface species. Corresponding to a
shorter surface for muscle adherence, the muscle attach¬
ment surface composed of the cleithrum and corcoideum
has a more flattened concave shape. In some partially
cave-dwelling species, the cleithrum and corcoideum are
greatly reduced and in some cases they disappear
completely. Additionally, the coracoid of surface species
is wide, and higher than the cleithrum, with shallow
ridges to accommodate muscle attachments. In
Sinocyclocheilus, the coracoid develops into a triangle with
the tip positioned anterior to the plane of the cleithrum.
In partially cave-dwelling species, the pectoral fin is
short, with a rear projection that does not reach the
starting point of the ventral fin, and an average number
of 15 branches in the fin ray. In contrast, the pectoral
fin of the cave-restricted species is long, with a rear
protraction that extends past the starting point of the
ventral fin, and an average of 13 branches in the fin ray.
Adipose Storage
Chinese cavefish restricted to caves store large
amounts of adipose tissue in various parts of the body,
such as the forehead, the horn, the base of the dorsal
fin, the base of the caudal fin, and the sides of the body
(Fig. 8). There is also a considerable amount of adipose
tissue stored in the eye socket (Fig. 9F). Fat deposition
in the eye sockets should not be dismissed simply as
padding, because the storage of fat is very important to
cave animals and allows them to survive during sea¬
sons in which food does not enter caves. The lack of a
primary food source that can be produced in darkness
may cause a food limitation and thus require improve¬
ments in food-finding and energy storage capabilities.
For this reason cave animals store fat wherever they
have space, such as in the empty eye sockets and in the
horn. Storing fat locally in the horn may function to
provide adjacent tissues, such as the brain, nourishment
when there is low food input into the cave.
Eye Degeneration
The adaptive changes in the optic system of Chinese
cavefish are mainly seen as changes in the size of
ENCYCLOPEDIA OF CAVES
118
c
FIGURE 8 The distribution of adipose tissue in S. hyalinus. S. hyalinus store fat in the horn (A), forehead (B,C), base of the dorsal fin (D),
lateral face of the body (E), and at the base of the caudal fin (F).
the eye, the degeneration of the eye's structure, and
in some species its complete disappearance. There
are three different eye structures seen in the genus
Sinocyclocheilus: normal, dot-eye, and blind (Fig. 4).
These variations are strongly correlated with differing
amounts of sunlight each species is exposed to in the
caves they inhabit. Those troglophilic species that are
sometimes active at the surface of the water in brighter
areas have eyes that are clear but underdeveloped
when compared to surface-dwelling species. The cir-
cumorbital structures of these fish are also modified,
such that the lacrimal and supraorbital bones retain
the shapes found in surface-restricted species, while
the infraorbital and postorbital bones regress to a tubu¬
lar shape. In some species, the supraorbital bone and
jawbones have a tendency to fuse, such as in S. tingi.
In species that live in half-enclosed caves with weak
light, such as S. macrophthalmus, eyes often tend to be
larger, which may be beneficial for detecting weak
light. Structural changes are even more pronounced in
troglobitic Chinese cavefish species that spend their
entire life cycle in deep caves or underground streams
and are only active near the mouth of these caves at
night, returning to the depths before daybreak. The
eyes of these fish are so small that they are barely
visible under the surface of the skin that grows over
the eye socket, and in many cases they even disappear
completely. The infraobital and postorbital bones
either form tubes or disappear as well, while the
supraorbital bone and jawbone fuse together. The sides
ENCYCLOPEDIA OF CAVES
CAVEFISH OF CHINA
119
FIGURE 9 Comparison of the eye structures of S. grahami (A,C,E) and S. hyalinus (B,D,F). (A) Lateral view of S. grahami with normal eyes;
(B) lateral view of S. hyalinus without eyes; (C) section of a S. grahami eye; (D) transversal section through the eye region of S. hyalinus showing
the region containing fat tissue and a flap of skin; (E) retinal structure of S. grahami; (F) enlarged image of S. hyalinus eye region.
of the jawbone are widened and the rima oculi nar¬
rowed. The skin of the jaw has elongated and fused with
the inferior skin to seal up the rima oculi completely. In
addition, adipose tissue is deposited around the eye as
it degenerates and can even replace it completely, as in
S. anatirostris and S. hyalmus (Fig. 9).
Here we use S. grahami and S. hyalinus to demon¬
strate the details of the various changes in the eyes
of Chinese cavefish (Fig. 9). On examination of the
external features of these fish, it can be seen that
the S. grahami has fully developed eyes located toward
the front section of the head (Fig. 9A), while the eyes of
S. hyalinus seem to disappear (Fig. 9B). The eye struc¬
ture of S. grahami is complete and includes the cornea,
iris, pupil, and retina, with its fully developed neural
structure (Fig. 9C,E). In S. hyalinus, the eye disappears
completely, to the point that only adipose tissue is
found in the eye orbit when the head is serially
ENCYCLOPEDIA OF CAVES
120
c
sectioned (Fig. 9D,F). In contrast, S. anophthalmus does
have small eyes buried deep within the orbits of the
skull, which are also reduced in size, as seen in skeletal
comparisons between species (Fig. 10D). Sometimes,
the level of degeneration is different even within
same species, as is known to occur in S. anophthalmus
and Triplophysa tianeensis. Most individuals of these
species have lost eyes completely; however, a few
individuals have vestigial eyes buried under the skin
(Fig. 11B) on one side or both sides. Histological sections
of S. rhinocerous showed only a single rod cell and no
cone cells, thus demonstrating that their eyes have lost
light sensitivity.
Loss of Pigmentation
Body pigmentation in teleosts is due to three
types of dermal chromatophores: black melanophores,
which contain melanin; silver iridophores, which con¬
tain purines; and yellow xanthophores, which contain
pteridines. There is a dramatic decline in the total
number of melanophores in Chinese cavefish, as well
as a strong reduction in the ability of these cells to
synthesize melanin. Those cavefish species that live ex¬
clusively in subterranean streams where there is no
light are often completely albino, whereas those species
that live around cave entrances and are only partial
cave dwellers often have light black-brown or dark
brown pigment on their bodies. This is seen in all
three of our sample species (Fig. 1): the surface species
S. grahami, which has a deep yellow body color; the
cave-dwelling species S. hyalinus, which is an albino
with no scales; and S. brevibarbatus, which is normally
semitransparent with milky white fins. Of greater
interest is the fact that individuals of the same species,
captured in a cave environment, are often lightly
colored or colorless, while those captured outside
of the cave often have a much darker body color
(S. microphthalmus). Some Chinese cavefish can gain
pigmentation when they enter a bright environment.
It seems likely that the disappearance of pigmentation
in Chinese cavefish is mainly caused by the inhibition
of gene expression in the pigment production pathway,
rather than by the total loss of pigmentation genes.
Disappearance of Scales
The development of scales in cave-adapted
Sinocyclocheilus takes on an obvious regressive trend.
FIGURE 10 Comparison between the head skeleton of S. grahami (A,B,C) and S. anophthalmus (D,E,F). (A,D) lateral views; (B,E) dorsal
views; (C,F) ventral views.
ENCYCLOPEDIA OF CAVES
CAVEFISH OF CHINA
121
(A)
fin-rays
pterygiophore
centrum
urostyle
fin-rays
haemal spine
pectoral girdle
rib pterygiophore
pterygiophore spine
centrum
urostyle _
fin-rays
FIGURE 11 Whole-mount staining of bone and cartilage of S. grahami (A) and S. anophthalmus (B). The cartilage was stained with Alcian
blue and the bone was stained with Alizarin red. Scale bars = 50 mm.
as seen in their reduced distribution over the body and
the relative size of the lateral line scales. Those species
that live in abundant light, such as S. grahami, have
large well-developed scales that are overlapping over
the entire body. The most extreme example of scale
regression is seen in cave-restricted species such as
S. tianlinensis, which have no scales on the body at all.
We can only find lateral line pores on the sides.
The rate of scale regression the lateral line due to
adaptation to the cave environment is much slower
than the speed at which general body scales are lost.
As previously mentioned, even those Sinocyclocheilus
species that have completely lost their body scales still
retain the scales of the lateral line, although they are
underdeveloped. For those species that have partial
regression of body scales, the lateral line scales retain a
larger shape than those general body scales above and
below them, while in species that have well-developed
body scales, those of the lateral line are just as well
developed or even larger. This indicates that the lateral
line scales of Chinese cavefish are more resistant to
regressive evolution during cave adaptation.
PHYLOGENESIS, SPECIATION
MECHANISMS, AND BIOGEOGRAPHY
Evolution is at the center of all Chinese cavefish
research. Due to the isolated nature of caves, it is
possible to combine geological data with the existing
records of cavefish distribution to reveal the synergic
relationship between cavefish evolution and cave
development. The origin of the genera Sinocyclocheilus
and Triplophysa is attributed to the ancient geology
and climate of China. Current theory suggests that the
primitive ancestor of Sinocyclocheilus may have origi¬
nally lived on the Yunnan-Guizhou Plateau during the
late Tertiary period. During the Quaternary period,
the Qinghai-Xizang Plateau underwent a sudden up¬
ward shift that caused the geological environment to
change greatly, while the Yunnan-Guizhou Plateau
also underwent an upheaval but settled at a different
elevation. At the same time, the temperature of the
Earth began to decline, such that the size of the polar
ice caps greatly increased. As a result of these drastic
environmental changes, the Sinocyclocheilus common
ENCYCLOPEDIA OF CAVES
122
c
ancestor was forced to live in caves and eventually
took on such adaptations as tetraploidy of chromo¬
somes, an increase in the number of body scales, and a
smaller body size in order to survive.
Sinocyclocheilus is a monophyletic group with four
clades, given the names jii, angularis, cyphotergous,
and tingi, named according to their most representa¬
tive species. The geographical distribution of these
different clades displays very little overlap. The phy¬
logenetic tree of Sinocyclocheilus based on combined
morphological and molecular data indicates that
different cavefish species invaded individual cave
waters multiple times and acquired their troglo-
morphic traits independently (Xiao et al., 2005;
Fig. 12). The hypothesized troglophilic ancestor of
modern Sinocyclocheilus species was presumably
distributed throughout the karst region of Yunnan
and Guizhou at the beginning of the plateau uplift.
After the violent sudden upheaval of the plateau
approximately 3.4 Ma BP, also known as Act A of the
Qinghai-Xizang Movement, the eastern population
became isolated from the other groups and evolved
into today's jii clade. Phylogenetic trees based on
maximum parsimony, Bayesian analysis, or morphol¬
ogy, all show that jii is the ancestral clade. Soon after¬
ward, some Sinocyclocheilus in the western region
became true troglobites, dwelling exclusively in caves
and/or subterranean rivers. These became the ecolog¬
ically isolated angularis clade, which became sepa¬
rated from other troglophilic relatives partially in
response to the various climate changes and struc¬
tural changes of the karst landscape. The angularis
clade emerged from this ancestral group. Most
species of angularis clade have a horn on the head
and a humpback. In addition to the appearance of
the horn and humpback, the species in this clade are
all troglobites, possessing the usual regressive charac¬
teristics typical of caves; their eyes are either reduced
in size or absent; and their pigmentation is decreased.
Almost at the same time, further upheaval of the
land during Act B (2.6 Ma BP) of the Qinghai-Xizang
Movement served to isolate the remaining two clades:
tingi and cyphotergous, which are sister groups.
These clades share a common characteristic in not
forming the parietal bone projection on the head
and back. The tingi clade, which with one exception
exclusively exhibits troglophilic levels of adaptation
to hypogean life, became limited to the watershed of
the Nanpanjiang River. This clade mainly shows a
troglophilic life cycle and will forage out of the
cave, although not far away from cave entrances.
The troglophilic and troglobitic members of the
cyphotergous clade are currently distributed among
the drainage sites of the Hongshuihe River. The
cyphotergous clade is characterized by a humpback
and a reduced lateral line scale number, normally
55 scales. Whether a given Sinocyclocheilus species
emerged as troglobite or a troglophile can be closely
tied to the geological characteristics of the Yunnan-
Guizhou Plateau as it exists today. Troglobitic species
are more heavily concentrated in the central area
of the distribution range, where violent upheavals
have produced steep slopes and a complicated karst
environment. In contrast, troglophilic species tend to
be found closer to the eastern and western edges of
the overall distribution area, which exhibit less eleva¬
tion change. Vicariance is the primary mechanism for
speciation in the genus Sinocyclocheilus, with diversifi¬
cation mainly resulting from genetic drift in isolated
populations (Zhao and Zhang, 2009).
Geographical isolation is the main cause of species
diversity in Sinocyclocheilus. The ancestor of Sinocyclocheilus
was likely to be a troglophilic cavefish. During its
subsequent evolutionary history, some troglophilic
cave Sinocyclocheilus became isolated in caves, which
gradually produced special characteristics adaptive to
the dark environment, and thus troglobitic cavefish
species appeared. The isolation of cave or subterranean
waters prohibited genetic communication between the
fish groups living in different caves. In the cave-rich
southern areas, which are affected by high temperatures,
abundant precipitation, and surface erosion, cave col¬
lapse or underground river diversion was frequent,
further enhancing the geographical isolation of species
that once had a continuous range. Thus, different popu¬
lations isolated by underground water evolved in differ¬
ent directions, finally resulting in the formation of
independent species.
A very interesting phenomenon not often encoun¬
tered in cave-adapted animals is that different
Sinocyclocheilus species can be spatially distributed
at the same underground river or same cave at
the same time. For example, S. anatirostris,
S. microphthalmus, and S. lingyunensis all live in
Guangxi Lingyun caves. This phenomenon may be
attributed to abundant rainfall in these areas, which
enters caves and isolates existing fish populations
spatially, thus eventually leading to the formation of
separate species in the same cave.
Although geographical isolation is the most impor¬
tant mechanism for Sinocyclocheilus speciation, small
population sizes also have a role in this process. In
small isolated populations there are powerful effects
of genetic drift as well as natural selection. In small
populations derived from a single founding popula¬
tion, genetic differentiation and speciation can occur
more rapidly due to the absence of migration and gene
pool exchange. In summary, speciation in this genus
is attributed to several different factors, including iso¬
lation and small population sizes, which result in
ENCYCLOPEDIA OF CAVES
CAVEFISH OF CHINA
123
Habitat type
—• surface • • • • cave
Eye
O normal <22> reduced
Scale radius
□ present in whole
scale
Frontal-parietal area
A upheaved non-
si gruficantly
intermediate
> blind
10 1 100 | S. macrocephalus 1 DAO
040/ * S. macroctphalus 2 DAO
S. kuaningensis A O
'i
S. guishanensis A O
10/100. . .
l S. oxycepkalus QAO
’S. luttanettsis QAO
10/97.
097/69
10/66
100 j
r—S.i
>9 HQ-56/
10/95
I only present in scale
apical region
A upheaved acutely
0.47/
0.79/
10/99
053/
094/
086/
10/98
10 /ICO
10/100
10/100
10/100
10/100
L 05/
S. malacopterus I A O
malacopterus 2AO
S. malacopterus 3 AO
S. angustiporus DAO
1Q / too f S. yangzong easts 1 □ A O
*S.yangzongeitsis2 0AO
10 /100 S. grahami 1 DA O
°^6J * S.grahami 3 DAO
10/IC0_ L _ S grahaM{ 2 QAO
S.tingi DAO
S. maitianheensic AO
S. artophthahnus □
• S. lateristritus 3 A O
| S. lateristritus 5 A O
S. lateristritus 2 A O
S. lateristritus 1 A O
S. lateristritus 4 A O
S. qujiitgensis A O
S. multipunctatus 1 □ A O
S. multipunctatus 2 DAO
S. cyphotergous
- S. macrolepis QAO
-f.
• S. lingyunensis 1 A O
10 / ICO t S. lingyunensis 2 A O
1 JO /ICO
10/100
10/100
10/99
m
* S. lingyunensis 3 A O
•• S. lortgibarbatus □ A O
• £ macrophthahnus a O
078/56 1^’ tnicrophtkalmus 1 ■ AO
5S. microphthalmos2 ■ AO
■ S. micropktkalmus 3 ■ AO
* S. halfibinthts ■ AO
057/-
10/86
10/78
10/82
r • • -
L[U.
10/100
IV
10/87 I
10
10/100
S. anatirostris ia«
* S./urcodorsalis A •
-S’. jiuxuensis A O
5". bicomutus » AO
• • • • S. tianeensis A O
S. altishoulierus a O
. S. rhinocerous AO
.£■. hya linus A 9
-S'.jYi 1 DAO
irf.
£/r# 2 QAO
Barboies laticeps
0.1
FIGURE 12 A combined phylogenetic tree of Sinocyclocheilus fishes constructed using both morphological and molecular data.
ENCYCLOPEDIA OF CAVES
124
c
inbreeding, rapid genetic differentiation, and fixation.
For example, S. anophthalmus and S. maitianheensis
have morphological variation, but their genetic differ¬
ence is small (only 0.36%), suggesting that enormous
and sudden selective pressures led to rapid evolution of
adaptive features. Small population speciation mechan¬
isms probably accelerate and amplify this process, and
have resulted in a particularly rich Sinocyclocheilus spe¬
cies diversity.
The distribution of Sinocyclocheilus in karst environ¬
ments is a consequence of drainage system distribution,
climate conditions, and karst development. As described
above, geographical isolation was probably the most
important factor in speciation; however, the distribution
of the tingi clade in the upstream parts of the
Nanpanjiang River may be the result of diffusion in this
area after the third diversity. And the ijj-type structure
in Guangxi karst landform, a Cenozoic downfaulted
basin and valley, has had a significant effect on this
pattern. The distribution of Sinocyclocheilus was sepa¬
rated coincidentally by both flanks and extension line of
Guangxi u-type structure at different areas. The jii clade
occupies the east part of the Guangxi ojMype structure's
east flank, including the Guijiang river and Hejiang
river; and no distribution records of Sinocyclocheilus exist
between the middle and east wing of Guangxi 'u-type
structure, which is due to the absence of caves in this
area. In contrast, the cyphotergous clade distributes
to the east and midline west areas of the Guangxi
i^-type structure. And the distribution of Sinocyclocheilus
has an obvious relationship to the vertical altitude
of karst from Yunnan to Guangxi. The tingi clade is
mainly distributed in eastern Yunnan Province, the
central eastern plateau, and the upstream parts of
the Nanpanjiang River at altitudes between 1500 m and
2000 m. Downstream, in the middle parts of the
Nanpanjiang River, the altitude of most areas is between
1000 m and 1500 m, and the angularis clade is found in
this region. The cyphotergous clade is located in the far
downstream parts of the river, where elevation falls to
between 500 m and 1000 m. Finally, the altitude where
the jii clade is found is only about 200 m. Thus, the
geographical distribution of Sinocyclocheilus clades has a
vertical basis.
RESEARCH AND CONSERVATION
Hypogean fishes are susceptible to the threats
from habitat degradation, hydrological manipula¬
tions, environmental pollution, overexploitation of
resources, and introduction of alien species. Because
most Chinese cavefish are endemic to small areas
and have small populations, any of these threats
FIGURE 13 The embryonic development of S. gra-
hami. (A) The zygote, a few minutes after fertilization;
(B) 16-cell stage; (C) 64-cell stage; (D) sphere stage; (E)
50% epiboly stage; (F) bud stage; (G) 5-somite stage;
(H) 22-somite stage; (I) 24-somite stage; (J) 28-somite
stage; (K) 42-somite stage; (L) long-pec stage.
ENCYCLOPEDIA OF CAVES
CHEMOAUTOTROPHY
125
could have serious consequences. Currently, some
populations have been considered "threatened" or
"vulnerable." In fact, most Chinese cavefish are cur¬
rently placed in the "vulnerable" status. In 2004,
14 species cavefish were placed on the "Chinese red
list" (Wang Song, Xie Yan (2004). China species red
list. Beijing: Higher Education), including 9 species
of Cyprinidae (8 species of Sinocyclocheilus ) and 5
species of Cobitidae.
Chinese cavefish are also threatened by the rapid
economic growth, which has depleted their habitats
for living and reproduction. To protect these valuable
and unique species of hypogean fish, more research
should focus on the present habitats that still exist,
while practical and effective protective measures
must be established for the karst caves and deep pools
where they live. Protection should also be put into
place so that predation by other animals and humans
is prevented. On April 3, 2008, the first "autonomous
district level cave rare fish natural area" was set up in
Guangxi Lingyun. This was the first natural reserve
area to focus on preserving the unique and rare cave¬
fish of China, and it covers an area 684 hectares
wide, including areas along an underground river
and six distinct caves. Even more promising is the
success of researchers at the Yunnan Endemic Species
Breeding Center, who in 2007 were able to stimulate
artificial reproduction in Sinocyclocheilus grahami
(Fig. 13). The study of Chinese cavefish has drawn
more people's attention to the urgent need for conser¬
vation measures to protect this fascinating group of
highly endangered animals.
See Also the Following Article
Astyanax mexicanus— A Model Organism for
Evolution and Adaptation
Bibliography
Chen, Y., Yang, J., & Zhu, Z. (1994). A new fish of the genus
Sinocyclocheilus from Yunnan with comments on its characteristic
adaptation (Cypriniformes: Cyprinidae). Acta Zootaxonomica Sinica,
19, 246-253.
Huang, Q., Cai, Y., & Xing, X. (2008). Rocky desertification, antide¬
sertification, and sustainable development in the karst mountain
region of southwest China. Ambio, 37, 390—392.
Li, W., & Tao, J. (1994). A new species of Cyprinidae from Yunnan
—Sinocyclocheilus rhinocerous sp. nov. Journal of Zhanjiang Ocean
University, 14, 1—3.
Li, W., Wu, D., Chen, A., & Tao, J. (1997). Histological study on the
horn-like projection of the head of Sinocyclocheilus rhinocerous.
Journal of Yunnan University, 19, 426—428.
Romero, A., Zhao, Y., & Chen, X. (2009). The hypogean fishes of
China. Environmental Biology of Fishes, 86, 211—278.
Wang, D., Huang, Y., Liao, J., & Zheng, J. (1995). Taxonomic revision
of the genus Gibbibarbus Dai. Acta Academiae Medicinae Zunyi, 18,
166-168.
Wang, S., & Xie, Y. (2004). China species red list. Beijing: Higher
Education.
Xiao, H., Chen, S., Liu, Z., Zhang, R., Li, W., Zan, R., & Zhang, Y.
(2005). Molecular phylogeny of Sinocyclocheilus (Cypriniformes:
Cyprinidae) inferred from mitochondrial DNA sequences.
Molecidar Phylogenetics and Evolution, 36, 67—77.
Zhao, Y., & Zhang, C. (2009). Endemic fishes of Sinocyclocheilus
(Cypriniformes, Cyprinidae) in China—Species diversity, cave adapta¬
tion, systematics and zoogeography. Beijing: Science Press.
CHEMOAUTOTROPHY
Annette Summers Engel
University of Tennessee
INTRODUCTION
Photosynthesis is not possible in the dark zone of a
cave. Consequently, the assumption has been that
nearly all life on Earth, and especially life in caves,
depends on organic carbon and energy derived from
photosynthesis. But, reactive rock surfaces and mineral-
rich groundwater provide energy sources for special¬
ized microorganisms that gain cellular energy from
chemical transformations of inorganic compounds—
such as hydrogen, reduced iron, or hydrogen sulfide,
present in groundwater or sediments—and convert inor¬
ganic carbon sources into organic carbon. Essentially,
microorganisms utilize the available chemical energy
that might otherwise be lost to a system. In this manner,
chemosynthesis provides a rich alternative energy
source for organisms, with the results being that ecosys¬
tem biodiversity and population densities are higher
compared to some nonchemosynthetically based ecosys¬
tems that rely on inconsistent and limited inputs of
organic carbon.
This article focuses on chemosynthesis and chemo-
synthetically based ecosystems in caves and karst.
In general, our knowledge of the evolution and metab¬
olism of major chemosynthetic microbial groups has
increased in recent years, partly due to advances in
molecular genetics methods, but also because of
expanded efforts to investigate habitats where chemo¬
synthetic microbes exist. Thanks to research from cave
and karst settings, as well as from research done on
significant chemosynthetic populations at deep-sea
hydrothermal vents (Deming and Baross, 1993) and
within the deep terrestrial subsurface (Stevens and
McKinley, 1995), the "world is green" view is beginning
to change. The article concludes with a discussion of
the importance of subsurface chemosynthetically based
Encyclopedia of Caves.
© 2012 Elsevier Inc. All rights reserved.
126
ecosystems and suggests relevant aspects of future
research in cave and karst systems.
TERMINOLOGY EXPLAINED
The physiological mechanisms for capturing chemical
energy during chemosynthesis are diverse, and there are
several descriptive qualifiers that define an organism
based on its carbon and energy sources. Microbial
metabolism terminology is reviewed in this book's arti¬
cle "Microbes." For the purposes of this article, microbes
that gain energy through chemosynthesis and fix inor¬
ganic carbon are chemolithoautotrophs (literally "self-feed¬
ing rock-eaters"), and energy is gained by transferring
electrons from one chemical (electron donor) to another
(electron acceptor). Typically, chemolithoautotrophs use
compounds present in rocks or groundwater. Chemical
electron donors include, but are not limited to, molecu¬
lar hydrogen, reduced sulfur compounds, metals, and
so on. Organisms that gain cellular energy from chemi¬
cal transformations but use organic carbon compounds
for their carbon source are chemoorganotrophs, and hetero-
trophs use organic carbon for cellular energy and carbon
sources. Several studies have shown that chemolithoau¬
totrophs can grow if organic carbon is present as mixo-
trophs, in which both chemolithoautotrophy and
heterotrophy are expressed simultaneously.
Microorganisms also have oxygen requirements, and
can respire aerobically, anaerobically, or ferment, all of
which relates to electron acceptor utilization. Oxygen is
the terminal electron acceptor for aerobic metabolic pro¬
cesses (Table 1). In reducing environments, microbes
that do not require oxygen ( anaerobes ) use a variety of
alternative electron acceptors for respiration in a
sequence of energetic, reduction reactions that occur
along thermodynamic (and redox) gradients, from
nitrate to carbon dioxide (Table 1). There are several
other energetically favorable, chemolithoautotrophic
pathways that occur in the absence of oxygen, including
TABLE 1 Examples of Chemolithoautotrophic Energy Reactions and Carbon Fixation Pathways
Energy Reaction
Metabolic Process
Major Genera
Electron
Donor
Electron
Acceptor
Carbon
Pathway 1
AEROBES
h 2 + ’/ 2 o 2 ^h 2 o
Hydrogen
oxidation
Alcaligenes, Aquifex
h 2
o 2
Calvin, rTCA
H 2 S + 20 2 -> SOi“ + 2H + or
2S° + 30 2 + 2H 2 0 -► 2H 2 S0 4
Sulfur oxidation
Sulfolobus, Thiobacillus
h 2 s, s°
O 2 , NO 3
rTCA, Calvin,
A-CoA
2 FeS 2 + 70 2 + 2H 2 0 ->• 2FeS0 4 + 2H 2 S0 4
Sulfur oxidation
T. ferrooxidans
S 2 “, FeS 2
o 2
Calvin
4Fe 2+ + 0 2 + 10H 2 O -> 4Fe(OH ) 3 + 8 H+
Iron oxidation
T. ferrooxidans
Fe 2+
O 2 , NO 3
Calvin
2Mn 2+ + 0 2 + 2H 2 0 -»2MnO z + 4H +
Manganese
oxidation
Shewenella
Mn 2+
o 2
Calvin
NHj + 20 2 ->■ NOJ + 2H + + H z O
Ammonium
oxidation
Nitrobacter,
Nitrosomonas
nhJ,no 2
o 2
Calvin
CH 4 + 20 2 -» C0 2 + 2H z O
Methane oxidation
Methanomonas
ch 4
o 2
RMP
ANAEROBES
4H 2 + C0 2 -»CH 4 + 2H z O
Methanogenesis
Methanoseata,
Methanococcus
h 2
co 2
A-CoA
2C0 2 + 4H 2 -> CH 3 COOH + 2H z O
Acetogenesis
Acetobacterium
h 2
co 2
A-CoA
4H 2 + SOT + 2H + ->• H 2 S + 4H 2 0
Sulfate reduction
Archeoglobus
h 2
sor,s 2 or
A-CoA
4H 2 + SOT + 2H + -> H 2 S + 4H 2 0
Sulfate reduction
Desulfovibrio,
Desulfobacter
h 2
sol -
A-CoA, rTCA
FeOOH + 2H + + VzH 2 -> Fe 2+ + 2H z O
Iron reduction
Geobacter, Sheumnella
h 2
Fe 3+ (as
FeOOH)
Calvin
NH 4 + NOJ -> N 2 + 2H 2 0
Ammonium
oxidation
Kuenenia
nh 4
no 2 ,soT
A-CoA
2 NOJ +5H 2 -> N 2 + 4H 2 0 + 20H“
Denitrification
T. denitrificons,
h 2/ h 2 s, s n
NO 3
Calvin, A-CoA
Pyrolobus
1 A-CoA = Acetyl-CoA pathway; rTCA = reductive tricarboxylic acid cycle; Calvin = Calvin—Benson cycle; RMP = Ribidose monophosphate cycle
ENCYCLOPEDIA OF CAVES
CHEMOAUTOTROPHY
127
sulfide oxidation via nitrate reduction and anaerobic
ammonium oxidation (anammox) (Table 1).
CARBON FIXATION PATHWAYS
Chemolithoautotrophy has been crucial in the trans¬
formation of chemical energy from one reservoir to
another during Earth's history. With photoautotrophy,
these two processes are the only mechanisms by which
the inorganic carbon is incorporated into the biosphere
(referred to as primary productivity). Most extant che¬
molithoautotrophy is accomplished by one of at least
four pathways: the Calvin—Benson—Bassham (CBB)
cycle (also known as the reductive pentose phosphate
cycle), the acetyl coenzyme A (acetyl-CoA) (also
known as the Wood—Ljungdahl) cycle, the reductive
tricarboxylic acid cycle (rTCA) (or Arnon), and the
3-hydroxypropionate cycle. Based on evolutionary
relationships of the 16S (small-subunit) ribosomal
RNA (rRNA) gene and the subsequent "Tree of Life"
(Woese, 1987), autotrophic carbon fixation pathways
are distinctly related to the phylogenetic position of
the respective organisms. Only members from the
domains of Bacteria and Archaea are capable of che¬
molithoautotrophy (i.e., plants are photoautotrophs).
Some researchers have speculated that the last com¬
mon ancestor between Bacteria and Archaea used
either the acetyl-CoA pathway or the rTCA cycle, with
the rTCA cycle most closely resembling what is con¬
sidered to be the first autotrophic pathway. Moreover,
chemosynthesis probably preceded photosynthesis on
Earth because none of the ancestral lineages are photo-
trophic (Schidlowski, 2001). Although both the acetyl-
CoA and rTCA pathways are used by diverse Bacteria
and Archaea, such as some sulfate-reducers and aceto-
genic bacteria, as well as methanogens (Archaea), little
is known about the ecology and evolution of the rTCA
cycle and the microbes capable of this important path¬
way. Instead, some of the most deeply branching bac¬
teria on the "Tree of Life" are thermophilic bacteria
Aquificales and Thermotogales, including the genera
Aquifex and Hydrogenobacteria, and fix carbon via the
CBB cycle, whereas the archaeal thermophiles, such as
Thermoproteus and Pyrodictium, fix carbon via the rTCA
cycle. The CBB cycle is found in many obligate and
facultative autotrophic microbes and is the dominant
pathway in the Eukarya ( e.g ., plants). Many research¬
ers have postulated that the key enzyme in the
CBB pathway, ribulose-bisphosphate carboxylase
(RuBisCo), is the most abundant protein on Earth. The
3-hydroxypropionate cycle was considered to be rele¬
gated to only one phylum of bacteria, the Chloroflexi,
but now it has been detected from Archaea and seems
to be more important than previously considered.
MAJOR CHEMOLITHOAUTOTROPHIC
MICROBIAL GROUPS
Electron donors are essential for chemolithoauto-
trophic growth. For microbes other than chemolithoauto-
trophs (e.g., heterotrophs, chemoorganotrophs), organic
carbon molecules are used as electron donors to gain cel¬
lular energy. One of the most energetic chemolithoauto-
trophic electron donors is molecular hydrogen (H 2 ).
Microbes that use H 2 also use C0 2 as an electron accep¬
tor, either to make methane (methanogenesis) or acetate
(acetogenesis). Sulfate-reducing microbes also compete
for H 2 , but some can use small organic molecules as elec¬
tron donors. Hydrogen- and ammonia-oxidizing bacteria
utilize H 2 aerobically. Additional electron donors include
reduced sulfur compounds (e.g., hydrogen sulfide, ele¬
mental sulfur), ammonia (NH 4 ), and reduced iron (Fe 2+ )
and manganese (Mn 2+ ). The following sections describe
chemolithoautotrophic microbial groups based on
required electron donors (Table 1 ).
Hydrogen
H 2 , hydrogen gas, accumulates from the anaerobic
breakdown of organic molecules by fermentative (het-
erotrophic) bacteria, and serves as an important energy
source for aerobic and anaerobic archaeal and bacterial
chemolithoautotrophs. In most cases, H 2 is rapidly
consumed under anaerobic conditions by methanogens
and sulfate-reducers. However, if anaerobic growth is
slower than production of H 2 by heterotrophs or che¬
moorganotrophs, then H 2 will diffuse into the aerobic
environment where it can be used by aerobic ^-utiliz¬
ing microorganisms. Chemolithoautotrophic aerobic
H 2 -oxidizing microbes oxidize H 2 as the electron
donor by a membrane-bound hydrogenase (enzyme),
oxygen as the electron acceptor, and fix C0 2 via the
CBB or rTCA cycle (Table 1). A wide variety of gram¬
negative and gram-positive bacteria are H 2 -oxidizers,
including Alcaligenes, Hydrogenobacter, Pseudomonas, and
Aquifex.
The anaerobic microbial groups that require H 2 as
an electron donor follow predictable redox chemistry
regarding the utilization of available terminal electron
acceptors. Reduction of nitrate, iron, or manganese
during anaerobic respiration is typically done by het¬
erotrophs or chemoorganotrophs, but iron or manga¬
nese reduction can also be accomplished by some
chemolithoautotrophs. Ferric iron is used as an exclu¬
sive electron acceptor by dissimilatory iron-reducing
microbes that use H 2 as their electron donor, although
some require organic carbon compounds, such as ace¬
tate or formate. One of the most studied iron-reducers,
Shewanella putrefaciens, is a chemoorganotroph, and can
ENCYCLOPEDIA OF CAVES
128
c
also use Mn0 2 as a sole electron acceptor. Biological
iron reduction in aquatic sediments is an important
anaerobic process.
Dissimilatory sulfate-reducing microbes are obligate
anaerobes that reduce sulfate to hydrogen sulfide (H 2 S)
(Table 1). Many sulfate-reducers are heterotrophic or
chemoorganotrophic, divided into two broad physiolog¬
ical subgroups. Group I dissimilatory sulfate-reducers,
including the genera Desulfovibrio, Desulfomonas, and
Desulfobulbus, use lactate, pyruvate, ethanol, or other
fatty acids as carbon sources. Group II genera oxidize
acetate, and include Desulfococcus, Desulfosarcina,
Desulfobacter, and Desidfonema. These bacterial groups
are found within the Deltaproteobacteria class of the
Proteobacteria phylum, although there are sulfate-
reducing Archaea that belong to the genus Archeoglobus.
Certain species are also capable of growing chemo-
lithoautotrophically with H 2 as the electron donor,
sulfate as the electron acceptor, and C0 2 as the sole
carbon source. Chemolithoautotrophs use the acetyl-
CoA pathway, while the rTCA cycle is used by che-
moorganotrophs. If sulfate concentrations are high, then
sulfate-reducing bacteria completely oxidize fermenta¬
tion by-products to C0 2 . In low-sulfate anaerobic envir¬
onments, however, sulfate-reducing bacteria compete
with methanogens for H 2 and organic compounds.
The reduction of elemental sulfur is also possible, and
chemolithoautotrophic sulfur-reducing Archaea include
Thermoproteus, Acidianus, and Desulfurolobus, which are
common in acidic environments. Although not exclusively
a chemolithoautotrophic pathway, the disproportion¬
ation of elemental sulfur is also possible, whereby ele¬
mental sulfur is split into two different compounds,
typically sulfite and thiosulfate, to produce both H 2 S
and sulfate. Desirtfocapsa is a common sulfur-dispropor-
tionating bacterial genus.
Acetogenesis results in less overall cellular energy
than methanogenesis, but both metabolic groups are
found in similar habitats. Acteogenic bacteria, such as
Clostridium and Acetobacterium, are obligate anaerobes
that form acetate from the oxidation of H 2 using the
acetyl-CoA pathway for C0 2 fixation. These organisms
are known as homoacetogens. H 2 is the common electron
donor, but donors can also be from sugars, organic
acids, and amino acids. Many acetogens also reduce
nitrate and thiosulfate, and are tolerant of low pH,
thereby making them more versatile than other
anaerobes.
Methanogens are exclusively Archaea, and are one of
the most common anaerobic microbes in highly reducing
conditions in close association with decomposing organic
material. Methanogens oxidize H 2 as the electron donor
while reducing C0 2 to methane using the acetyl-CoA
pathway (Table 1). These methanogens are usually found
in close association (syntrophy or interspecies hydrogen
transfer) with fermenting microbial groups because of
the need for the continuous H 2 supply provided through
fermentation. Chemolithoautotrophic methanogenesis is
less efficient than those utilizing formate, acetate, metha¬
nol and other alcohols, carbon monoxide, or even ele¬
mental iron, as alternative electron donors. The
physiologies of these different types of methanogens
(chemolithoautotrophic versus chemoorganotrophic) are
distinct, and there are seven major groups of methano¬
gens based on this physiology, including the genera
Methanobacterium, Methanosaeta, Methanococcus, and
Methanolobus.
Reduced Inorganic Sulfur Compounds
High concentrations of reduced sulfur compounds
are toxic to most organisms. H 2 S gas, in particular,
reacts with operational biomolecules to form nonfunc¬
tional complexes that inhibit respiration. However,
H 2 S, elemental sulfur, thiosulfate, polythionates, metal
sulfide, and sulfite can serve as electron donors for
chemolithoautotrophic sulfur-oxidizing bacteria and
Archaea (note: for the purposes of this discussion, any
microbe capable of oxidizing any reduced sulfur com¬
pound is referred to as a sulfur-oxidizer) (Table 1).
One difficulty for sulfur-oxidizing microbes is compet¬
ing with chemical oxidation of reduced sulfur com¬
pounds. As a result, most sulfur-oxidizing bacteria
occupy aerobic/anaerobic interfaces where sulfide and
oxygen meet, and are therefore chemotactic. Interface
or gradient growth is important in some ecosystems
where sulfur-oxidizing bacteria form symbiotic rela¬
tionships with animals, such as at deep-sea hydrother¬
mal vent sites.
The earliest microbiological research regarding
chemo-lithoautotrophic metabolism was done with
sulfur-oxidizing bacteria in the late 1880s. Some sulfur-
oxidizers thrive in low pH environments (acidophiles),
while others require neutral pH conditions. Sulfur-
oxidizing bacteria can form large microbial mats or
thick biofilms in a range of habitats, including acid
mine drainage and mine tailings, sulfidic thermal
springs, marine sediments, and sewage sludge. There
are several different types of sulfur-oxidizers, including
those belonging to the genus Thiobacillus and the
"morphologically conspicuous sulfur bacteria" such as
Beggiatoa, Thiothrix, Thiomicrospim, and Thiovulum,
among others. Virtually all of these colorless sulfur bac¬
teria (as compared to the "green" and "purple" sulfur
bacterial groups that are anoxygenic phototrophs) are
gram-negative and deposit intracellular elemental
sulfur when grown on H 2 S. Although most sulfur-
oxidizing bacteria are chemolithoautotrophic, using
either the CBB or rTCA cycles for carbon fixation, some
can also be chemoorganotrophic, heterotrophic, or
ENCYCLOPEDIA OF CAVES
CHEMOAUTOTROPHY
129
mixotrophic with respect to their carbon source.
Typically, molecular oxygen is the terminal electron
acceptor, but some sulfur-oxidizers use nitrate as an
electron acceptor ( e.g ., Thiobacillus denitrificans).
Ammonia and Nitrite
Nitrification includes the oxidation of both ammonia
and nitrite, and results in the formation of nitrite and
nitrate, respectively. The formation of nitrate requires a
two-step process in which ammonia is first oxidized to
nitrite by ammonia-oxidizers, and then nitrite is oxi¬
dized to nitrate by nitrite-oxidizers (Table 1). The two
groups of nitrifying bacteria, belonging to the family
Nitrobacteraceae, typically work synchronously in a habi¬
tat as chemolithoautotrophs that use the ammonia or
nitrite as electron donors and CO 2 as the sole source of
carbon. In some rare instances, small organic com¬
pounds can also be used. Genera of ammonia-oxidizers
have the prefix Nitroso-, including Nitrosomoms and
Nitrosospira, whereas genera of nitrite-oxidizers have
the prefix Nitro-, including Nitrobacter and Nitrospira.
Nitrification occurs at the aerobic/anaerobic interface,
but nitrifying bacteria have a high affinity for oxygen.
Microbiological ammonia and nitrite oxidation can
occur over a broad range of conditions, including in
acid soils, and at low and high temperatures.
The oxidation of ammonia anaerobically, referred to
as anammox, involves coupling ammonia oxidation to
nitrite reduction under anaerobic conditions (Table 1).
The end product is nitrogen gas (N 2 ), but an intermedi¬
ate by-product of this reaction is hydrazine (rocket fuel)
that is contained in unique anammoxasomes, or special¬
ized intracellular organelles. If ammonia is limited,
these microbes can use organic acids to convert nitrate
and nitrite into N 2 . Anammox microbes fix carbon via
the acetyl-coA pathway and have been identified from
marine and freshwater habitats. So far, anammox
microbes belong to the Planctomycetes phylum with
four provisionally described bacterial genera from fresh¬
water ( Brocadia , Kuenenia, Anammoxoglobus, and Jettenia),
and Scalindm from marine habitats.
Iron and Manganese
As with sulfur oxidation, microbes can also gain
energy from ferrous iron (Fe 2+ ) or manganous manga¬
nese (Mn 2+ ) oxidation. In the past, Fe 2+ - and Mn 2+ -oxi¬
dation was attributed to many different microbes based
on the accumulation of iron or manganese minerals
associated with cellular material. However, recent
investigations have found that most of these microbes
do not gain energy from mineral buildup. Iron rapidly
oxidizes to ferric (Fe 3+ ) iron at neutral pH, so successful
iron-oxidizers live in low pH environments, such as
acid mine drainage, acid springs, mine tailings, or acid
soils containing sulfide minerals such as pyrite.
Chemolithoautotrophic iron-oxidizing bacteria include
the acidophilic aerobe Thiobacillus ferrooxidans. Despite
research attempting to isolate other chemolithoauto¬
trophs, the most common iron-oxidizers, Leptospirillum
and Gallionella, and some manganese-oxidizers, such as
Arthrobader and Hyphomicrobium (except H. manganoxi-
dans), are chemoorganotrophs. Ecologically, iron- and
manganese-oxidizers are important for detoxifying the
environment by lowering the concentration of dis¬
solved toxic metals.
One-Carbon Compounds
There has been some controversy regarding whether
or not microbial processes based on reduced one-carbon
(C-l) compounds, as the sole carbon and energy source,
represent true chemolithoautotrophy; C-l compounds
used as electron donors are biogenic, thereby not
fulfilling the litho component to chemolithoautotrophy.
However, cycling C-l compounds by potential chemo¬
lithoautotrophs is gaining appeal because understand¬
ing CO 2 and CH 4 dynamics is important for carbon
budgets of aquatic ecosystems in surface and subsur¬
face environments.
Methanotrophs and methylotrophs utilize methane
or methanol, respectively, for cellular energy and
growth under aerobic conditions (Table 1). Aerobic
methanotrophs sequentially oxidize methane to a
series of intermediates, including methanol, formalde¬
hyde, and formate, before generating CO 2 enzymati¬
cally with methane monooxygenase. Some of the
intermediates can be incorporated into the cell, which
is one of the reasons why C-l metabolism is not con¬
sidered to be strictly autotrophic. There are two
physiological groups of methanotrophic bacteria,
both obligate aerobes. Type I methanotrophs include
the genera Methylomonas, Methylosphaera, and
Methylomicrobium within the Gammaproteobaderia, and
fix carbon by assimilating formaldehyde in the ribu-
lose monophosphate pathway. Type II methanotrophs
fix formaldehyde in the serine pathway, and genera
include Methylocystis and Methylosinus within the
Alphaproteobaderia. Methanotrophy had been previ¬
ously associated with neutrophilic pH conditions, but
bacteria from the phylum Verrucomicrobia have been
described recently that oxidize methane below pH 5
and optimally around pH 2.
Methane can also be oxidized anaerobically by a
syntrophic consortium of bacteria and anaerobic
methanotrophic (ANME) Archaea, referred to as
anaerobic oxidation of methane or AOM. Sulfate is the
electron acceptor for sulfate-reducing bacteria. ANMEs
ENCYCLOPEDIA OF CAVES
130
c
are closely related to methanogenic Archaea, and
AOM is the enzymatic reverse of methanogenesis. This
process has been identified from marine and freshwa¬
ter lake sediments and water columns. Anaerobic oxi¬
dation of methane with nitrite as the electron acceptor
has recently been identified by the new microbe
"Candidatus Methylomirabilis oxifera" in the absence
of a syntrophic association.
Growth on carbon monoxide by carboxidobacteria
(carbon monoxide-oxidizing bacteria) occurs with a
range of electron acceptors, including oxygen, elemen¬
tal sulfur, sulfate, nitrate, or ferric iron. Members of
the genera Pseudomonas, Bacillus, Alcaligenes, and
Clostridium are known for carbon monoxide oxidation.
CHEMOLITHOAUTOTROPHY IN CAVES
AND KARST SETTINGS
Cave ecosystems have been considered to be both
energy- and nutrient-limited because organic matter,
assumed to be derived primarily from photosyntheti-
cally produced material on the surface, needs to be
brought into the subsurface. Knowledge of chemo-
lithoautotrophic microbial processes, however, has
changed our understanding of the types of ecosystem-
level processes that can occurs when surface input of
organic matter is limited. Studies of chemolithoauto-
trophic processes in caves have been done since the
1960s, with intensive efforts since the mid-1990s.
At a glance. Table 2 shows that the sulfur cycle in
karst has been intensively studied. Research efforts have
been sustained for decades for some systems, including
the Movile Cave (Romania), Frassasi Caves (Italy), and
Lower Kane Cave (Wyoming, U.S.A.), but several new
systems have been discovered and extensively investi¬
gated (Table 2). To be fair, much of the recent research
has been done in systems most likely to harbor chemo-
lithoautotrophic microbes (Table 2), such as those
having sulfidic or methanogenic waters, metal-laden
sediments, or obvious microbial mats and biofilms
formed under extreme geochemical conditions.
In general, most studies have provided a census of
microorganisms and their metabolic processes, with
very few ecosystem-level studies that attempt to inte¬
grate the activities of different microorganisms with
respect to each other, or studies of cave or karst sys¬
tems where chemolithoautotrophy may not be a domi¬
nant metabolism. From recent studies, it is clear that
more diverse chemolithoautotrophic microbial groups
are present than previously considered. The following
section summarizes our current knowledge of chemo¬
lithoautotrophic microbial groups in cave and karst
settings, divided into the sulfur, nitrogen, methane,
and metal (Fe, Mn) cycles.
Sulfur Cycle
Caves with active microbial sulfur cycling typically
have Fl 2 S-rich springs that discharge into passages,
termed sulfidic caves. Because of H 2 S oxidation, either
abiotically or biotically, sulfuric acid is generated and
promotes local limestone bedrock dissolution. This
cave formation process is known as sulfuric acid speleo-
genesis, described in detail from Lower Kane Cave by
Egemeier (1981) (Fig. 1). These types of caves, the
microbial communities, and geochemical and geologi¬
cal processes have been reviewed recently by Engel
(2007).
Prior to and just after the 1986 discovery of the
Movile Cave ecosystem, there were few studies
describing sulfur-based microbial populations in caves.
Sulfur-oxidizing bacteria occur in both aqueous and
subaerial habitats, with filamentous groups forming
thick microbial mats, rope-like structures, biofilms,
and webs within cave spring orifices, pools, or
streams. In Movile Cave, there is a thin microbial mat
that floats at the water surface (Fig. 2) and an anaero¬
bic biofilm mixed with clay that covers the submerged
cave floor. For most studies, the main goals have been
to understand the microbial diversity and spatial dis¬
tribution of microbes, but rates of primary productivity
have been quantified from a few systems (Table 2).
Recent advances in integrative isotopic, molecular, and
microscopy techniques are revealing how higher level
organisms in the cave ecosystems are dependent on,
and even in potential symbiotic associations with,
chemolithoautotrophic sulfur-oxidizing microbes (e.g.,
Dattagupta et al., 2009).
16S rRNA gene sequence surveys have certainly
increased our understanding of the microbial diversity
from these systems. Phylogenetic analyses from several
of the cave systems reveal that sulfur-oxidizing bacteria
closely related to the genera Thiovulum, Thiothrix,
Tlriobacillus, Beggiatoa, and Thiomicrospira occur. However,
unclassified (due to the inability to culture these groups
to date) microbial groups belonging to the Epsilon-proteo-
bacteria class of the Proteobacteria phylum dominate
many of the microbial communities in the sulfidic sys¬
tems. Sequences from many other different bacterial
lineages, including candidate divisions, have also been
retrieved from sulfidic karst systems, suggesting a richer
microbial diversity than previously considered. Sulfate-
reducing and sulfur-disproportionating microbes have
also been identified from 16S rRNA gene surveys and
culture-based methods of water and sediments (Table 2).
Sulfate-reducers are important for recycling sulfur com¬
pounds by generating H 2 S that sulfur-oxidizers can use.
In cave streams, this interactions among microbes across
redox gradients in microbial mats, which are cycling ele¬
ments from oxidized to reduced forms (e.g., sulfate to
ENCYCLOPEDIA OF CAVES
CHEMOAUTOTROPHY
131
TABLE 2 Chronological Summary of Studies Describing Chemolithoautotrophic Microorganisms Found in Various Caves Worldwide
Cave or Karst System
Date(s) of
Study 1
Chemolithoautotrophic Metabolic Processes 2
Approach 3
Caves in France and Iowa (U.S.A.)
1963, 1987
Iron oxidation
M, C
Florida Aquifer karst, U.S.A.
1965,
1992-1999;
2006
Sulfur oxidation; methanogenesis and methanotrophy
M, C, DNA, BA, SI
Bungonia Caves, Australia
1973-1994
Iron reduction, sulfate reduction, sulfur oxidation
M, C
Mammoth Cave, Kentucky, U.S.A.
1977; 2009
Ammonium and nitrite oxidation
C
Edwards Aquifer, Texas, U.S.A.
1981,1986;
2005-2011
Sulfur oxidation; methanogenesis and methanotrophy; sulfate
reduction, ammonium oxidation, nitrite oxidation, anaerobic
ammonium oxidation
SI, DNA, FG
Movile Cave, Romania
1986-2011
Sulfur oxidation, methanotrophy, methanogenesis, ammonium
oxidation, nitrite oxidation
M, C, SI, DNA, FG,
BA, RI, nanoSIMS
Cesspool Cave, Virginia, U.S.A.
1986, 2001;
2009
Sulfur oxidation; sulfate reduction
M, C, DNA, RI
Grotta Azzurra, Italy
1986-2007
Sulfur oxidation, sulfate reduction
C, M, SI
Parker Cave, Kentucky, U.S.A.
1988,1998
Sulfur oxidation
M, DNA
Karst springs, Spitsbergen
1994
Sulfate reduction, sulfur oxidation
M, SI
Cueva de Villa Luz, 4 Mexico
1994-2007
Sulfur oxidation, sulfate reduction
M, C, DNA
Kugitangtou Caves, Turkmenistan
1994, 1997,
2001
Sulfur oxidation, sulfate reduction, iron oxidation
M, C, DNA
Zoloushka Cave, Ukraine
1994, 2001
Sulfate reduction, denitrification, sulfur oxidation,
iron oxidation
C, SI
Frasassi Caves, 5 Italy
1996-2011
Sulfur oxidation, sulfate reduction
M, C, SI, DNA, RI,
nanoSIMS
Anchihaline caves, Mexico
1997
Ammonium oxidation, sulfur oxidation, methanotrophy
SI
Maltravieso Cave, Spain
1997
Ammonium oxidation
C
Bundera Sinkhole, Australia
1999
Sulfur oxidation
SI
Lower Kane Cave, Wyoming,
U.S.A.
1999-2011
Sulfur oxidation, sulfate reduction, iron reduction,
methanogenesis, iron oxidation
M, C, SI, DNA, FG,
BA, RI, FISH
Glenwood Cavern, Colorado,
U.S.A. (and nearby mine adit)
2005-2007
Sulfur oxidation, hydrogen oxidation, ammonium oxidation
M, DNA, FG
Ayalon Cave, Israel
2006-2008
Sulfur oxidation, methanotrophy, methanogenesis
M, SI
Acquasanta Terme caves, Italy
2009-2011
Sulfur oxidation, sulfate reduction
DNA, FISH
Sistema Zacaton, Mexico
2007-2011
Sulfur oxidation, sulfate reduction, methanogenesis, anaerobic
ammonium oxidation
M, DNA, FG
Coldwater Cave, Iowa and
Minnesota (U.S.A.)
2010
Ammonium oxidation, nitrite oxidation, anaerobic ammonium
oxidation
C, DNA, FG
1 Representative years for which published results can be found in the literature, thus far.
2 Listed in order of dominance, if known.
3 Approach used to identify and to verify chemolithoautotrophic processes. M = microscopy; C = culture techniques; SI = stable isotopes; DNA = phylogenetic relationships based
on 16S rRNA analyses; FG = functional gene surveys; BA = biomarker assays; RI = radioisotope experiments; FISH = fluorescence in situ hybridization; nanoSIMS = nano¬
secondary ion mass spectroscopy. For a description of the approaches, refer to this book's article "Microbes."
4 Also known as Cueva del Azuffre or Cueva de las Sardinas.
includes the Grotte di Frasassi and Grotta Sulfurea systems.
ENCYCLOPEDIA OF CAVES
132
c
FIGURE 1 Main trunk passage in Lower Kane Cave, Wyoming
(U.S.A.). White, filamentous microbial mats dominated by sulfur-
oxidizing bacteria colonize shallow sulfidic water, beginning at the
lower right corner (water flows from the lower right to upper left).
This microbial mat extends for approximately 20 m, with an average
thickness of 5 cm. Piles of gypsum surround the stream (especially on
the left), formed from the replacement of limestone during sulfuric
acid speleogenesis.
FIGURE 2 Partially flooded passage in Movile Cave, Romania. A
white filamentous microbial mat floats on the surface of the water. A
grid was constructed to encourage additional mat growth.
sulfide, then sulfide to sulfur to sulfate, or C0 2 to bicar¬
bonate to organic carbon and back to C0 2 ), as well as
involving abiotic and biotic transformation processes,
has been proposed as an example of nutrient spiraling
of sulfur (and carbon).
Biofilms on subaerial cave-wall surfaces, referred to as
microbial draperies, cave-wall biofilms, or snottites, have
been described from the several active sulfidic caves.
Although some of the longest draperies reportedly reach
up to 10 cm in length or more, cave-wall biofilms typi¬
cally occur as discontinuous patches of insoluble crusts
with subcentimeter long, mucus-like droplets suspended
from the crusts (Fig. 3). Condensation droplets associ¬
ated with the biofilms have pH values between 0 and 3.
FIGURE 3 Discontinuous cave-wall biofilms and crusts (dark
patches) from Lower Kane Cave, Wyoming. Crusts form on gypsum
(light areas), and condensation droplets and mucus-like drops are
suspended from the crusts. Elemental sulfur is typically associated
with the crusts.
Acidity is maintained, in part, because the replacement
of carbonate with gypsum in the sulfuric acid speleogen¬
esis process that physically separates the cave-wall solu¬
tions from the underlying carbonate that would buffer
pH to more neutral conditions, and also because of con¬
tinued microbial sulfur oxidation. Laboratory isolates
and phylogenetic studies demonstrate that microbes
are closely related to extremophiles, including those
belonging to the genera Acidithiobacillus, Sulfobacilhis,
and Acidimicrobium, as well as the Thermoplasmales,
actinomycetes, and the bacterial candidate lineage TM6.
Filamentous fungi and protists have also been described
from cave-wall biofilms.
Nitrogen Cycle
Nitrogen enters the cave in various forms, including
as atmospheric gas, nitrates from fertilizers, organic
matter in soil, and from bat and rat guano. Ammonium
is produced by nitrogen-fixing bacteria, which would
be present in soils overlying limestone. The ammonium
could be brought into the caves via bedrock fissures or
washed in with stream sediments. Many anaerobic
microbes (e.g., sulfate-reducers) can also fix nitrogen.
Nitrosomonas oxidizes ammonia to nitrite, and
Nitrobacter oxidizes nitrite to nitrate. This two-step pro¬
cess can influence nitrate mineral precipitation, such as
niter and nitrocalcite (saltpeter). Whether microbes are
involved in the precipitation of these minerals or not is
still debated. Nitrate minerals are typically found in dry
cave sediments, such as in the southeastern United
States. Isolates of Nitrobacter have been obtained from
Mammoth Cave (Kentucky) sediments rich in saltpeter
and suggest a geomicrobiological link, but more work is
needed to confirm this.
ENCYCLOPEDIA OF CAVES
CHEMOAUTOTROPHY
133
Nitrification is an important chemolithoautotrophic
pathway in nitrogen-limited freshwater streams and
lakes because it enhances overall nitrogen availability.
Limited work has been done from caves. One study
based on stable nitrogen isotope ratio analysis from
anchihaline caves in Mexico indicates that nitrification
may be the primary energy source to that ecosystem,
in addition to sulfur- and methane-oxidation. Recently,
from research done in a mine adit at Glenwood Springs,
Colorado, ammonium oxidation by Crenarchaea may be
occurring (Table 2). Characterization of nitrogen cycling
from sediments impacted by atrazine (sourced from
agricultural land runoff) in Coldwater Cave, Iowa and
Minnesota, indicates that atrazine may negatively affect
nitrogen cycling, and cause an increase in nitrification
at the expense of nitrogen fixation (Iker et al., 2010).
Methane Cycle
Methane cycling is beginning to interest karst
researchers because of its potential significance to eco¬
system development and primary productivity in
some cave systems. Methanotrophy and methanogen-
esis, while expected to occur in groundwater have
only been studied from a few cave and karst systems,
likely because of a limited understanding of what the
dissolved methane concentrations are. Studies, particu¬
larly in Movile Cave, have been based on enrichment
and isolation techniques for methanotrophs and
methanogens, 13 CH 4 -labelling incorporation, and DNA
analysis following separation of 12 C from 13 C (a pro¬
cess referred to as stable isotope probing) (Table 2).
Very low carbon isotope values for dissolved inorganic
and organic carbon in Movile Cave, groundwater from
the Floridan Aquifer (Opsahl and Chanton, 2006), and
groundwater from the Edwards Aquifer in Central
Texas (Longley, 1981), are suggestive of active metha¬
notrophy. Moreover, higher trophic levels (e.g., ani¬
mals) have isotopic values that indicate their food
sources were organic carbon produced from these
microbial processes.
Iron and Manganese Bacteria
According to Caumartin (1963), iron plays a signifi¬
cant role in the microbiology of a cave, and most cave
sediment contains iron bacteria. Indeed, microbes asso¬
ciated with iron and manganese cycling have been
found in cave sediments and speleothems based on
microscopy and culturing, with chemolithoautotrophic
Gallionella, and heterotrophic Leptothrix and Crenothrix
being the most common genera described. Fossil or
encrusted sheaths and stalks of these bacteria have
been observed in stalactites, sediments, and corrosion
crusts from several caves, including caves from Iowa,
the Black Hills (South Dakota), Lechuguilla Cave (New
Mexico), Kugitangtou caves (Turkmenistan), Grand
Caymen Island caves, the Nullarbor Plain caves in
Australia, as well as others. However, it is difficult to
distinguish between chemical and biological iron/
manganese precipitates, and it is not evident from
mineralization, fossil or active, whether an organism
was chemolithoautotrophic. Therefore, the ecology of
metal cycling remains speculative and the roles of
these microbial groups as potential primary producers
remain unclear.
CHEMOLITHOAUTOTROPHICALLY-
BASED CAVE ECOSYSTEMS
Although chemolithoautotrophic metabolism was
described from sulfidic springs more than 100 years ago,
it has been estimated that chemolithoautotrophs do not
produce efficient energy, and therefore, could not be sig¬
nificant primary producers for an ecosystem when com¬
pared to photosynthetically produced energy. However,
in the subsurface, photosynthesis is impossible. For
research done in deep-sea hydrothermal vents, the deep
terrestrial subsurface, and caves and karst aquifers, which
can also be restricted from an influx of organic matter
from the surface, the significance of chemolithoautotro-
phy to ecosystem-level processes is being revised.
Part of the revision is because of the extremely rich
biodiversity in these ecosystems. From marine environ¬
ments, chemolithoautotrophically based ecosystems
include deep-sea hydrothermal vents (Deming and
Baross, 1993), hydrocarbon cold seeps, and estuarine
sediments. Gammaproteobacteria and Epsilonproteobacter-
ia are in chemosymbiotic associations with animals,
such as the vestimentiferan tube worms ( Riftia pachypila)
and clams ( Calyptogena magnified). Deep-sea sediments
exposed to hydrocarbon seeps have mussels with
methanotrophic symbionts, and some mussels ( Bathy-
modiolus thermophilus) also have sulfur- and methane-
oxidizing bacterial symbionts. Clams, including Sole-
mya and Thyasira, living in brackish water and estua¬
rine sediments have sulfur-oxidizing symbionts. The
relationships between chemolithoautotrophic microor¬
ganisms and animals are beneficial to both organisms,
as the animals provide protective habitat and the bac¬
teria supply nutrients and energy.
From caves and karst aquifers, our knowledge of
chemosymbiotic associations in freshwater systems is
limited. But, the recent discovery of a potential chemo¬
symbiotic association between the sulfur-oxidizing bac¬
teria Thiothrix spp. ( Gammaproteobacteria ) and Niphargus
spp. (amphipod) in the Frasassi Caves (Italy)
(Dattagupta et al., 2009) provides an incentive to keep
investigating other possible associations. The highly
ENCYCLOPEDIA OF CAVES
134
c
diverse Movile Cave ecosystem, with 33 new cave-
adapted taxa identified from 30 terrestrial invertebrate
species (24 are endemic) and 18 species of aquatic ani¬
mals (9 endemic), is another cave system to investigate
possible chemosymbiotic associations (Sarbu et al.,
1996). The recent discovery of the Ayalon Cave ecosys¬
tem in Israel, with possibly a new sulfur-based, che-
mosynthetic ecosystem and eight invertebrates new to
science, is testimony that substantial and unknown
subterranean diversity in karst still exists. Other
opportunities to expand our knowledge of chemo-
lithoautotrophically based karst systems include the
Edwards Aquifer, where the role of chemolithoautotro-
phy is just beginning to be investigated, as well as the
Upper Floridan aquifer.
CHEMOLITHOAUTOTROPHY
IN THE FUTURE
Caves are important and relatively accessible habi¬
tats to study chemolithoautotrophic metabolism. In the
absence of organic carbon compounds, reactive rock
surfaces and mineral-rich groundwater in the subsur¬
face provide an assortment of potential energy sources
for chemolithoautotrophic microorganisms. In compar¬
ison to the research done with macroscopic organisms
in caves, the distribution of chemolithoautotrophic
microbial groups and associated biogeochemical pro¬
cesses occurring in caves have infrequently been
addressed, although the methodology exists. Of those
investigations describing cave microbial communities,
many have been simply observational and the physio¬
logical capabilities of microorganisms were inferred
based on the utilization of a specific substrate in cul¬
tures or from 16S rRNA gene sequence phylogenetic
affiliation. To date, there have been few stable isotope
or radioisotope studies done to describe chemolithoau-
totrophy in cave ecosystems, and combined micros¬
copy and molecular studies have just recently been
initiated that should continue to expand our under¬
standing of chemolithoautotrophic processes in the
subsurface. From these investigations, potential che-
mosymbiosis between microbes and animals provide
exciting future research opportunities.
See Also the Following Articles
Microbes
Cave Ecosystems
Bibliography
Caumartin, V. (1963). Review of the microbiology of underground
environments. National Speleological Society Bulletin, 25(1), 1—14.
Dattagupta, S., Schaperdoth, I., Montanari, A., Mariani, S., Kita, N.,
Valley, J. W., et al. (2009). A novel symbiosis between chemoauto-
trophic bacteria and a freshwater cave amphipod. The ISME
Journal, 3(8), 935-943.
Deming, J., & Baross, J. (1993). Deep-sea smokers: Windows to a
subsurface biosphere? Geochimica et Cosmochimica Acta, 57(14),
3219-3230.
Egemeier, S. J. (1981). Cave development from thermal waters.
National Speleological Society Bulletin, 43, 31—51.
Engel, A. S. (2007). Observations on the biodiversity of sulfidic karst
habitats. Journal of Cave and Karst Studies, 69(1), 187—206.
Iker, B. C., Kambesis, P., Oehrle, S. A., Groves, C., & Barton, H. A.
(2010). Microbial atrazine breakdown in a karst groundwater sys¬
tem and its effect on ecosystem energetic. Journal of Environmental
Quality, 39(2), 509-518.
Longley, G. (1981). The Edwards Aquifer—Earth's most diverse
groundwater ecosystem. International Journal of Speleology, 11
(1-2), 123-128.
Opsahl, S. P., & Chanton, J. P. (2006). Isotopic evidence for methane-
based chemosynthesis in the Upper Floridan aquifer food web.
Oecologia, 150(1), 89—96.
Sarbu, S. M., Kane, T. C., & Kinkle, B. K. (1996). A chemoautotrophi-
cally based cave ecosystem. Science, 272(5270), 1953—1955.
Schidlowski, M. (2001). Carbon isotopes as biogeochemical recorders
of life over 3.8 Ga of Earth history: Evolution of a concept.
Precambrian Research, 106(1—2), 117—134.
Stevens, T. O., & McKinley, J. P. (1995). Lithotrophic microbial eco¬
systems in deep basalt aquifers. Science, 270(5235), 450—455.
Woese, C. R. (1987). Bacterial evolution. Microbiological Reviews, 51(2),
221-271.
CLASTIC SEDIMENTS IN CAVES
Qregory S. Springer
Department of Geological Sciences, Ohio University
BASIC PHYSICAL PROPERTIES
Clastic sediments are broken fragments of preexisting
rocks that have been transported and redeposited.
Clastic sediments can be derived from any rock type,
including rocks enclosing caves and originating out¬
side of caves (Fig. 1), as well as secondary deposits
such as travertine. Clastic sediments are volumetrically
the most common deposits in caves. The majority of
clastic particles in cave sediments are detrital grains
eroded from land surfaces and carried into caves by
streams, mass movements, wind, wave action, and
ice (Table 1). Geomorphologists use clastic cave sedi¬
ments to understand the processes that shape caves
and landscapes, including system responses to climate
change, while hydrologists may focus on the driving
and resisting forces that affect sediment movement
through karst systems. In contrast, archaeologists and
paleontologists seek to know how objects came to rest
in a cave, why they were buried, and whether the
objects have been disturbed or altered since
Encyclopedia of Caves.
© 2012 Elsevier Inc. All rights reserved.
CLASTIC SEDIMENTS IN CAVES
emplacement. The diversity of sediments and their
uses means that research goals vary markedly from
one study to another, but all cave studies share a com¬
monality in that subterranean research techniques are
modified versions of techniques developed for surface
deposits and landforms. Analyses begin with sediment
description.
Scientists studying cave sediments focus on their
stratigraphy and are typically interested in mineralogy,
median grain sizes, sorting, and sedimentary structures.
The mineral content of sediment grains is largely a func¬
tion of their source. Grains eroded from cave surfaces or
FIGURE 1 Clastic sediments slumping into a short cave passage
connecting a sinkhole and stream passage. The cave stream occasion¬
ally pushes rounded stream cobbles uphill and out of the tube, but
cobbles are falling back into the cave along with angular blocks shed
by cliffs and hillsides in the sinkhole. Together, the sediments are a
gravity deposit.
TABLE 1 Types of Clastic Sediments Commonly Found in Caves
135
the overlying landscape will contain minerals found in
the surrounding rocks ( e.g ., calcite or chert). These
locally derived sediments are defined as autochthonous
and contrast with allochthonous grains derived from
neighboring nonkarst landscapes. Allochthonous sedi¬
ments commonly include insoluble silicate minerals,
such as quartz, feldspar, or mica, which are easily dis¬
tinguished from soluble carbonate and evaporite
minerals. Perhaps surprisingly, clay-sized sediments in
caves draining silicate rocks are typically dominated
by quartz and contain very little to virtually no carbon¬
ate grains. Allochthonous sediments can be used to
identify cave stream sources if local streams drain
rocks with distinctive mineral compositions.
The median grain size is a diameter for which 50%
of all grains are larger and 50% of all grains are smal¬
ler (Fig. 2). Sorting refers to the relative proportions of
different grain sizes. As examples, a well-sorted sand
deposit may consist almost entirely of 1-mm grains,
but a poorly sorted deposit may contain similar pro¬
portions of clay, silt, sand, and pebbles. Cumulative
grain size distributions are useful for understanding
sediment textures, which are functions of size and sort¬
ing, and how sediments were deposited. For instance,
the middle flat regions in the two quietwater curves in
Figure 2 mean the sample has significant amounts of
sand and significant amounts of clay, but very few
grains of intermediate sizes. This is the result of attri¬
tion, whereby a sandy limestone ceiling was slowly
dissolved and quartz sand grains fell onto clays being
deposited by slow-moving water.
Sediment grains are often organized into distinctive
patterns, called sedimentary structures, during deposi¬
tion. For instance, sand ripples contain thin inclined
Water-Borne (Stream) Facies Class: Detrital Sediments Transported by Flowing Water
Bedload: Comprised of grains mobilized by tractive forces at the base of stream flows
Slackwater: Comprised of grains carried in suspension and deposited floodwaters
Quietwater: Comprised of grains carried in suspension and deposited within standing bodies of water
En mass: Comprised of grains transported as semicoherent units by floodwaters (sliding beds) or debris flows
Gravity Facies Class: Sediments or Blocks Transported by Gravity, but Independent of Fluid Flow
Collapse: Comprised of blocks detached from cave ceilings, walls, and entrance areas
Infiltrates: Comprised of detrital grains transported by gravity through discrete, subvertical pathways
Residual Facies Class: Sediments Derived from Weathering of Primary or Secondary Deposits
Residuum: Bedrock weathering products preserved in situ with primary grains or sedimentary structures
observable
Surface Equivalent
Channel lag deposit
Overbank deposit
Lacustrine (lake)
deposit
Debris flow deposit
Talus and scree
Saprolite
Sources: Springer et al. (1997); Bosch and White (2007).
ENCYCLOPEDIA OF CAVES
136
c
Clay Silt Sand Gravel
Grain size (millimeters)
FIGURE 2 Seven example grain-size distributions plotted by par¬
ticle diameter (mm) and cumulative percent finer. Two curves are
derived from ancient stream bedload deposits and 50% of their
grains are sand size or smaller (orange lines). In contrast, slackwater
sediments deposited by floodwaters have median grain sizes in the
silt range (green). The floodwaters were moving more quickly than
those responsible for the quietwater sediments (blue), but the quiet-
water sediments have comparatively large median grain sizes. This
unexpected finding reflects sand grains falling from a sandy lime¬
stone ceiling onto clays being deposited in a permanently flooded
cave passage; median grain size is not always a reliable indicator of
past water velocities.
FIGURE 3 Sediments deposited since the 1940s in a West
Virginia cave. White lines separate three major stratigraphic
packages. The basal gravel bed was deposited prior to a large logjam
forming just downstream of this photograph. The logjam created a
large pool during floods, wherein slackwater sediments consisting of
silt and organic detritus were deposited to form the middle strati¬
graphic unit. Eventually, the cave floor aggraded to the top of the
logjam and this allowed gravel to move once again. The resulting
channel deposit (behind student) was deposited along with silt
banks.
layers (laminae) arranged as cross-bedding (Fig. 3)
formed as grains rolled off ripple crests. Cross-bedding
is evidence of deposition by moving water or air, but
many other sedimentary structures exist and can be
used to infer other sediment transport and deposition
processes. A bed lacking sedimentary structures is
generically referred to as massive and this too may
have meaning, perhaps recording sediments having
been transported chaotically within gravity flows (e.g.,
mudflows and other landslides).
STRATIGRAPHY
Stratigraphy describes how individual beds of
sediment are arranged relative to one another, includ¬
ing their stacking orders and ages. Each bed represents
a different chapter in cave passage evolution, so
deciphering how and why beds are associated with
one another is valuable for understanding cave and
landscape histories. For instance, the stratigraphy seen
in Figure 3 includes streambed gravels ( bedload ) over-
lain by finer sediments that accumulated in a large
pool behind a logjam. By slowing floodwater veloci¬
ties, the logjam prevented gravel from being moved
and created the conditions necessary for finer sedi¬
ments to settle out. The channel bottom aggraded (filled
with sediment) over several decades until it was nearly
level with the top of the logjam, allowing a new gravel
layer to be deposited a meter above the old one. The
stream is now incising through the accumulated sedi¬
ments because cave explorers removed the logjam.
Caves can possess multiple generations of deposits
scattered among passages whose ages may differ by
thousands or millions of years. Long geomorphic records
can be constructed by combining the stratigraphic
records of many individual deposits. This requires
placing deposits in chronological sequence, which is
typically accomplished through paleomagnetic, radio-
metric, cosmogenic isotope, and relative dating techni¬
ques. Cave sediments generally adhere to the three laws
of traditional stratigraphic studies, including the law of
superposition stating that the oldest sediments are at the
bottom of a deposit. The law of original horizontality
states that all sedimentary beds are deposited flat-lying,
but clay and fine silt commonly adhere to inclined cave
surfaces to create similarly inclined lamina (thin layers)
in a process called parallel accretion (Fig. 3). So, tilted
lamina in caves should not be automatically assumed
disturbed. The law of cross-cutting relationships is
violated in unique ways and objects cutting through
a deposit may be older than the sedimentary deposit.
This occurs when sediments are deposited around pro¬
jections (e.g., travertine) on cave surfaces. Care is needed
ENCYCLOPEDIA OF CAVES
CLASTIC SEDIMENTS IN CAVES
137
with regard to the three laws, but violations are not so
frequent as to void standard stratigraphic techniques.
A widely encountered stratigraphic sequence is the
abandonment suite, which is comprised of stream sedi¬
ments that become finer toward the top of a section.
Active stream tiers commonly carry ample bedload,
which accumulates to varying depths on passage
floors. A passage ceases to receive bedload when a
lower level passage forms. Flood water velocities in the
older, higher tier decrease because it mostly experi¬
ences backflooding or overflow from the lower
passage. In this lower velocity setting, suspended sedi¬
ments are deposited atop the older sediments, hence
from the bottom to top, a typical abandonment succes¬
sion is gravel, sand, silt, clayey silt. Abandonment
suites can be mistaken for channel infill created during
aggradation of a streambed and the silts and clayey
silts can be deposited long after a passage has been
abandoned, so fining-upward sequences should be
interpreted carefully.
Long-term deposition ( aggradation ) may be recorded
as banks of sediments on channel margins and thick
deposits beneath streambeds. This can occur when cave
passages are obstructed (Fig. 3), climate changes, or
excess sediment is supplied by tributary streams, gla¬
ciers, or landslides. Aggradation can be inferred where
the evidence is consistent with paragenesis or ascending
erosion. Paragenesis occurs when clastic sediments
shield cave floors from erosion and ceiling erosion cre¬
ates vertical accommodation space. Over time, stream-
bed and ceiling elevations increase (ascend) and a thick
aggradational deposit is created. The process is unique
to caves and leaves distinctive evidence, including ceil¬
ing pendants and meander notches that propagated
upstream during passage growth. Lateral erosion can
detach pendants from ceilings and leave them other¬
wise intact and upright in an aggradational sequence,
which is convincing evidence of long-term erosion and
sediment transport, rather than simple infilling as part
of passage abandonment.
The stratigraphic facies concept is widely used for recon¬
structing depositional relationships and geomorphic his¬
tories (Table 1). Facies are distinctly unique bodies of
sediment identifiable on the basis of appearance, compo¬
sition, texture, or sediment sizes. A depofacies possesses
unique sedimentary structures and stratigraphic succes¬
sions whose origins are derived from a common deposi¬
tional process or environment. Examining Figure 3, the
uppermost gravel layer is channel fill adjacent to a
silt-dominated point bar. The two stratigraphic units
have very different textures and appearances, but flu¬
vial (stream) processes deposited them more or less
simultaneously, so they represent a stream depofacies
traceable throughout much of the cave. In contrast,
organic-rich sediments below the gravel layer were
deposited in a large, intermittent pool and lack cross
bedding (Fig. 3). Hence, a lacustrine facies is recog¬
nized (Table 1), but the facies cannot be traced
throughout the cave because it is restricted to the
area upstream of the logjam. We can estimate the
extent backflooding created by the logjam by map¬
ping the extent of pool slackwater sediments.
SEDIMENT PRODUCTION
Clastic sediments are produced by the physical dis¬
aggregation of preexisting rocks during weathering
and mechanical erosion. Chemical weathering weakens
rocks by altering mineral compositions and by remov¬
ing the minerals cementing them together. Physical
weathering involves fragmentation due to tensional or
compressive stresses, including expansion of clay
minerals within rocks, frost shatter, and root wedging.
The latter two are frequently sources of angular rock
fragments, including breakdown, in and near cave
entrances (Fig. 1). Breakdown blocks are clastic sedi¬
ment and they easily exceed the total volume of all
other clastic deposits in a cave, occurring where pas¬
sage widths exceed the tensile strength of ceiling or
overhanging rock beds.
Cave streams mechanically erode conduit walls
through a combination of plucking and corrasion.
Plucking adds sediment to karst streams when intense
hydraulic forces pull or wedge blocks of rock from
conduit surfaces. Corrasion occurs when sediment
grains carried by floodwaters strike bedrock surfaces
with sufficient force to break off fragments. This pro¬
cess is generally referred to as abrasion when the result
is a smooth or polished cave surface and percussion
when it results in obvious flaking and rock breakage.
Percussion generally implies impact of large grains,
such as cobbles, whereas abrasion is most commonly
associated with "sand blasting" by suspended sedi¬
ment. As reported by Malcolm Newson (1971), solid
carbonate grains created by abrasion can collectively
weigh more than the dissolved solids in those same
cave-fed floodwaters. Whether corrasion plays a major
role in cave development elsewhere is a matter of
ongoing research, but it is undoubtedly a nontrivial
source of sediment to high-energy cave streams where
large grains impact cave walls.
Carbonates may contain chert or other low-solubil¬
ity minerals, which enter cave passages as the sur¬
rounding rock is eroded. Where abundant, chert can
be the major or only significant autochthonous bedload
supplied to cave streams. In other cases, comparatively
high concentration of autochthonous grains may accu¬
mulate in clays and silts when sedimentation rates
are otherwise very slow (Fig. 2), perhaps recording
ENCYCLOPEDIA OF CAVES
138
c
protracted pipe-full conditions in a quietwater setting
(many hundreds of years) (Table 1).
SEDIMENT TRANSPORT AND
DEPOSITION
Stream Sediments
Sediment moves through cave streams in suspen¬
sion, whereby turbulence keeps grains aloft in the
water column with little or no contact with the stream-
bed, and as bedload. The latter moves by rolling, top¬
pling, sliding, and saltation (similar to bouncing).
Grains moving in suspension and as bedload are col¬
lectively referred to as the stream's sediment load, a
quantity often expressed in kilograms per second.
Load increases during floods and little or no load is
carried between floods because stream waters are
moving too slowly to entrain (put in motion) grains.
Intuitively, high water velocities are needed to
entrain and transport large sediment grains, but this is
also true of clay and silt-sized particles because they
are cohesive. The relationship between size and the
velocities needed to entrain particles is shown in
Hjulstrom's Diagram (Fig. 4A). Particles will continue
to be transported at velocities less than what was
needed to entrain them and coarse sediments are
deposited first because of their greater weights. Clay
and silt are deposited last because they are easily
suspended. The velocity-dependent deposition of sedi¬
ment grains causes them to be sorted into beds con¬
taining similar-sized particles. This is often expressed
as a fining upward sequence consisting of a basal
gravel bed beneath a sand bed capped with a silt bed.
Hjulstrom's Diagram is useful for presenting ideal¬
ized relationships between grain sizes and sediment
transport. But the mechanics of sediment transport are
complex and a number of factors must be considered,
including those describing conditions at the scale of
individual grains (Fig. 5). These individual particles
shield one another from the full brunt of floodwaters,
normal forces pinning particles in place, and friction
and gravity. Those resisting forces compete against
buoyancy, drag, lift, and shear stress. All submerged
objects are buoyant and the effective weight of a
submerged quartz grain is only 62% of its dry weight,
a significant difference, although not enough for it
to float. Lift and drag are created by vertical and
horizontal pressure differences. Fluid pressure is
inversely proportional to velocity, so higher pressures
are reached on the upstream and undersides of grains
(Fig. 5C). Much like an airplane wing, the pressure dif¬
ference pulls grains toward lower pressure regions. In
contrast, shear stress is generated by differences in
Grain Diameter (mm)
(b) Shield's Diagram
OJ
£
£ 0.1
CD
s 0.01
FIGURE 4 Hjulstrom's Diagram is based on laboratory experi¬
ments and is divided up into three zones. Sediments will be
entrained when water velocities reach the uppermost dividing line
and, as indicated by arrows, fine sand is the easiest to move because
finer sediments are cohesive. Sediments will continue to move even
after velocities fall into the transport zone because their inertia has
been overcome, but grains will eventually be deposited when veloci¬
ties cross the lower dividing line. However, sediments finer than
sand are easily suspended and may remain in motion for long time
periods (left side of diagram). The Shield's Diagram is more com¬
plex, but is preferred by sedimentologists. The Reynolds shear stress
and Shield's parameter are calculated based on a variety of variables
discussed in the text and take into account fluid and grain
properties.
-►jHydraulicsh*-
r*-► /
1 10 100
Reynolds Shear Stress
(a) Flow direction
FIGURE 5 (A) Shows imbricated sediment grains on a channel
bed. The forces acting on a grain are shown in (B) with purple indi¬
cating a resisting force and green representing driving forces. The
latter act to entrain particles. Lift and drag are the result of pressure
differences with higher pressures forming at the upstream and
undersides of particles (C). Shear stress is derived from the momen¬
tum of moving water and transferred through the water's viscosity.
ENCYCLOPEDIA OF CAVES
CLASTIC SEDIMENTS IN CAVES
139
water velocity above grain surfaces. The water on
grain surfaces has an effective velocity of zero, but the
water's viscosity helps transmit the momentum of
floodwaters to grain surfaces and accelerates them in a
downstream direction. Greater differences in velocity
translate to larger shear stresses.
The Shield's Diagram incorporates many more fac¬
tors than Hjulstrom's and is widely used in bedload
studies. The Shield's parameter accounts for the com¬
peting influences of particle weight and shear stress.
Motion is expressed as a function of the Reynolds
shear stress and Shield's parameter (Fig. 4B). The for¬
mer is calculated using particle diameter, kinematic
viscosity, and shear velocity. The latter incorporates
additional variables, including shear stress exerted on
the streambed, and altogether the calculations help
express the effects of turbulence around a grain. The
importance of turbulence at grain and macroscopic
scales should not be underestimated because it creates
large lift and drag forces by driving fast-moving water
close to grain surfaces, thus creating large velocity
gradients. So, together the Shield's parameter and
Reynolds shear stress take into account the driving
and resisting forces discussed previously.
Gravity Sediments
A gravity deposit results from sediment moving as
a semicoherent or highly viscous mass of solid parti¬
cles, as opposed to a large volume of water moving a
much smaller volume of sediment. Gravity deposits
are easily distinguished from stream sediments (allu¬
vium) by a lack of sorting. Breakdown is, technically, a
gravity deposit, but many other gravity deposits are
found in caves, including landslides of many sizes.
Clastic sediments and highly degraded rock can col¬
lapse to form fast-moving landslides or be slowly
churned as they creep downslope. Debris flows, simi¬
lar to rapidly moving liquid concrete, are known to
enter caves in mountainous regions, but hillslope-
derived gravity deposits (colluvium) are most common
in caves. Colluvium may enter a cave via impassable
openings and spread out to form alluvial fans made of
many different sizes of material (Fig. 1).
GEOMORPHIC PERSPECTIVE
As major contributors to landscape denudation,
streams carry solutes and sediment produced during
weathering and erosion. Sediment quantities and
calibers (sizes) are sensitive to changes in climate, tec¬
tonics, and base-level elevations, which make old clas¬
tic sediments useful for understanding past changes in
earth systems, especially when considered together
with passage morphologies and elevations. This is well
documented by studies of the Mammoth Cave System
in Kentucky, U.S.A., and its local base level, the Green
River. Drs. Darryl Granger, Art Palmer, and Vic
Schmidt age-dated clastic sediments in Mammoth
Cave to reconstruct landscape responses to the
Pleistocene Ice Ages (Schmidt, 1982; Granger et al.,
2001). Early in the cave's history, the Green River was
apparently incising slowly because cave streams had
time to create large, low gradient trunk passages.
Smaller but steeper canyon passages formed soon after
glaciers rearranged river paths farther to the north and
caused the Green River to incise rapidly, increasing
local hydraulic gradients. Subsequently, cycles of sedi¬
ment infilling and incision occurred in Mammoth
Cave as the Green River alternately aggraded and
incised in response to Ice Age climate fluctuations.
We would know considerably less about the effect of
Ice Ages on the region's largest rivers without the
unique ability of caves, such as Mammoth, to protect
clastic sediments for millions of years.
Clastic cave sediments can also be used to recon¬
struct flood histories of surface rivers where caves are
found in riverbanks. Paleoflood studies are most prac¬
tical when a surface river is the only significant source
of water and sediment to a cave because this simplifies
the identification and interpretation of slackwater sedi¬
ments (Springer and Kite, 1997). Once identified, a
minimum floodwater depth can be inferred from
slackwater sediment elevations and this stage can in
turn be used to estimate flood discharge. Such studies
of clastic sediments offer the potential to significantly
extend flood records and improve flood recurrence
interval calculations. Clastic cave sediments have
many other potential applications and the bibliography
includes references to a broad range of studies.
See Also the Following Articles
Breakdown
Cosmogenic Isotope Dating of Cave Sediments
Bibliography
Bosch, R. F., & White., W. B. (2007). Lithofacies and transport of clas¬
tic sediments in karstic aquifers. In I. D. Sasowsky & J. Mylroie
(Eds.), Studies of cave sediments: Physical and chemical records of
paleoclimate (pp. 1—22). Dortrecht: Kluwer. doi:10.1007/978-1-
4020-5766-3_l.
Granger, D. E., Fabel, D., & Palmer, A. N. (2001). Pliocene-Pleistocene
incision of the Green River, Kentucky, determined from radioac¬
tive decay of cosmogenic 26 Al and 10 Be in Mammoth Cave sedi¬
ments. Geological Society of America Bulletin, 113(7), 825—836.
doi:10.1130/0016-7606(2001)113<0825:PPIOTG>2.0.CO;2.
ENCYCLOPEDIA OF CAVES
140
c
Newson, M. D. (1971). A Model of Subterranean Limestone Erosion
in the British Isles Based on Hydrology. Transactions of the
Institute of British Geographers, 54, 55—70.
Schmidt, V. A. (1982). Magnetostratigraphy of Sediments in
Mammoth Cave, Kentucky. Science, 217(4562), 827—829.
doi:10.1126/science.217.4562.827.
Springer, G. S., & Kite, J. S. (1997). River-derived slackwater sedi¬
ments in caves along Cheat River, West Virginia. Geomorphology,
18(2), 91—100. doi:10.1016/S0169-555X(96)00022-0.
Springer, G. S., Kite, J. S., & Schmidt, V. A. (1997). Cave sedimentation,
genesis, and erosional history in the Cheat River Canyon, West
Virginia. Geological Society of America Bulletin, 109(5), 524—532.
doi:10.1130/0016-7606(1997)109<0524:CSGAEH>2.3.CO;2.
CLOSED DEPRESSIONS
IN KARST AREAS
Ugo Sauro
Department of Geography, University of Padova
TYPES OF CLOSED DEPRESSIONS
Closed depressions are the dominant and most distinc¬
tive forms of the karst landscape. The following types
of closed depressions can be observed in karst areas
(listed according to increasing size):
1. Rain pits or rain craters (1 cm to about 3 cm in size)
2. Solutions pans or kamenitza (centimeter- to meter-
scale features)
3. Dolines ; also called sinkholes (diameters ranging
from some meters to over 1 km)
4. Compound depressions (hundreds of meters to a few
kilometers)
5. Poljes (reaching up to some tens of kilometers)
Dolines, compound depressions, and poljes link the
surface and the underground fissure and cave systems.
SOLUTION DOLINES AS THE
DIAGNOSTIC FORMS
OF KARST SURFACES
The solution doline (also called a sinkhole) is
considered to be the diagnostic form of the karst
landscape and is also defined as the index form
(White, 1988; Ford and Williams, 2007). Dolines are the
most common closed depressions in the Carso di
Trieste, which is known as the classical karst because it
is regarded as the most typical karst region. Dolines
commonly have circular to subcircular plan geometry,
and a bowl- or funnel-shaped concave profile. Their
depths range from a few decimeters to a few hundred
meters, and their inner slopes vary from subhorizontal
to nearly vertical.
The Slav term doline means "depression" and in a
broad sense includes channels and hollows of dif¬
ferent types, such as fluvial valleys, dry valleys, blind
valleys, uvalas, poljes, and karst dolines. The name dolina
was first applied in 1848 by Morlot to a circular, closed,
karst depression. Subsequently, Cvijic, in 1895, extended
use of the term to all circular, closed depressions in karst
areas. Gams (2000) further suggested that the word doline
could be replaced by the term kraska, to highlight its
importance as a morphological marker of the karst
landscapes.
Scientific investigation of dolines began in the
middle of the nineteenth century, during construction
of the Southern Railway in the Austro-Hungarian
Empire through the Karst of Trieste. At that time, the
most common theory was that dolines resulted from
the collapse of cave roofs. If that were true, the natural
hazard for a railway crossing the karst was high.
The railway engineers, on the other hand, came to a
different conclusion: that most dolines originated
through dissolution.
Dolines link the surface and the underground drain¬
age systems. The absence of still water, such as lakes
or ponds, at the bottoms of most dolines reveals that
water is lost to the underground karst system, most
commonly to caves. Through the sole observation of
surface forms, a doline seems very simple: it is a bowl-
or funnel-like depression with one or a few swallow¬
ing cavities at the bottom which are covered by soil.
Cross sections outcropping in road cuts and quarries,
however, reveal that dolines are actually rather com¬
plex. To understand how dolines and, in general, all
large, closed karst depressions develop, it is necessary
to examine (1) their morphology and size; (2) their
locations and relationships with the topographic and
geomorphologic settings; (3) their structures; (4) their
hydrological behavior and related solution processes;
(5) other processes that play a role in their evolution;
(6) examples of evolution; and (7) peculiar morpholo¬
gies that occur under specific environmental condi¬
tions. In the following sections, all of these aspects will
be discussed.
Morphology and Size of Dolines
Most dolines show a circular or slightly elliptical
plan geometry (commonly, the main axis/minor axis
ratio is <1.5 and the depth/width ratio is about 0.1).
The three-dimensional form of dolines may be com¬
pared to a bowl (nearly hemispherical depression),
a funnel (nearly conical depression), or a well (nearly
cylindrical depression). In some doline populations.
Encyclopedia of Caves.
© 2012 Elsevier Inc. All rights reserved.
CLOSED DEPRESSIONS IN KARST AREAS
141
more complex forms, such as truncated conical depres¬
sions with flat bottoms and star shapes, are common.
The bowl shape is the most typical doline shape
and is four to ten times more common than funnel-
shaped dolines (Cvijic, 1895). The well-shaped doline
is uncommon. Soil-covered forms show an almost
circular plan geometry. Depressions that developed
on bare rock are more irregular and commonly
show sharp contours that follow the main fracture
systems.
The dolines described by Cvijic as "holotypes" were
subsequently carefully analyzed by Sustersic (1994),
who recognized that "... the dolines ... are not as reg¬
ular as many simple morphometric methods pre¬
sume. .. some of these irregularities are probably due
to the roughness of the neighboring terrain... the
others are probably due to the greater dynamism of
the whole doline which does not permit the 'bowl' to
achieve a regular shape." Sustersic (1994) also identi¬
fied the presence of three distinct concentric areas
within each single doline: (1) a flat central area, which
is generally covered by soil; (2) a ring of steep slopes;
and (3) an outer belt with a gentler slope.
A population of dolines observed in a large-scale
topographic map rarely shows the entire spectrum of
possible sizes attained by these karst depressions,
because smaller forms are not represented. The mor¬
phometric parameters of the dolines measured in the
Carso di Trieste indicate that there is not a continuous,
asymptotic size distribution; therefore, it is possible to
recognize different subpopulations. The smaller forms
are relatively shallow, and it is difficult to identify
them through field observation alone. Within a doline
population, there are a few forms that can be distin¬
guished from the typical doline of the population
because of some particular characteristics, such as
elongation, asymmetry, steepness of the slopes, depth,
and so on. Morphometrical methods and analyses of
various doline populations are illustrated in a number
of papers (Bondesan et ah, 1992).
Locations and Relationships with the
Topographical and Geomorphological Settings
Dolines commonly occur in populations with
variable numbers of individuals and with different
densities, ranging from a few to over 200 per square
kilometer. Within a population, the density depends on
the surface slope. On subhorizontal or gentle slopes, the
density of dolines is higher than on steep slopes.
Dolines are commonly lacking on very steep slopes.
Two major types of karst plateaus can be distinguished
on the basis of the spatial distribution of dolines: one
that is characterized by isolated dolines (the classical
karst type) and another that is characterized by honey¬
comb systems of dolines. In many karst plateaus, inter¬
mediate situations can be observed. With respect to
the distribution of dolines, it is possible to recognize the
influence of major fracture systems. The center of the
doline is commonly set at the intersection between two
or more fractures. Rows of dolines commonly follow
fault lines, especially when the dislocations are linked
to extensional tectonics that bring different rock units
into contact. In karst plateaus where a net of dry valleys
can be distinguished, dolines are aligned along the
valley bottoms.
The Structure of Dolines
In the bare karst, such as at elevations above the
timberline, it has been observed that the structure of
closed depressions is controlled by the macro- and
micro-structural properties of the soluble rocks. For
example, in Monte Canin (Julian Alps, between Italy
and Slovenia) the properties of the rock yield a high
density of steep-sided, rocky, nival depressions, or
kotlici. In the typical bowl- or funnel-shaped dolines,
the rock is normally covered by soil and other
surface deposits. The structure of the dolines in the
soil-covered karst is visible only through artificial
cuts. Some basic characteristics of dolines can be
inferred from the analysis of many cross-sections. In
particular, it can be clearly observed that the doline
structure begins to develop before the onset of the sur¬
face depression. Many small depressions in the hard
rock, which are completely filled by soil and surface
deposits, go undetected if one looks only at the sur¬
face. These small forms have been termed embryonic
cryptodolines or subsoil dolines (Nicod, 1975). Some of
these structures host paleosols, beside the present-day
soils. The presence of paleosols in the filling sediments
indicates that the development of cryptoforms took
place over a long time and that cryptodolines have
acted as sediment traps since the beginning of their
formation. In the early phases, the trapping action was
effective for soils, which show less erosion than in the
surrounding area in correspondence to the cryptode¬
pression. When the cryptodolines evolved as surface
forms, the trapping action also became efficient for
other materials, such as rock debris from the slopes,
soil sediments, aeolian dust, volcanic ashes, and so on.
In the Mediterranean karst, these fillings are com¬
monly red (terra rossa, or "red soil"), and the color is
the result of weathering processes. In the dolines of
the classical and dinaric karst, the filling commonly
consists of a slightly reddish brown silt, which is pri¬
marily composed of weathered, loess-like sediment of
aeolian origin. The filling may be up to 10 m thick. The
profile of some dolines exhibits multilayered filling.
ENCYCLOPEDIA OF CAVES
142
SUBSOIL KARREN
C
FIGURE 1 Schematic showing the structures of a doline and a cryptodoline. The epikarst in the host rock, the fillings, and the soil are
outlined. The secondary porosity of the soluble rock is greater at the bottom of the dolines. The main morphodynamic and hydrologic
processes are also provided.
The sketch of the structure of an ideal middle
latitude doline (Fig. 1) shows that the solid rock/filling
materials interface is not a simple boundary, but rather
a complex, involute surface that consists of many
rounded bulges in the soluble rock separated by
crevices and fissures that become narrower inward.
These covered solution forms, called rounded karren,
develop through the slow flow of the water solution at
the rock/filling interface. The structures of dolines in
different karst areas show large variability depending
on the role that is played by different factors, such as
lithology, structural conditions of the rock, topogra¬
phy, climate conditions, and features inherited from
previous morphodynamical events.
Hydrological Behavior and Related Solution
Processes in Dolines
A simple doline is a hydrological form that can be
compared to a first-order valley segment. Its concavity
is an expression of the convergence of water toward
the bottom which is a transitional point between
surface and underground hydrologic networks. The
physical structure of the doline is the framework of its
hydrological functioning and, because this functioning
is also related to the genesis and evolution of the form,
the doline may be considered the surface expression of
a peculiar three dimensional hydrostructure. The geo-
morphological setting and the structure of the dolines
allow the distinction of three major hydrostructural
types (Fig. 2): (1) point-recharge doline, (2) drawdown
doline, and (3) inception doline.
A point-recharge doline begins to form when a fluvial
net loses water to the cavities of a soluble rock. It is nec¬
essary for protocaves to begin to form inside the rocky
mass, connecting surface recharge points to an under¬
ground network. Once developed, a protocave focuses
both drainage and solute removal at the points of transi¬
tion from the surface to the underground network, thus
causing the development of a surface depression. The
point-recharge doline is commonly located along a
ENCYCLOPEDIA OF CAVES
CLOSED DEPRESSIONS IN KARST AREAS
143
SUBSOIL KARREN
O
>
"0
>
3
oo
>
73
73
m
73
SUBSOIL KARREN
CAPILLARY BARRIER
cr
ill
£
03
Z)
o
111
CL
FIGURE 2 The three main types of solution dolines. (A) In the point-recharge doline, the strongly asymmetric hollow is linked by one
main sink, fed primarily by surface and soil water. (B) In the drawdown doline, the depression results from focusing of the dissolution inside
the water infiltration zone of the rock through centripetal convergence of the mainly subsurface water held inside the epikarst (i.e., the upper
zone of the soluble rock presents a greater secondary porosity). (C) In the inception doline, the depression also originates from a centripetal
convergence of water, but this occurs inside a preexisting hydrogeological structure and is triggered by a change of hydraulic conductivity of
the rocky mass, influenced by lithological and structural factors.
ENCYCLOPEDIA OF CAVES
144
c
FIGURE 3 A line of funnel shaped dolines in the Monti Lessini
along the bottom of a dry valley (Venetian Pre-Alps, Italy). These
inception dolines develop just above a lithology change.
rather dry hydrographic net and shows transitional fea¬
tures between the depression of a blind valley and a
common drawdown doline (Fig. 2A). It is often elon¬
gated and skewed in the direction of the flow, with the
upstream side being gentler than the downstream side.
The drawdown doline is the most common and
typical karst form. To understand the drawdown
dolines it is necessary to consider the behavior of the
water in the rock zone just beneath a nearly horizontal
or gently sloping surface. Near the surface, the rock
fractures tend to open by tensional relaxation; therefore,
the soil water seeps through the fractures and enlarges
them through dissolution. With time, secondary
porosity develops inside the shallow rock layer near the
surface which then becomes saturated in wet periods.
If, in some points of this layer, the water is able to find
and enlarge routes to an underground network, flow
paths converging to the leakage zones are thus
established. As a consequence, the water velocity
increases centripetally toward each leakage point, and
more water molecules are brought into contact with the
rock surface near the major drain. There, the rock is
dissolved more rapidly, secondary porosity increases
faster than in the peripheral zone, and the hydraulic
conductivity increases. Consequently, a depression
tends to form at each leakage zone, the center of which
deepens faster than in the surrounding areas. A draw¬
down doline is, therefore, the product of the positive
feedback resulting from the interrelations between
the hydrological and the solute processes acting inside
the outer layer of a fractured soluble rock.
Gams (2000) calls the dissolution process at the
central part of a doline accelerated corrosion. The hydro¬
structure corresponding to this central part is compara¬
ble to the funnel-shaped depression created by a
pumping well in the water table of a phreatic aquifer.
The outer rock layer, which is characterized by high
secondary porosity, is shallower in the peripheral areas
and thicker in the central parts of the dolines. It hosts a
hanging aquifer, which is nearly saturated during wet
climate periods and nearly empty during dry periods.
The existence of the epikarst with the characteristics of
a water reservoir was recognized through the study of
the hydrological regime of karst springs.
The inception dolines develop from the interception
by the epikarst of a hydrogeological structure that
developed inside the rock and previously triggered by
a change in hydraulic conductivity of the rocky mass.
A change of lithology or the presence of an imperme¬
able layer, such as clay or chert, may cause the
formation of a hanging aquifer. If connections such as
fractured zones exist and allow water to overcome an
obstacle, a hydrostructure similar to the drawdown
dolines develops, yielding a focused dissolution both
above and below the bottleneck (Fig. 3). Above, the
fractures of the rock are enlarged to fissures; below,
small shafts develop.
This type of doline differs from the drawdown
doline because of the more marked lithological and
structural control and because of an early evolution of
the hydrostructure, which is not necessarily dependent
on the epikarst.
The models illustrated in Figure 2 do not consider
the role of surface deposits or filling materials. These
components are important, because the water is
absorbed by soil, surface materials, and filling sedi¬
ments before it reaches the fissures in soluble rock.
The fillings may host small aquifers nested above the
epikarst and act as pads that slowly release water
downward, thus influencing with their physical
character both the flow velocity at the cover—rock
interface and the water regime of the epikarst and the
main karst springs.
ENCYCLOPEDIA OF CAVES
CLOSED DEPRESSIONS IN KARST AREAS
145
Other Processes in Doline Evolution
Even if the major process of doline formation is
dissolution through a differential mass wasting of the
rock, the following processes also play roles in the
shaping of the depression: soil forming processes and
other weathering processes, slope processes, overland
flow processes, the capacity of the closed depression to
trap different types of sediments, and the processes of
evacuation of filling materials.
The soil acts as a filter and as an insulating layer
with respect to the outer environment. Soil releases
solutions rich in humic acids and air with a high
content of carbon dioxide. It also yields fracture-filling
material.
Frost shattering of the rock is a very important
weathering process. In the fillings of most midlatitude
dolines, a variable amount of angular rock debris is pro¬
duced through this mechanism. During the cold phases
of the Pleistocene, some dolines were completely filled
with rock debris. Slope processes influenced by the
gravitational force, such as creep and solifluction, are
responsible for soil thinning, for the formation of a
debris cover along the slopes, and for its thickening in
the bottom areas. The overland flow processes are
responsible for erosion, the washing of soils and loose
material, and for the deposition of colluvial sediments
in the bottom area of the depression. Once formed, a
closed depression acts as a trap for different types of
materials that are carried by wind, rain, and so on.
In the midlatitude karst, the doline filling contains
variable amounts of loess-like deposits that were trans¬
ported by the wind during cold Pleistocene phases, vol¬
canic ashes, sands, and silt deposited by rain, and so
on. The fine-grained material is commonly deposited on
the entire karst surface and then accumulates by over¬
land flow into the central part of the depression. Many
authors consider the filling material to be a residue of
limestone dissolution. Even if in some karst areas this
may be true, most of the midlatitude doline fillings con¬
sist of allochthonous materials, with the exception of
angular rock debris due to in situ frost shattering. The
filling material may be evacuated through the karst con¬
duits through different processes, such as mass wasting
of the rock debris by solution, subsurface flow, liquefac¬
tion of the loose material, or piping.
Examples of Evolution
Some examples will help to understand the evolu¬
tion of dolines. Point-recharge dolines are present in
the gypsum karst of the Santa Ninfa Plateau (Sicily,
Italy). There, chains of dolines follow the pattern of a
fluvial network that developed on the impermeable
cover and was overprinted on the gypsum. Inside
some chains, the bottom of the upstream doline is at a
lower altitude than that of the progressively down¬
stream doline. The last upstream doline marks the end
of a blind valley. It is clear that each closed depression
became less active after the development upstream
of a new swallow hole. So, the speed of bottom
deepening is strictly linked to the activity of the
swallow hole.
In the Waitomo district of New Zealand, popula¬
tions of point-recharge dolines occur in the framework
of interstratal karstification. It is possible to reconstruct
the transitions from a fluvial network in an impervious
rock to a honeycomb system of dolines that developed
in the underlying limestone units and the subsequent
recovery to fluvial network in the impervious rock unit
previously overlain by the limestones, following their
chemical erosion. Gunn (1986) proposed a model based
on five sample areas, each representing a different
stage of the transition. The model outlines the impor¬
tant roles played by both allogeneic recharge and the
location of previous subterranean drainage structures.
A small, nearly horizontal area of the classical karst,
which was surveyed at very high resolution (scale
1:1000; contour interval 1 m; spot elevation, at the
bottoms of some depressions, checked at the resolution
of 0.1 m), is characterized by drawdown dolines,
including some very small and shallow forms
(Fig. 4A). The main morphometric parameters reveal
three main subpopulations represented by the follow¬
ing types of dolines: (1) small and shallow dolines less
than 20 m in diameter and 0.4 to 2 m in depth; (2)
medium-sized dolines 12 to 50 m in diameter and 2 to
7 m in depth; and (3) large dolines, 50 to 120 m in
diameter and 8 to 15 m in depth. From the ratios
between the numbers of individuals of each group it is
possible to infer that most of the embryonic surface
dolines will probably abort prior to becoming
medium-sized forms. Their hydrological functioning is
at the limit between triggering or not the positive feed¬
back that would allow them to evolve into typical
dolines. The coexistence of the three subpopulations
also suggests different ages for the forms, which are
probably related to the starting times of their evolution
during favorable morphoclimatic phases. The smaller
and shallower dolines could be the result of the karst
morphogenesis that began at the end of the last cold
period, which was triggered by the preexistence of
cryptodolines.
Seven rock terraces were cut in the Montello
(Venetian Pre-Alps of Italy) neotectonic anticlinal mor-
phostructure consisting of Upper Miocene conglomer¬
ate during the tectonic uplift. The terraces
subsequently became the sites of doline morphogenesis
(Fig. 4B). Depending on the ages of the terraces, it is
possible to analyze populations of drawdown dolines
that developed in very similar geomorphological
ENCYCLOPEDIA OF CAVES
146
c
(A)
500 m
361 m
311 m
750 m
Classical karst near Villa Opicina
m
m
FIGURE 4 Digital elevation model of two doline areas. (A) Classic Karst near Opicina (Triest). (B) Seventh terrace in Montello Hill
(Venetian Pre-Alps). Morphometric parameters and distribution of dolines are more homogeneous in the Montello area than in the Opicina
example. In the latter, the dolines show a larger variability of morphometric values. Modified from Ferrarese and Sauro, 2001.
ENCYCLOPEDIA OF CAVES
CLOSED DEPRESSIONS IN KARST AREAS
147
environments and now represent different evolution¬
ary stages. A "standard" doline for each paleosurface
was reconstructed from morphometrical analysis of the
dolines of each terrace (with the exception of the lowest
one, where no recognizable dolines exist) (Fig. 5). The
series show that the standard forms are not the expres¬
sion of a regular and linear growth, probably because
of the different influence of some local factors on each
terrace. Nevertheless, there is a general trend toward
increasing sizes and depths of the dolines, from the
lower and younger terraces to the upper and older
ones. What is clear is that the deepening process was
faster than the widening process. What was increasing
in a perfectly linear way was the sum of the volumes of
the dolines for an area unit. The depressions tended to
evolve from bowl-like to dish-like forms. The flat bot¬
toms of the older and larger depressions result from
thick fillings that formed inside the hollows. In the
lower terraces, the dolines are isolated spots on a
nearly flat surface, but on the upper terrace their
boundaries are shared and a honeycomb karst mor¬
phology is almost reached (Fig. 4B).
In the southern Monte Baldo (Venetian Pre-Alps of
Italy) are dolines that differ in size and character.
Many host very thick fillings; some open dolines filled
up to the rim look like amphitheaters with flat bot¬
toms, which are open on one side. The fillings com¬
monly consist of angular rock debris supported by a
loess-like silt matrix. Some relicts of large egg-carton¬
like structures show that a population of big dolines,
which probably developed in the late Tertiary, had
been largely dismantled by the periglacial processes
during the cold phases of the Pleistocene. During
the Interglacials, the karst morphogenesis prevailed
against other competitive processes, and so, many
dolines survived up to the present day, even though
their morphology has been strongly modified (Sauro,
1995).
The Velebit Mountains of Croatia have different
types of dolines. In the central plateau, the dolines are
very large and deep and are funnel shaped, almost
without covers and fillings and with open shafts on the
slopes and at the bottom. In the southern plateau, the
depressions are smaller and partially filled by rock
debris and loess-like deposits. The main difference
between the two groups is represented by the lithology
of the host rocks: the larger rocky dolines developed in
a massive limestone breccia characterized by low sensi¬
tivity to frost shattering, while the smaller and partially
filled dolines developed in a limestone more susceptible
to gelivation. The larger depressions remained, there¬
fore, filling free, and during the winter they trapped
large amounts of wind-transported snow. The large
amount of meltwaters accelerated the evolution of the
forms, which reach huge sizes.
In the Eyre Peninsula of South Australia, clusters
and lines of dolines developed on a stack of fossil
coastal fore-dunes of Middle and Late Pleistocene age,
now consisting of calcarenites. Some dolines occur high
in the local topography and are also aligned in groups.
Their distribution is influenced by the diversion
of groundwaters in fractures that cut into the pre-
Pleistocene basement (granites) and the predominant
action of solution processes in the limestone above such
zones. This is probably related to an earlier mass wast¬
ing of the sand during its diagenesis in the open fis¬
sures of the underlying rock with a local increase of
porosity of the newly formed rock above. The influence
of a buried surface on the development of hydrostruc¬
tures and related forms is referred to as underprinting
by Twidale and Bourne (2000).
The above examples demonstrate the important
roles played by various factors and processes such as
the morphostructural setting of the soluble rock, its
lithology and density of discontinuities, its susceptibil¬
ity to frost shattering, the qualities of the rock units
above and below, the characters of the erosional sur¬
face on which the karst forms began to develop, the
morphoclimatic system and its relative changes, and
the occurrence of such events as falling volcanic ashes,
among others.
Populations of Dolines Linked to Specific
Environmental Conditions
In humid tropical climates, some populations of closed
depressions show peculiar characteristics that distinguish
them from doline populations of the midlatitude karst. In
particular, the closed depressions are often larger than
typical dolines and in the topographical maps show a
star-shaped figure. The cartographic representation of a
tropical karst landscape with closed depressions resem¬
bles a classical karst landscape with dolines, but with an
inverted relief. In fact, most of the rounded, concentric
figures drawn by the map contour lines are conical or
tower-like hills encircling the concentric star-shaped con¬
tours of a closed depression (Fig. 6). The egg-carton
shape of the basin is characterized by a great difference
between minimum and maximum depths (differences in
elevation between the bottom of the depression and the
lowest and highest points of the perimeter). This type of
tropical doline, called a cockpit from their Jamaican name,
has very steep slopes, usually from 30 to 40°, and is
coated by a thin soil cover. Often, the rock outcropping
on the steeper slopes shows evidence of biokarstic weath¬
ering, such as honeycomb alveolar cavities, and locally of
carbonate deposition. At the bottom, the fillings are shal¬
low, and sometimes it is possible to observe a flat,
sponge-like rocky bottom, blackened by biokarstic weath¬
ering, and a sharp change of gradient at the contact with
ENCYCLOPEDIA OF CAVES
148
FIGURE 5 Models of the standard doline for every Montello terrace, ordered from the youngest (T2) to the oldest (T7). The model is drawn
using the average value of each morphometric parameter. The horizontal scale is the same for all drawings, while the vertical scale is slightly
decreasing from T2 to T7.
ENCYCLOPEDIA OF CAVES
CLOSED DEPRESSIONS IN KARST AREAS
149
FIGURE 6 Sketch of a tropical cockpit. The star-shaped depression is encircled by conical hills separated by dry valleys and gaps. The pla¬
nar shape of the watershed line is polygonal.
the surrounding slopes. During the wet season, the shal¬
low holes at the bottom of some depressions behave like
springs, and temporary ponds may form.
The morphological differences between the normal
solution dolines and the tropical cockpits are mainly
due to the presence of the valley-like depressions
entrenched on the slopes of the cockpits, which deter¬
mine the star-like form of these sinks (Williams, 1985).
These incisions in the slopes, even if similar to some
morphological types of the fluvial landscapes, are solu¬
tion forms. Their presence in the tropical dolines, and
not in the midlatitude dolines, is due to such different
factors as: (1) the larger dimensions of the slopes of the
tropical dolines; (2) the poorer development and role
of the epikarst; (3) the minor influence of limestone
debris and soil cover; and (4) the more active overland
flow and biokarstic weathering that act on the slopes.
Similar to tropical dolines, some populations of
midlatitude dolines, such as those of the coastal belt of
Dalmatia (Velebit, Croatia) and of Cantabria (Spain),
may be interpreted as inherited forms, which first
developed in subtropical humid conditions during the
late Tertiary. The survival of inherited forms was pos¬
sible where the dismantling processes competing with
the solution processes, such as frost shattering, have
played a secondary morphogenetic role due to both
peculiar environmental conditions and a low suscepti¬
bility of the rock to gelifraction. Also similar to the
tropical dolines, some very large dolines of the Central
Velebit plateau can be explained as old, inherited
forms that were not considerably altered by the peri-
glacial processes of the cold Pleistocene phases because
they developed in a limestone breccia resistant to frost
shattering.
ENCYCLOPEDIA OF CAVES
150
A
SANDSTONE
FIGURE 7 The four main types of dolines that do not directly originate by solution of rock at or near the surface: (A) collapse doline,
resulting from the collapse of the roof of a cave; (B) subsidence doline, formed by the settling of an insoluble rock following solution of
the underlying soluble rock; (C) intersection doline, originating from the emptying of fillings of an old fossil cave due to intersection with
the topographical surface; (D) cover doline, which has developed in incoherent rocks burying a soluble rock or partially filling a karst
depression.
Summary of Solution Dolines
The solution doline may be considered the most
typical karst geo-ecosystem that links the surface and
the underground networks. A doline is both a three-
dimensional form, and a multicomponent geosystem
consisting of several elements of variable thickness,
such as soil, surface deposits, fillings, and epikarst. If
the development of a doline is triggered by a positive
feedback in the synergical interactions of hydrological
and solute processes, other cooperative and competi¬
tive processes may also play an important role in its
evolution. In particular, the soil plays a cooperative
function by enriching the water with carbon dioxide.
The fillings, on the contrary, commonly act as a
competitive factor by slowing water seepage into
the epikarst and, if carbonates are present in their com¬
position, by increasing the water pH, thus decreasing
potential dissolution of the underlying solid rock.
Without the presence of soil cover it is not possible for
a doline to attain a typical bowl- or funnel-like shape,
as documented by the irregular depressions of the bare
high mountain karst. Soil and other surface deposits,
therefore, play a fundamental role in the evolution of
the forms. Dissolution is the dominant morphogenetic
process in the cockpits of the tropical karst. If the soil
cover is very thin on the steep slopes, its role can be
replaced by a rock layer that is strongly corroded by
biological processes.
OTHER TYPES OF DOLINES
In karst areas it is possible to find other types of
dolines that do not directly originate in the solution
of rock at or near the surface. The main types are
(1) collapse dolines ; (2) subsidence dolines-, (3) intersection
dolines ; and (4) cover dolines (Fig. 7).
Collapse Dolines
Collapse dolines result from the collapse of a cave.
They show different shapes, with vertical or overhanging
walls. Collapse dolines that allow entry into a subterra¬
nean system are also called karst windows, because they
open into the underground environments. The best
known karst windows are the dolines of Skocjanske
Jame in Slovenia (Fig. 8), where the Reka River enters
into a large cave feeding the aquifer of the classical karst
and the spring of Timavo 40 km away.
ENCYCLOPEDIA OF CAVES
CLOSED DEPRESSIONS IN KARST AREAS
151
FIGURE 8 The karst windows on the first part of the subterra¬
nean course of the Reka river in the Skocjanske Jame system
(Slovenia).
FIGURE 9 A huge collapse doline called Red Lake near Imotski
(Croatia). The depth from the rim to the lake is over 250 m, and the
depth of the lake itself is greater than 250 m.
The range of sizes of the collapse dolines varies
from a few meters to hundreds of meters. The Velika
dolina (large doline) of Skocjanske is about 500 m in
diameter and 164 m in depth. The Blue Lake and Red
Lake dolines of Imotski (Croatia) are among the most
spectacular collapse dolines in the world (Fig. 9). The
main axis of the Blue Lake doline is about 1.5 km; the
diameter of the Red Lake doline is about 400 m at
the lake surface, and its total depth, including the
submerged part, is 520 m.
The Chinese name of the largest of the collapse
dolines is tiankeng. A tiankeng is a collapse doline at
least 100 m long, wide and deep, and with perimeter
walls that are close to vertical (Zhu and Waltham, 2006).
The greatest of these depressions exceed 100 million m 3
in volume and have developed in correspondence of large
cave chambers where underground rivers are flowing.
These forms are not the result of a unique collapse
episode, but of a number of collapses often related to
oscillations of the water table (Beck, 1984). Most of the
collapse dolines do not provide access to a subterra¬
nean system because of collapse debris at the cavity
bottom.
A particular type of collapse doline is the cenotes,
open depressions in the coastal karst belts giving
access to the water table and resulting from the col¬
lapse of the roof of submerged cavities. The best-
known cenotes are in Yucatan and in the Caribbean
Islands. The evolution of cenotes and of the related
cave systems is often due to both efficient dissolution
by the brackish waters of the coastal belts and oscilla¬
tions of the water table. Some cenote populations
could be interpreted to be intersection dolines.
Subsidence Dolines
Subsidence dolines are closed depressions caused
by the subsiding of an area. They may be due to
upward propagation of an initial collapse of the roof of
a deep cavity with development of a breccia-pipe
ENCYCLOPEDIA OF CAVES
152
c
structure; when this structure reaches the surface a
closed depression with an unstable bottom is formed.
Subsidence closed basins are also found on insoluble
rocks overlying soluble rocks. If the underlying rocks
are eroded by solution processes, the rocks above
subside. These hollows, which may host lakes, are
frequent on rocks overlying very soluble minerals,
such as gypsum and salt.
Intersection Dolines
Intersection dolines are depressions formed when
an old cave system, partially or totally filled with sedi¬
ments, is cut by the topographic surface as a conse¬
quence of the lowering of such a surface by chemical
erosion. The opening of such fossil caves reactivates
the old hydrostructures, leading to evacuation of the
fillings and development of closed depressions. These
forms are relatively common in the classical karst,
where chains of depressions or slender depressions
also some hundred meters long exist. On the bottom of
such a doline, it is easy to find relicts of cave fillings
and stalagmites. The recent building of a highway has
cut many of these forms and they have been labeled
roofless caves by Mihevc (2001); here, the term intersec¬
tion dolines is preferred because they are no longer
underground forms.
Cover Dolines and Other Types
Cover dolines are the closed depressions formed in
incoherent materials such as alluvial deposits, glacial
drift, superficial deposits, soil sediments, and so on.
These forms, also commonly called alluvial dolines,
have different characteristics according to the type,
continuity, and thickness of the deposits and the pres¬
ence of a well-developed karst relief under the depos¬
its or of a still incipient karst hydrology, among other
factors. The alluvial plains of a large river and the flat
bottoms of large solution dolines, retaining filling
deposits, may be considered as end members of the
various development environments for this type of
doline. The influence of buried karst hydrostructures
on the formation of cover dolines is a clear example of
underprinting. Various processes may play a role in
the evolution of these forms, such as water infiltration,
washing and piping, suffosion (in the sense of erosion
from below), suction of the sediments caused by oscil¬
lations in the water table, liquefaction of the sediments
and the linked mud flows, and changes in volume
caused by freeze—thaw cycles.
The two main models for the development of cover
dolines are (1) upward migration of a cavity and (2)
subsidence, as determined by the gradual rearrange¬
ment (a sort of inner creeping) of the cover material
induced by its mass wasting through the karst net.
If the cover material is sufficiently cohesive, the mass
wasting by the water infiltrating underground pro¬
duces a cavity with an arched roof; the arch gradually
migrates upward, where it reaches the surface and
causes the sudden opening of an ephemeral cover
doline with nearly vertical walls. The subsidence cover
dolines form in less cohesive sediments.
In addition to the forms of dolines described above
are other types originating by different processes, such
as dolines along fault lines caused by seismotectonic
movements and anthropogenic dolines. Also, craters
similar to dolines were created by shelling during
World War I in the chalky limestone of the Venetian
Pre-Alps and are an interesting anthropogenic type. The
crater-like depressions behave as drawdown dolines
because of the rock fracturing caused by the explosions..
UVALAS AND COMPOUND AND
POLYGENETIC SINKS
In the old literature, a large closed depression that
does not show a doline morphology is also referred to
as an uvala. This name is one of the first classical ones
applied to a type of karst depression that has been
abandoned because it was erroneously utilized by
some authors in combination with the theory of the
karst cycle. A recent review about the uvala concept,
based on the analysis of a population of large karst
basins of the dinaric karst, revalues this term (Calic,
2009). Indeed, there are forms that match the original
concept such as closed basins mostly of elongated and
irregular plan shape ranging in size from approxi¬
mately one up to several kilometers without a flat bot¬
tom and never flooded, generated mainly by the karst
process, and so without clear evidences of fluvial and
glacial shaping. In the dinaric karst, development of
such depressions of uvala type is mostly guided by
tectonically broken zones of regional extension, in
which corrosion is intensified due to rock fracturing.
So, from the genetic point of view, uvalas could be
considered forms of accelerated corrosion triggered by
local tectonic settings and dynamics.
Beside the uvalas, both compound and polygenetic
closed depressions can be observed in the karst land¬
scapes. A compound hollow is a form that originated
from the fusion of simpler forms. If more dolines coa¬
lesce together, an irregular depression develops, some¬
times with a lobate perimeter.
A polygenetic sink is a closed depression that
clearly evolved through both the karst process and
another morphogenetic process. The most frequent
types are the (1) tecto-karstic hollows; (2) the fluvio-
karstic hollows ; and (3) the glacio-karstic hollows.
ENCYCLOPEDIA OF CAVES
CLOSED DEPRESSIONS IN KARST AREAS
153
A tecto-karstic hollow is a closed basin that devel¬
oped inside a tectonic depression that also evolved
through karst processes. A fluvio-karstic hollow may
be considered the closed ends of blind valleys that
evolved through both fluvial and karstic processes.
Glacio-karstic basins are common forms in the alpine
high mountain karst. These forms were formed by
both karst processes and glacial abrasion and are of
two main types: (1) the high plateau type, often elon¬
gated and similar to a bathtub; and (2) the glacial
cirque type, which developed in the bottom of a glacial
cirque. Larger glacio-karstic depressions can be over
1 km long. Their bottoms are often occupied by rocky
hummocks, drumlins and till deposits..
DEPRESSIONS IN KARST AREAS
Poljes are the largest closed depressions observed in
the karst terrains. In 1895, Cvijic defined a polje as a
"large karst depression, with a wide, flat and nearly
horizontal floor, completely enclosed between steep
slopes." This simple definition is insufficient to explain
the genesis of this type of form and the role of the
karst processes in its characterization. To understand
this form it is necessary, as for dolines, to consider
both the structure of the form and its dynamics;
in other words, to consider the entire geosystem
expressed by the form. Poljes present a large variabil¬
ity, and the development of most of them cannot be
explained through the karst process only, because they
are polygenetic forms, resulting from the combination
of a number of processes.
Typical of this form are the poljes of the dinaric karst,
which show large, flat, and nearly horizontal floors,
from one to some tens of kilometers long. The sides,
usually sloping at about 30°, connect with the bottoms
at a sharp angle. Closed depressions of this type are
also present in other karst regions of the world and are
referred to by various names, such as campo or piano in
Italy, plans in France, and hojos in Cuba.
The main environmental peculiarity of this type of
form is the common absence of a permanent lake
inside the basin; an ephemeral lake without surface
outlets may form and disappear during the seasonal
cycle in relation to the precipitation regime. The lake,
if present, represents the transition point between the
surface and the underground hydrology and serves as
a window to the underground aquifer. In other words,
the polje floor is affected by the oscillations of the local
water table. Although the lake may be fed by both sur¬
face runoff and underground circulation, the draining
of it occurs exclusively through the subterranean karst
network (Fig. 10). The dynamics of the lower part of
the basin is strictly linked with the seasonal cycle of
water input, throughput, and output. The temporary
lake is responsible for the solution of both the floor
and the base of the slopes, leading to the planation of
the bottom and to its lateral enlargement by marginal
corrosion. The forming, standing, and dissipation of
the water body is also the cause of a redistribution and
leveling of the filling materials and of their volumetric
reduction by solution and erosion via the subterranean
circulation system (Fig. 11).
In the structure of the poljes it is possible to recog¬
nize a large variability in the thickness of the filling
deposits, ranging from a few decimeters to more than
100 m. Swallowing cavities and springs may open at
both the bases of the slopes and inside the floor area.
Of these, some are real caves, and others are cover
dolines developed in the fillings above cavities in the
underlying rock. From the hydrological point of view,
some behave permanently as springs or as swallowing
cavities, and others may invert their functioning from
springs to swallow holes and vice versa. This last type
is referred to as an estavelle in France.
FIGURE 10 A typical blind valley: the Biogradski ponor (Bosnia-
Herzegovina).
FIGURE 11 A typical karst polje in the dinaric range: the
Dabarsko Polje (Bosnia-Herzegovina).
ENCYCLOPEDIA OF CAVES
154
c
FIGURE 12 The structural polje of Piano del Cansiglio in the Venetian Pre-Alps (Italy). The polje is now inactive and hosts smaller depres¬
sions nested inside the main floor. This form has been described as a lithological contact polje.
The classification of poljes may be based on both the
geomorphological character and the hydrodynamics.
Gams (1994) used these two criteria to distinguish five
main types of poljes: (1) border poljes, (2) piedmont poljes,
(3) peripheral poljes, (4) overflow poljes, and (5) base-level
poljes. A border polje is located at the transition between
a nonkarstic and a karstic area. The recharge of the basin
is, therefore, mainly due to surface streams and the
escape of water through stream-sinks. A piedmont polje
is located downslope of a mountain area from where a
large amount of debris has been received, filling the
depression and hosting a local aquifer. Here, the
recharge is also in the main made by surface streams. A
peripheral polje is a depression fed by a large internal
area of impermeable rocks with a centrifugal stream net¬
work. The sinks are located around the periphery of the
inlier. An overflow polje is underlain by a belt of rela¬
tively impermeable rocks that act as a hydrological bar¬
rier to water that emerges at springs on one side of the
polje floor and escapes via stream-sinks on the other
side of the basin. A base-level polje is a polje with a floor
cut entirely across karst rocks and is affected by the ver¬
tical oscillations of the water table; consequently, it is
inundated during high-level periods.
Based on the books of Ford and Williams (1989,
2007), it is possible to reduce the categories to three
basic types: (1) border poljes, (2) structural poljes, and
(3) base-level poljes. From the geomorphological point
of view, most poljes correspond to tectonic depressions
such as graben, faultangle depressions, pull-apart
basins, and so on. These dislocations often lead to
contact between rocks with different permeabilities,
thus leading to conditions of lithological contact. Many
structural poljes have also been described as lithologi¬
cal contact poljes (Fig. 12), a situation corresponding,
at least partially, to that of the border poljes.
Some structural poljes have trapped hundreds of
meters of sediments hosting local aquifers. An example is
the Piano del Fucino in the Central Apennines (Italy), the
bottom of which, during most of the Holocene, was occu¬
pied by a lake subject to numerous vertical variations in
the level. In early Roman and modern times tunnels were
excavated to drain the lake and to reclaim the floor for
agriculture. Today, the intensive agriculture is supported
by overexploitation of the alluvial aquifer.
Some poljes have developed along typical fluvial
valleys. A specific example is represented by Popovo
Polje in Bosnia-Herzegovina, which is 40 km long and
only 1 to 5 km wide. The floor slopes gently through¬
out its length as demonstrated by the course of the
river Trebinjcica. Before the construction of a draining
tunnel, the bottom was occupied from October to May
by a seasonal lake.
The best examples of true karst poljes can be found in
tropical karst regions, such as in southern China. Here,
the floors of certain populations of large cockpits are in
the vertical oscillation zone of the water table. The base-
level floors of these cockpits expand by marginal corro¬
sion, leading to the gradual dismantling of the ridges and
to fusion of the floors. The resulting compound depres¬
sions represent true karst poljes, probably a unique type
of polje created by karst processes only.
ENCYCLOPEDIA OF CAVES
COASTAL CAVES
155
See Also the Following Article
Soil Piping and Sinkhole Failures
Bibliography
Beck, B. F. (1984). Sinkholes: Their geology, engineering and environmen¬
tal impact. Rotterdam: A. A. Balkema.
Bondesan, A., Meneghel, M., & Sauro, U. (1992). Morphometric
analysis of dolines. International Journal of Speleology, 21, 1—55.
Calic, J. (2009). Uvala—Contribution to the study of karst depressions,
with selected examples from Dinarides and Carpatho-Balkanides.
PhD thesis, Postojna.
Cvijic, J. (1895). Karst. Beograd: Geografska Monografija.
Ferrarese, F., & Sauro., U. (2001). Le doline: Aspetti evolutivi di forme
carsiche emblematiche [The doline: Evolution aspects of the
emblematic karst form], Le Grotte d'ltalia, 5(2), 25—38 (in Italian).
Ford, D., & Williams, P. W. (1987). Karst geomorphology and hydrology.
London: Unwin Hyman.
Ford, D., & Williams, P. W. (2007). Karst hydrogeology and geomorphol¬
ogy. Chichester, U.K.: Wiley.
Gams, I. (1994). Types of poljes in Slovenia, their inundation and
land use. Acta Carsologica, 23, 285—300.
Gams, I. (2000). Doline morphogenetical processes from global and
local viewpoints. Acta Carsologica, 29, 123—138.
Gunn, J. (1986). Solute processes and karst landforms. In S. T. Trudgill
(Ed.), Solute processes (pp. 363—437). New York: John Wiley & Sons.
Mihevc, A. (2001). Speleogeneza divaskega Kraza [The speleogenesis of the
Divaca Karst], Ljubljana, Slovenia: Zalozba ZRC, ZRC SAZU.
Nicod, J. (1975). Corrosion de tipe crypto-karstique dans le karst
mediterraneen. Bulletin Association Geographique Franqois, 428,
284—297 (in French).
Sauro, U. (1995). Highlights on doline evolution. In I. Barany-Kevei
(Ed.), Environmental effects on karst terrains (Vol. 34, pp. 107—121).
Szeged, Hungary: Universitatis of Szegediensis.
Sustersic, F. (1994). Classic dolines of classical sites. Acta Carsologica,
23, 123-156.
Twidale, C. R., & Bourne, J. A. (2000). Dolines of the Pleistocene dune
calcarenite terrain of western Eyre Peninsula, South Australia: A
reflection of underprinting? Geomorphology, 33, 89—105.
White, W. B. (1988). Geomorphology and hydrology of carbonate terrains.
New York: Oxford University Press.
Williams, P. W. (1985). Subcutaneous hydrology and the develop¬
ment of doline and cockpit karst. Zeitschrift fiir Geomorphologie, 29,
463-482.
Zhu, X., & Waltham, T. (2006). Tiankeng: Definition and description.
Speleogenesis and Evolution of Karst Aquifers, 4(1), 1—8.
COASTAL CAVES
John E. M ylroie
Mississippi State University
INTRODUCTION
Caves that form in coastal environments are controlled
by different factors from caves that form in traditional
inland settings. The first and most obvious factor is the
physical and chemical power of waves and salt water
acting on coastal rocks. Second, and less obvious, is
the fact that sea level can change, and with that
change, the position of the coastline moves. Therefore
the position of cave development by coastal processes
will also move. Sea level can change in a variety of
ways, but there are two ways that are of particular
importance to cave formation on coasts. First, sea level
can change on a global scale, called eustatic sea level
change. The most common reason for this change is
the amount of ice on the continents. From 10,000 to
2,600,000 years ago—the Pleistocene Epoch—the Earth
underwent a series of ice advances called glaciations
(the "Ice Ages") and a series of ice retreats called inter¬
glacials. As ice sheets grew, evaporated seawater falling
as snow on land was trapped as ice, and sea level
dropped worldwide. When the ice melted as an inter¬
glacial occurred, sea level rose as the meltwater flowed
back into the ocean basins. Eustatic sea level change of
this type is called glacioeustatic sea level change.
Evidence indicates that the Earth went through at least
15 of these glacial cycles, and hence sea level changes,
in the Pleistocene. Second, sea level position can also
shift as isolated events at specific locations because the
land is either subsiding or being uplifted. Such sea-
level change is called local sea level change, as only
that local area is affected, and is commonly caused by
tectonic movements of the Earth, or compaction and
subsidence of sediments, as in river deltas.
Caves found in coastal areas fall into two major
categories, karst caves and pseudokarst caves. The lat¬
ter can form in almost any rock material by a variety of
mechanisms. The most common pseudokarst cave in a
coastal environment is the sea cave, produced by wave
action. Sea caves are usually single chambers or a small
collection of chambers and fissures open to the sea. The
precise scientific name is littoral cave, meaning a cave
formed within the range of tides. Coastal weathering
and erosive conditions commonly create coastal cliffs,
which in turn allow production of other pseudokarst
cave types, such as tafoni (subaerial weathering pock¬
ets), talus caves (boulder and block rubble deposits),
and fissure caves (cliff failure fractures). While tafoni
are only isolated pockets, talus and fissure caves can
sometimes be extensive. Flowever, talus and fissure
caves are a direct result of coastal erosion, and as a
result are in an environment of continued modification
and destruction, which results in loss of these cave
types through time. Sea caves, being in solid bedrock,
persist longer.
SEA CAVES
The most common of the coastal caves are sea caves,
and they are found the world over. Sea caves are caves
that form by wave erosion in coastal areas that contain
Encyclopedia of Caves.
© 2012 Elsevier Inc. All rights reserved.
156
c
exposed bedrock. They can develop in almost any type
of bedrock, with wave energy utilizing fractures and
other preexisting weaknesses in the rock to quarry out
voids by mechanical action. The chemical action of salt
water can also exploit rock weakness. The compression
of air caused by water flowing forcefully into cracks
and fractures in the rock can break rock, including
rock above sea level. In addition to exploiting weak¬
nesses in the rock, waves also interact with themselves
and the sea floor to create constructive interference
patterns that can focus wave energy at specific sites
along a uniform coast. As a result, seas caves can form
in rock where no obvious weakness exists.
Sea caves can vary from arches and small voids
only a few meters across to very large chambers up to
100 m deep and wide (Fig. 1). The sea caves seen on
coastlines today have formed rapidly, as sea level has
only been at its present elevation for perhaps 3000 to
5000 years, following melting of the large continental
glaciers at the end of the last glaciation. In areas such
as Alaska and Norway, where the Earth's crust was
depressed by large masses of ice during glaciation, the
shoreline is now rising as the crust rebounds to a
stable position following melting of the ice. In so
doing, sea caves formed many thousands of years ago
have now been carried high above modern sea level.
The Earth is currently in an interglacial, or between
glaciations, so glacial ice is at a minimum and sea level
is high. The last interglacial occurred 131,000 to
119,000 years ago, and during that time the ice melted
back a bit more than present conditions, and sea level
was about 6 m higher than it is today. On some rocky
coasts, sea caves produced at that time are still visible,
6 m above the ocean, if more recent erosion has not
obliterated them.
Sea caves have had a long history of interaction
with people, especially wherever sailors have used the
ocean on rocky coastlines. Sea caves were particularly
favored by smugglers to hide stolen goods, and to also
hide the small, fast sailing ships that carried such
cargo out of reach of the taxman. Pirates allegedly bur¬
ied treasure in sea caves, but most sea caves are in an
active erosional environment, and anything buried
would not survive long. Pirates who chose old sea
caves above modern sea level would have had better
success. Some sea caves formed during a past sea level
higher than today contain significant archaeological
and paleontological remains.
Sea caves are ubiquitous on the rocky coasts of the
world. Fingal's Cave in Scotland, the Blue Grotto of
Capri in the Mediterranean, Sea Lion Cave on the coast
of Oregon, and Arcadia Cave on the coast of Maine
are well-known sea caves visited by tourists on a regu¬
lar basis. Many organisms use sea caves as a refuge,
particularly seals, sea lions, and other marine mam¬
mals, as well as birds, which roost in the ceiling ledges
above the reach of waves. From the viewpoint of
cave exploration, sea caves are not of major category,
primarily because they are short in length. In areas
where other types of caves are rare, such as in
Southern California, sea caves offer the best cave
exploration option. Occasionally, sea caves can have
spacious chambers and over 500 m of passages.
Exploration of sea caves can be very dangerous for
those not experienced in handling strong waves, tides,
and currents.
FLANK MARGIN CAVES
The coastal environment creates a very unusual
type of cave when limestones are present. The interac¬
tion of freshwater and seawater produces a unique
geochemical situation that allows caves to form by dis¬
solution that are very different from both sea caves,
made by mechanical wave action, and most limestone
caves located in the interior of continents, which are
usually underground stream conduits formed by fresh¬
water dissolution.
FIGURE 1 Left, looking out of a large sea cave on Eleuthera Island, Bahamas. Right, Sea arches and waves, Barbados.
ENCYCLOPEDIA OF CAVES
COASTAL CAVES
157
Freshwater is slightly less dense than seawater,
because of the extra salt dissolved in seawater. Average
freshwater has a density of 1.0 g/cm 3 and average sea¬
water has a density of 1.025 g/cm 3 . The difference in
density is only one part in 40, but it is sufficient that
when freshwater flows toward the ocean inside an aqui¬
fer, it floats on top of the seawater that has invaded
the aquifer from the ocean. The boundary between the
freshwater and salt water is called the halocline ("halo"
meaning salt, "dine" meaning boundary) if it is a sharp
boundary. If the boundary is broad, containing water of
brackish salinity, it is called a mixing zone. The freshwa¬
ter flows toward the ocean because rainfall infiltrates
the land behind the coast, piling up in the aquifer
until there is sufficient slope to drive the water toward
the ocean. Where the freshwater is piled up inland,
because of buoyancy, the halocline sinks downward
into the seawater, much like a piece of wood floating
in water. Because the difference in density is one part
in 40 for each centimeter (or each meter) that the fresh¬
water piles up above sea level, it sinks 40 cm (or 40 m)
into the seawater. As the freshwater flows down this
small slope toward the ocean, its elevation above sea
level decreases, and in buoyant response, the halocline
rises up toward sea level, a 40-cm rise for each 1 cm of
elevation loss of the water table. At the coast, the fresh¬
water discharges to the sea as a thin sheet. This configu¬
ration of freshwater over seawater is called the
freshwater lens, because seen in cross section in an island
(where the water discharges to coasts on either side),
the freshwater body is seen to have the shape of a lens,
similar to a lens in a magnifying glass. Figure 2 shows
this relationship in a diagrammatic fashion, with verti¬
cal exaggeration. Understanding the freshwater lens is
critical to successful exploitation of fresh groundwater
in islands and coastal areas.
When the freshwater lens is formed within a lime¬
stone aquifer in a coastal region, a very unique type of
FIGURE 2 Cartoon diagram of a freshwater lens in a limestone
island, showing flank margin cave development in the lens margin,
under the flank of the enclosing landmass. The lens is drawn with
vertical exaggeration.
cave, called a flank margin cave, can develop. Limestones
are made up of CaC0 3 , either as the mineral calcite, or
its close polymorph (i.e., alternate crystal structure), ara¬
gonite. Seawater is usually saturated with CaC0 3 , and
cannot dissolve limestone very well. Freshwater that
has had a long residence time in a limestone aquifer is
also commonly saturated with CaC0 3 and also cannot
dissolve more limestone. However, because the fresh¬
water and seawater became saturated with CaC0 3
under different initial conditions, when they mix they
are capable of doing more dissolution, a process called
mixing corrosion or mixing dissolution. Where the fresh¬
water lens meets the seawater at the halocline, it is pos¬
sible to dissolve out large voids, or caves that otherwise
would not be able to form.
The top of the freshwater lens is also a place where
waters can mix. In this case, the freshwater at and
below the water table (in this case, the top of the lens)
is called phreatic water, and the water descending from
the ground surface above is called vadose water. It is
common for both the phreatic and vadose waters to be
saturated with respect to CaC0 3 , but as with the case
of mixing seawater and freshwater, the phreatic and
vadose waters are saturated at different initial condi¬
tions, so that when they mix, the water can dissolve
more CaC0 3 . Therefore both the top and bottom of
the freshwater lens are favorable environments for the
dissolution of CaC0 3 .
The top of the lens (the water table) and the bottom
of the lens (the halocline) represent density interfaces.
Organic particulate material transported by the vadose
water flow from the land surface commonly floats on
the top of the water table. Some of this organic mate¬
rial may then become waterlogged, and work its way
to the bottom of the lens where it floats on denser sea¬
water at the halocline. The decay of the organic matter
at these interfaces creates CO 2 , which dissolves in the
water to make carbonic acid that promotes CaC0 3 dis¬
solution. In both situations, if the amount of organic
material becomes too much, its decay will use up the
local oxygen supply to create anoxic conditions. If the
anoxic conditions persist, anaerobic bacteria will create
H 2 S, which can later encounter water with oxygen in it
to create H 2 SO 4 or sulfuric acid, a very powerful acid
which can dissolve even more CaC0 3 .
The mixing, and the possible anoxic conditions, that
promote dissolution can occur at the top and bottom
of the lens throughout the lens area. However, these
environments are superimposed on each other at the
edge, or margin of the freshwater lens where the top
of the lens slopes down to sea level, and the bottom on
the lens rises up to sea level. Because the lens is thin at
this point, its flow velocity is high because the entire
lens discharge is being forced through a thinning
wedge. The combination of increased flow and
ENCYCLOPEDIA OF CAVES
158
c
FIGURE 3 Interior of flank margin caves, showing the complex dissolutional morphology. (A) Small pores, vugs, and pockets (ruler in
lower right is 10 cm long). Cumulus Cave, Crooked Island, Bahamas. (B) Small pockets bordering a chamber in Cueva del Agua Sardinera,
Isla de Mona, Puerto Rico. (C) Chamber with wall pockets and pores. Ten Bay Cave, Eleuthera, Bahamas. (D) Large chamber with pockets,
arches, and tubes, 1702 Cave, Crooked Island, Bahamas.
superposition of the favorable geochemical environ¬
ments for CaCC >3 dissolution results in large voids
forming very rapidly at the margin of the lens. As
noted earlier, this discharge occurs at the flank of the
land, so the caves thus developed are called flank mar¬
gin caves. In addition, in tropical islands and coasts, the
limestone is commonly very young. Unlike the ancient
limestones of the mid-continent regions, these young
limestones have not been buried, squeezed, or greatly
altered. The rocks still have a high degree of primary
porosity, and water moves through them easily. Such
rock is said to be eogenetic. True conduit flow is diffi¬
cult to develop, as the rock has a high degree of per¬
meability, meaning that the water has many flow
routes to choose from. So how do caves form in this
setting?
Flank margin caves are not true conduits, like caves
formed by sinking streams in continental interiors.
They are instead mixing chambers. Conduit caves
form by turbulent water flow, but flank margin
caves develop in the laminar (nonturbulent) flow of a
porous, eogenetic limestone aquifer. The freshwater
enters the developing flank margin cave as diffuse
flow and exits, after mixing with seawater, as diffuse
flow. Flank margin caves form without human-sized
entrances. As a result of their nonconduit origin, flank
margin caves do not have long tunnels or a dendritic
pattern found in most conduit caves. The flank margin
caves are a series of oval rooms from pore size up to
large chambers, which tend to be extensive in the hori¬
zontal direction, but limited in the vertical direction,
a result of developing in the thinning margin of the
freshwater lens (Fig. 3). The chambers can connect in
a somewhat random manner, creating caves that are
unpredictable in their pattern. Maze-like areas are
common, indicating regions where chamber develop¬
ment did not go to completion. As the caves were
growing, the mixing zone advanced into them, such
that the back wall of the cave (the wall farthest from
the ocean) is the youngest. Complex cross-connections
between chambers can develop, and the caves can be
quite complex despite their simple mode of develop¬
ment. The typical flank margin cave consists of one or
more large chambers located just inside the edge of the
island. The cave may trend parallel to the coast for
some distance, but rarely penetrates very far inland, as
its development site was restricted to the margin of
the lens (Fig. 4). The caves form entrances when the
erosion of the hillside that contains them breaches into
the underlying cave. Initially, this entrance may be a
small opening, but through time it can enlarge as more
of the outer wall of the cave erodes away. Throughout
ENCYCLOPEDIA OF CAVES
COASTAL CAVES
159
FIGURE 4 Map of Sistema del Faro on Isla de Mona, Puerto Rico,
a very large flank margin cave. Note how the cave wraps around the
edge of the island, following the margin of a past freshwater lens.
Cartography by Marc Ohms.
the Bahamas and other carbonate island areas, there
are flank margin caves in all states of erosional
destruction, from almost intact to almost entirely
removed. On islands such as Bermuda, where weath¬
ering and erosion of the limestone hills is more rapid
than in the Bahamas, most flank margin caves have
been entirely removed.
Because they develop in the freshwater lens, flank
margin caves are sensitive to sea level change. If sea
level falls, the caves are drained as the freshwater lens
follows sea level downward, and cave development
ceases. If sea level goes up, the lens is pushed higher,
the caves become flooded with pure seawater from
below, and dissolution and cave enlargement stop.
The Bahama Islands are tectonically stable, meaning
they are not rising or falling because of tectonic forces.
The dry flank margin caves that explorers enter there
today resulted from the freshwater lens being 6 m high¬
er than today 131,000 to 119,000 years ago, during the
last interglacial. That sea level highstand lasted only
about 12,000 years, but caves with individual chambers
up to 14,000 m 3 in volume developed in that time frame,
windicating how rapidly this mixing dissolution process
can occur (Fig. 3). On stable limestone coasts around the
world are many flank margin caves that developed dur¬
ing the last interglacial sea level highstand.
Isla de Mona, halfway between Puerto Rico and the
Dominican Republic, is a small island that has been
uplifted by tectonic forces. It has huge flank margin
caves that formed almost 2 million years ago (Fig. 4).
These caves are very large because they developed in a
freshwater lens before the Pleistocene glaciations began
and so had a longer time to dissolve before sea-level
fluctuation caused the freshwater lens to change posi¬
tion. When the initial glaciation began early in the
Pleistocene, these caves were drained and speleothems
such as stalactites and stalagmites developed. When a
subsequent interglacial sea level highstand occurred, sea
level rose and the caves were reinvaded by the freshwa¬
ter lens, partially dissolving the speleothems produced
during the dry phase. The caves were then uplifted by
tectonics well beyond any further glacioeustatic sea level
changes, and have been preserved for exploration today.
In this case, local sea level change affected only this one
island. Some of the cave chambers are over 400,000 m 3 in
volume, much wider than they are high, but with many
complex connections with adjacent chambers. The cham¬
bers have ancient speleothems much modified by attack
from an invading freshwater lens, and more modern
speleothems that have grown since the last uplift event,
and which are in pristine condition. The longest flank
margin cave in the world, Sistema del Faro, is located on
Isla de Mona, and has over 19 km of survey on its map
(Fig. 4). Hank margin caves commonly have numerous
entrances, and in tropical settings are warm and friendly.
The pleasant conditions and many entrances make them
easy to explore. When movies and TV dramas show
pirates in caves, it seems they always have enough light,
and they move easily through large open passages.
Flank margin caves actually do look a bit like this
show-biz characterization (Fig. 5).
BLUE HOLES
Blue holes are names for large, deep pits that form
on islands and lagoons in tropical waters. They are
named blue holes because their great depth gives them
a very dark blue color (Fig. 6). They commonly connect
to cave systems at depth. The name blue hole was first
published in 1725, and later appeared on British
Admiralty charts from the Bahama Islands in the 1840s.
Blue holes became popularized in the early 1970s when
cave divers began to make the first serious investiga¬
tions of their depths. Since that time, blue holes have
been the subject of a number of major scientific
ENCYCLOPEDIA OF CAVES
160
c
investigations, including the discovery in 1979 of a new
class of crustaceans.
Blue holes are defined as "subsurface voids that
are developed in carbonate banks and islands; are
open to the Earth's surface; contain tidally-influenced
waters of fresh, marine, or mixed chemistry; extend
below sea level for a majority of their depth; and
may provide access to submerged cave passages"
(Mylroie et ah, 1995, p. 225). Blue holes can be addi¬
tionally characterized as being found in two settings:
(1) ocean holes, which open directly into the present
marine environment and contain marine water, usually
with tidal flow; and (2) inland blue holes, which are iso¬
lated by present topography from marine conditions,
open directly onto the land surface or into an isolated
pond or lake, and contain tidally influenced water of a
variety of chemistries from fresh to marine (Fig. 6).
While blue holes are best known from the Bahama
Islands, they are found in a wide variety of tropical
FIGURE 5 Hamilton's Cave, Long Island, Bahamas. Spacious,
well-lit chambers like this one are common in flank margin caves.
coasts and islands. Their origin is tied to the coast,
island, and lagoon environments where they are
found, which means that their development has been
influenced by glacioeustatic sea level changes of the
Pleistocene. Blue holes commonly contain many stalac¬
tites and stalagmites (known as speleothems) that are
now under water, indicating that the blue holes were
drained by glacioeustatic sea level lowstands, allowing
the speleothems to form from drip water, then were
flooded by return of sea level as the ice sheets melted
on the continents at the end of the last glaciation.
Some of these speleothems are more than 350,000 years
old, indicating that the blue holes containing them are
very old, and have undergone repetitive sea level low-
stands and highstands.
There are three main hypotheses of how blue holes
form: 1) drowning of surface karst features such as
pits and sinkholes; 2) collapse of deep-seated phreatic
dissolution voids; and 3) bank margin fracturing. Blue
holes come in a variety of morphologies, and may rep¬
resent features of polygenetic (i.e., many origins) devel¬
opment, in which case a combination of the above
hypotheses may be correct.
Exploration of blue holes generally involves cave
diving to great depths. Such exploration is at the lead¬
ing edge of technology and stamina, requiring the use
of mixed gases, long decompression stops, total dark¬
ness, the danger of silt-out (stirring up silt so that the
way out cannot be seen), and tight passages. Cave div¬
ing in blue holes is extremely dangerous, and many
cave divers and scientists have lost their lives trying to
penetrate into the unknown. Unlike many other types
of exploration, however, there is no substitute for
direct human exploration. The blue holes and their
associated caves, and the contents of those caves, can¬
not be viewed, measured, or sampled without some¬
one going there.
FIGURE 6 Blue holes, so named for their deep blue color. Left, an inland blue hole, Andros Island, Bahamas. Right, an ocean hole. Dean's
Blue Hole, the world's deepest at 200+ m. Long Island, Bahamas.
ENCYCLOPEDIA OF CAVES
CONTAMINATION OF CAVE WATERS BY HEAVY METALS
161
CONCLUSION
Coastal caves are important to science as they
contain information about present and past sea level
conditions. Their utilization as a habitat makes them
important for many organisms over the breadth of the
Animal Kingdom. The speleothems contained within
blue holes and flank margin caves contain within their
layers evidence about changes in the Earth's climate
over hundreds of thousands of years. Flank margin
caves reveal past conditions of the freshwater lens, an
essential component of island habitation. While flank
margin caves are generally easy to explore, sea caves
can be dangerous for the unwary, and blue holes
exceptionally dangerous even for the well trained.
See Also the Following Articles
Anchihaline Caves and Fauna
Underwater Caves of the Yucatan Peninsula
Bibliography
Bunnell, D. (1988). Sea caves of Santa Cruz Island. Santa Barbara, CA:
McNally & Loftin.
Mylroie, J. E., Carew, J. L., & Moore, A. I. (1995). Blue holes:
Definition and genesis. Carbonates and Evaporites, 20(2), 225—233.
Mylroie, J. E., & Mylroie, J. R. (2007). Development of the Carbonate
Island karst model. Journal of Cave and Karst Studies, 69(1), 59— 75.
Palmer, R. (1997). Deep into blue holes. Nassau, Bahamas: Media
Publishing.
Vacher, H. L. (1988). Dupuit-Ghyben-Herzberg analysis of strip-
island lenses. Geological Society of America Bulletin, 100(4),
580-591.
Waterstrat, W. J., Mylroie, J. E., Owen, A. M., & Mylroie, J. R. (2010).
Coastal caves in Bahamian eolian calcarenites: Differentiating
between sea caves and flank margin caves using quantitative
morphology. Journal of Cave and Karst Studies, 72(2), 61—74.
CONTAMINATION OF CAVE WATERS
BY HEAVY METALS
Dorothy J. Vesper
West Virginia University
Heavy metals are ubiquitous throughout nature, includ¬
ing within caves and karst environments. Evaluating the
accumulation and transport of metals in cave waters
requires understanding the governing physical and
chemical processes. While the presence of heavy metals
in speleothems and cave deposits has been investigated
in some detail, the general metal cycling through the
karst system is less well known but can be inferred from
analogous investigations in surface systems. The term
heavy metals is poorly defined and has been used incon¬
sistently through time and in the scientific literature. In
the context of this discussion, the metals and metalloids
discussed are those defined as potentially toxic by the
U.S. Environmental Protection Agency and the World
Health Organization (Table 1). It should be noted that
many of these metals, while toxic in large quantities, are
essential nutrients in small quantities.
NATURAL AND ANTHROPOGENIC
SOURCES
Metals are omnipresent in atmospheric, marine, and
terrestrial settings. In karst environments, they are most
likely to be found in three primary compartments:
soils, the matrix—fracture—conduit system, and springs
(Fig. 1). Within conduits and caves, heavy-metal rich
minerals may be found incorporated into speleothems,
coatings, fillings, rinds, and other cave deposits.
Additionally, both caves and springs may have metals
present in water or associated with suspended and bed
sediments.
Metals may be part of the natural background
or anthropogenic (Table 2). Spectacular deposits of
metal-rich speleothems can occur when caves exist in
proximity to natural geologic sources. For example,
Cupp—Coutunn Cave in Turkmenistan has spe¬
leothems rich in manganese, iron, lead, and zinc due
to the presence of overlying bituminous coal and sub¬
sequent hydrothermal alternation. Mbobo Mkula Cave
in South Africa also boasts unusual metal-rich spe¬
leothems thanks to the presence of overlying ore
minerals and a sulfide-rich black shale. The host rock
itself may contribute to the metal load. Low levels of
trace metals in spring waters in Nevada and California
have been attributed to the paleomarine chemistry at
the time of carbonate deposition. Given that the
groundwater feeding the springs was thousands of
years old, the water chemistry was attributed to disso¬
lution of the carbonates. Metals may also be found as
detrital material within the host rock. The manganese
source for Jewel Cave in South Dakota has been attrib¬
uted to such detritus.
Anthropogenic sources of metals are widespread
(Table 2). Sources may be diffuse, such as emissions
from fuel combustion, or more localized, such as
point-source discharges from manufacturing facilities.
Acid mine drainage (AMD) is a common metal source
in many karst regions (the Appalachians, Kentucky,
and Tennessee in the United States and in southern
China). Studies of metals in AMD in Tennessee have
shown that Fe and Mn concentrations decrease in
water when introduced to karst systems. The most
likely explanation is that metals precipitate in the pres¬
ence of alkalinity. In cave stream/spring water, the
drop in metal concentrations is associated with the
Encyclopedia of Caves.
© 2012 Elsevier Inc. All rights reserved.
162
c
production of floes of Fe and Mn hydroxides. There is
some suggestion that the metals may precipitate as a
metal armor on the host rock, thereby limiting the
interaction between water and rock. Manufacturing
sources also contribute to metal contamination. Horse
Cave in Kentucky and its associated aquifer and
springs have been impacted by discharge from a
metal-plating factory that began operations in 1970.
Although some dilution occurred, elevated concentra¬
tions of Cr, Ni, and Cu were found in both cave and
spring waters within the Hidden River basin.
ENVIRONMENTAL METAL CHEMISTRY
General Chemistry
Once metals are introduced into karst settings, their
storage and transport depend on physical processes,
the specific metal chemistry, and the chemistry of the
TABLE 1 Potentially Toxic Metals and Their Abbreviations
Metal
Abbreviation
Metal
Abbreviation
Arsenic
As
Nickel
Ni
Beryllium
Be
Lead
Pb
Cadmium
Cd
Antimony
Sb
Chromium
Cr
Scandium
Sc
Cobalt
Co
Selenium
Se
Copper
Cu
Titanium
Ti
Iron
Fe
Thallium
T1
Mercury
Hg
Vanadium
V
Molybdenum
Mo
Zinc
Zn
Manganese
Mn
surrounding environment. Metal speciation influences
solubility and the likely mode of transport through the
aquifer. Metal speciation also controls the bioavailabil¬
ity and toxicity of the metal. Metal chemistry can be
complex and depends on many competing variables.
Although a brief description is provided herein, the
reader is referred to the bibliography, which lists a few
of the excellent texts on the topic.
In general, metals are present in three forms: min¬
eral, otherwise associated with solids, or dissolved
(Fig. 2). Mineral-bound metals exist as source materials
in soils in bedrock, as secondary cave deposits, and as
detrital material throughout karst systems. Heavy
metals may also be associated with the surfaces of
solids and are often associated with suspended and
bed sediments. Metal interactions at solid surfaces
range in intensity from exchangeable (loosely bound)
to incorporation into the near-surface mineralogy
(tightly bound). Metals may also be incorporated into
surface coatings (both organic and inorganic) or
attached to insoluble organic matter. It is difficult to
distinguish between metals that are electrostatically
bound to the surface, specifically adsorbed, or copre¬
cipitated with the surface mineral or coating; therefore,
they are often loosely referred to in combination as
sorbed metals. Metals are present in water as free metal
ions, soluble metal complexes, or associated with col¬
loids. Metals may also form complexes with soluble
or colloidal organic compounds and thereby increase
their solubility dramatically. Copper, in particular, has
a strong affinity for organic compounds. Metals may
move between these compartments via chemical reac¬
tions such as dissolution, sorption, or desorption or via
physical processes such as deposition and entrainment
(Fig. 2).
Two key variables that control metal solubility and
chemical form are pH and reduction—oxidation (redox)
Soil/
Residuum
Epikarst
Matrix/
Fractures
Conduit/
Cave
System
Soil and Bedrock Aquifer
FIGURE 1 Locations for heavy metal storage in karst settings.
ENCYCLOPEDIA OF CAVES
CONTAMINATION OF CAVE WATERS BY HEAVY METALS
163
TABLE 2 Sources of Heavy Metals
Type of
Source Examples of Specific Sources
Natural Original marine deposition; detrital materials in bedrock; ore bodies; hydrothermal deposits; black shales; coals
Anthropogenic Mining and mineral processing; agriculture (fertilizer, pesticides, preservatives, irrigation); emissions and solid wastes
from fossil fuel combustion; sewage and solid wastes; manufacturing (metallurgical, electronic, ceramic, chemical,
pharmaceutical); sport and military shooting; breakdown of metal alloys and paints
FIGURE 2 Storage compartments for heavy metals in karst settings. Arrows indicate some of the chemical reactions and physical processes
that transfer metals between compartments.
state. Many metals are more soluble in acidic waters
than in neutral waters and thus tend to precipitate in
well-buffered karst waters. Redox reactions occur when
electrons are transferred between metal species. For
example, ferric iron (Fe 3+ ) is reduced to ferrous iron
(Fe 2+ ) by the addition of one negatively charged elec¬
tron. In actuality, free electrons do not exist, and the
reaction occurs by combination with another "half reac¬
tion" in which an electron is lost. Redox reactions are
extremely important because they control the solubility
of many metals. While some metals are mobile in
reducing conditions (As, Mn, Fe, Mo), others are mobile
in oxidizing conditions and immobile in reducing
environments (Zn, Cu, Hg, due to sulfide mineraliza¬
tion in reduced settings). For many metals, both the pH
and the redox conditions must be known to predict
what species should be present.
Redox states change spatially and through time.
Spatial changes may occur on a microscale, and tempo¬
ral changes may occur on a scale as short as hourly
(e.g., with storms and changing hydraulic conditions).
The degree to which metal redox reactions in caves are
mediated and catalyzed by microorganisms is a topic
of increasing interest. Microbes can be either electron
acceptors or electron donors and can therefore either
oxidize or reduce metals as part of their metabolic pro¬
cess. Hence, both precipitation and dissolution of
metals can be induced by microbial action. Much of
the microbial action may occur along redox gradients,
such as those found at the edges of caves.
Other factors may influence the redox state, such as
introduction of oxygen during a storm and degradation
of organic compounds. The breakdown of either natural
organic matter or organic contaminants in the overlying
soil, karst aquifer, or spring sediments can create locally
reducing and acidic environments capable of solubiliz¬
ing and mobilizing metals.
Iron and Manganese
Iron and manganese can be found in many locations
in caves. While oxide and hydroxide forms are the
most common, heavy-metal carbonates, phosphates,
nitrates, and sulfates have also been identified in caves.
In general, while reduced iron (Fe 2+ ) is soluble in
water, when it oxidizes to Fe 3+ it forms oxide and
hydroxide precipitates such as goethite (FeOOH) and
ferrihydrite (Fe(OH) 3 ). Limonite is a generic term that
refers to both mineral and amorphous forms of iron
oxides and hydroxides. Likewise, reduced manganese
(Mn 2+ ) is soluble, but its oxidized forms are not. Fully
oxidized Mn often forms the common cave mineral bir-
nessite, which can be poorly crystallized (f-Mn0 2 ). In
reality, however, the chemistry is far more complicated
ENCYCLOPEDIA OF CAVES
164
c
than this. The reactions also depend on metal concen¬
tration, solution pH, CO 2 partial pressure, and presence
of organic compounds. Both reduced and oxidized
metals can be mobilized by organic complexation, and
reduced metals can precipitate into metal—carbonate
minerals.
Trace and Contaminant Metals
Iron and manganese oxides and solids also play a
key role in trace-metal chemistry because of their abil¬
ity to scavenge trace metals from solution and
because—given their relative concentrations—they
often control the overall redox chemistry of the solu¬
tion or sediment. Iron oxides are commonly associated
with As, Cu, Ni, Mn, and Zn; manganese oxides com¬
monly contain Co, Fe, Ni, Pb, and Zn. Therefore, the
transfer of Fe and Mn between mobile and immobile
forms ultimately controls the trace-element chemistry,
even if the trace elements are themselves not redox
sensitive. For example, while precipitation of Fe and
Mn oxides may remove trace and contaminant metals
from solution, dissolution of the same oxides may
rerelease trace metals back into solution.
Studies of soils and marine and lake sediments have
shown that Fe and Mn oxides are typically the control¬
ling factor in determining trace-metal concentrations.
Trace metals may also be associated with organic com¬
pounds (both soluble and insoluble) and inorganic
coatings on particulates.
METAL STORAGE AND TRANSPORT
Schematic Scenarios
Heavy metals are introduced into karst aquifers
from surface runoff and spills, reactions in the soil
zone, and dissolution of overlying geologic units and
the host rock (Fig. 1). The metals can be either dis¬
solved or associated with colloids and particulates
(Fig. 2). Dissolved metals may be transported through
the system or may change chemical form and precipi¬
tate as speleothems, onto walls or onto sediments. The
fate of the dissolved metals may depend on the type of
recharge. If dissolved metals are introduced slowly
from fractures in the cave ceiling they may—given the
right chemical conditions—form speleothems. If the
metals arrive as part of a cave stream, they may be
more likely to be transported through the system or to
form coatings on stream sediments. Metals introduced
in particulate form can either travel through the sys¬
tem as suspended sediments and be discharged at a
spring or can be deposited within the aquifer.
Metals deposited within the fractures and conduits
of the aquifer may be stored for extended periods. It is
also possible, however, that they will either be later
dissolved (minerals and coatings) or re-entrained (sedi¬
ments) and be flushed from the system. One of the
final possible storage locations on the flowpath is
spring sediments. Depending on spring morphology
and hydrology, it is possible for some metals to accu¬
mulate in the springbed sediments.
Speleothems and Cave Deposits
Coatings of Fe and Mn are not uncommon in caves.
Although far less common, extensive decorations
incorporating Fe, Mn, and trace metals do occur.
Examples of manganese deposits exist as powders and
coatings in Matts Black Cave in West Virginia and
Rohrer's Cave in Pennsylvania, as stream-cobble coat¬
ings in Butler Cave in Virginia, and as floor fills in
Jewel Cave in South Dakota. Examples of iron-rich
speleothems also exist in Rohrer's Cave, which has
stalactites, stalagmites, and columns of limonite.
Spectacular Fe and Mn speleothems exist in Mbobo
Mkula Cave in South Africa. The metals in this cave
are mobilized when acidic water, produced in the
overlying shales, infiltrates through chert layers in the
limestone. Trace metals are often found in association
with the Fe and Mn deposits. Samples from Rohrer's
Cave have been shown to contain up to 20% heavy
metal oxides. Deposits of trace-metal speleothems also
exist, although they are more rare. For example, mala¬
chite, a copper-carbonate, has been observed in cave
crusts and speleothems. Recent studies indicate that
metal incorporation into speoleothems may be an indi¬
cator of water flow. Speleothems in Grotta di Ernes in
Italy have distinct layers of metals in stalagmite lami¬
nations; their source has been interpreted as associated
with organic-colloids during infiltration.
Suspended and Bed Sediments
Given the near-neutral pH of karst water and the
input of oxygenated surface waters, neither Fe nor Mn
is likely to be present in appreciable concentrations as
dissolved metals in water. This is one factor that dis¬
tinguishes the transport of heavy metals from the more
commonly studied alkaline earth metals ( e.g ., Ca, Mg).
The alkaline earth metals, present from the dissolution
of the carbonate host rock, are almost solely present in
the dissolved form. If dissolved heavy metals are pres¬
ent, they are likely to be oxidized and precipitated as
cave deposits or directly onto particulates in the sedi¬
ments. Once deposited in a cave or spring, sediment
Fe and Mn may be influenced by small-scale transi¬
tions in redox state. This is analogous to metal behav¬
ior in lake and marine sediments where redox
gradients, and the associated Fe and Mn chemistry,
change over short vertical distances.
ENCYCLOPEDIA OF CAVES
CONTAMINATION OF CAVE WATERS BY HEAVY METALS
165
The importance of particulate metal transport has
long been established in surface-water systems where
it has been shown that nearly all heavy metals are
transported in association with solids. Groundwater
studies in the karst aquifer at the Oak Ridge National
Laboratory have also shown that most of the metals
are associated with colloidal or particulate matter.
Metal and radionuclide transport in granular and frac¬
tured aquifers is often attributed to colloidal transport.
Karst aquifers, however, are able to transport much
larger sizes of particles; therefore, metal transport is
not limited by the size of particulate to which the
metal is adhered or incorporated. Data from springs in
Kentucky and Tennessee, in which concentrations of
digested and filtered samples were compared, demon¬
strated that Fe, Mn, and the trace elements were pres¬
ent in particulates larger than 0.45 pm.
Storm-Enhanced Transport of Sediment-
Associated Metals
The ability of the aquifer to transmit sediment is
largely a function of flow velocity; therefore, the flushing
of particles from the overlying soil, entrainment of sedi¬
ments already within the aquifer, and the deposition of
suspended sediments within conduits and in springs are
all key physical processes controlling metal transport
and are closely linked to the groundwater velocity.
Groundwater velocity is not constant through time
in karst aquifers. Systems that rapidly transmit
recharge water exhibit increases in velocity during
storm events. Research has shown that the sediment
transport that occurs during storms controls the trans¬
port of heavy metals as well. Storm-event samples
collected from springs in Kentucky and Tennessee
demonstrated clearly that the heavy-metal concentra¬
tions increased dramatically during storm conditions
(Fig. 3). Data from different-sized storms and different
types of springs in the same area suggest that this rela¬
tionship is consistent and that total metal transport is
episodic and enhanced by storms.
While some "dissolved" (less than 0.45 pm) heavy
metals were present throughout the storm events, their
concentrations were relatively constant in comparison
to the particulate metals. This suggests that "dis¬
solved" metal transport through the aquifer occurs
continuously at low concentrations while particulate
metal transport occurs primarily during storms.
Data from Matts Black Cave indicates that the distri¬
bution of manganese coatings may also be influenced
by the same physical flow processes. The coatings are
much thicker on the ceilings (up to 10 mm) than on
the walls (up to 2 mm) or along the stream (poorly
coated), suggesting that metal deposition and stream
erosion are in competition; therefore, the thickness and
the distribution of metal coatings within caves may
change temporally with storm events.
SUMMARY
Heavy metals are present throughout caves and
karst systems due to both natural and anthropogenic
sources. The solubility and transport of iron and
manganese, two of the most common heavy metals,
are controlled by pH and redox conditions. Although
typically soluble in water in reducing conditions, the
oxidized metals tend to precipitate as cave deposits
8000-1
6000-
g 4000-
0)
2000 -
o-J
cn
3
.Q
0_
■a
c
CO
Cfl
<
(—40
-30
T3
c
co
-10 6
1-0
FIGURE 3 Concentrations of heavy metals in spring water (Beaver Spring). Symbols indicate digested sample concentrations and represent
the total metal transport (both dissolved and particulate components). 1999 data from Fort Campbell Army Base, Kentucky/Tennessee.
ENCYCLOPEDIA OF CAVES
166
c
or on sediments. While speleothems of Fe and Mn
are unusual, coatings of the same are not. Many
metals are never present in the dissolved state and
are introduced, stored, and transported through the
system as colloids or larger particles. Particulate
metal transport is enhanced during storms when
high groundwater velocities permit the metals to be
entrained and suspended. The transport and storage
of trace and contaminant metals are closely linked to
the iron and manganese chemistry. Although some
trace metals are not sensitive to redox conditions,
their behavior is governed by their association with
iron and manganese so they are impacted by redox
conditions.
Bibliography
Allard, B. (1994). Groundwater. In B. Salbu & E. Steinnes (Eds.),
Trace elements in natural waters (pp. 151—176). Boca Raton, FL:
CRC Press.
Borsato, A., Frisia, S., Fairchild, I. ]., Somogyi, A., & Susini, J. (2007).
Trace element distribution in annual stalagmite laminae mapped
by micrometer-resolution X-ray fluorescence: Implications for
incorporation of environmentally significant species. Geochimica et
Cosmochimica Acta, 71, 1494—1512.
Hill, C. A. (1982). Origin of black deposits in caves. National
Speleological Society Bulletin, 44(1), 15—19.
Hill, C. A., & Forti, P. (1997). Cave minerals of the world (2nd ed.).
Huntsville, AL: National Speleological Society.
Horowitz, A. J. (1991). A primer on sediment—trace element chemistry
(2nd ed.). Chelsea, MI: Lewis Publishers.
McCarthy, J. F., & Shevenell, L. (1998). Processes controlling colloid
composition in a fractured and karstic aquifer in eastern
Tennessee, USA. Journal of Hydrology, 206(3—4), 191—218.
Northrup, D. E., & Lavoie, K. H. (2001). Geomicrobiology of caves: A
review. Geomicrobiology Journal, 18(3), 199—222.
Salomons, W., & Forstner, U. (1984). Metals in the hydrocycle. Berlin:
Springer-Verlag.
Sasowsky, I. D., & White, W. B. (1993). Geochemistry of the Obey
River Basin, north-central Tennessee: A case of acid mine drain¬
age in a karst drainage system. Journal of Hydrology, 146, 29—48.
Siegel, F. R. (2002). Environmental geochemistry of potentially toxic
metals. Berlin: Springer-Verlag.
Vesper, D. J., & White, W. B. (2003). Metal transport to karst springs
during storm flow: An example from Fort Campbell, Kentucky/
Tennessee, U.S.A. Journal of Hydrology, 276, 20—36.
White, W. B., Vito, C., & Scheetz, B. E. (2009). The mineralogy and
trace element chemistry of black manganese oxide deposits from
caves. Journal of Cave and Karst Studies, 71, 136—143.
CONTAMINATION OF CAVE WATERS
BY NON AQUEOUS PHASE LIQUIDS
Caroline M. Loop
Groundwater Management Associates
Due to their unique properties, nonaqueous phase
liquids (NAPLs) are widely used as solvents, insulators,
and fuels. Some of the same valuable properties.
however, make NAPLs toxic and difficult to remove
from soils and groundwater once they have been
spilled. In the heterogeneous karst subsurface, the rate
of movement of NAPL contamination can vary by
orders of magnitude. Tailoring a conceptual model to a
particular NAPL release and subsurface characteriza¬
tion is necessary to develop multiple working hypothe¬
ses and better guide detection and monitoring
techniques.
NAPL CHARACTERISTICS
AND SOURCES
Four characteristic chemical properties that influence
the behavior of NAPLs in the environment are solubility,
density, vapor pressure, and viscosity. Nonaqueous
phase liquids must, by definition, have limited solubility
in water, allowing them to remain in a separate phase
(Fig. 1). Thus, the dissolved, or aqueous phase, concen¬
tration of a nonaqueous phase liquid is moderate to low.
Unfortunately, even the limited aqueous phase concen¬
tration is often much higher than the maximum concen¬
tration level established by the U.S. Environmental
Protection Agency (Table 1). Solubilities are typically
less than 5000 mg/L for chlorinated solvents, less than
2000 mg/L for most gasoline components, and less than
1.0 mg/L for polychlorinated biphenols (PCBs).
Nonaqueous phase liquids are divided between
those that are less dense than water (LNAPLs) and
those that are more dense than water (DNAPLs).
Hydrocarbons, including benzene and ethylbenzene,
are common LNAPLs, as are vinyl chloride and
styrene. LNAPLs are often used in the synthesis of
plastics, and as gasoline. With a density less than water,
they will be present as a separate phase above the water
table (Fig. 2). The water table will be depressed in pro¬
portion to the thickness and density of the LNAPL
immediately above it. DNAPLs include some very toxic
chemicals such as perchloroethylene (PCE), PCBs, and
insecticides and herbicides such as lindane and atra-
zine. PCE and its breakdown products were commonly
used as solvents, especially in dry cleaning. (Their use,
however, is decreasing due to their highly toxic nature.)
PCBs, for example, the Aroclor formulations, were
commonly used as liquid insulators in capacitors and
transformers. DNAPLs will sink below the water
table and continue to migrate deeper in an aquifer until
they meet sufficient resistance (Fig. 3).
NAPLs can be characterized based on their vapor
pressure, which is a measure of the tendency of the
compound to evaporate from a pure liquid of the com¬
pound. NAPLs are either volatile, with a vapor pres¬
sure greater than 10 4 atm, or semivolatile, with a
vapor pressure between 10 4 and 1 0 _11 atm. Benzene
Encyclopedia of Caves.
© 2012 Elsevier Inc. All rights reserved.
CONTAMINATION OF CAVE WATERS BY NONAQUEOUS PHASE LIQUIDS
167
P = 1
rri;
FIGURE 1 Diagram of important properties for characterizing organic contaminants, p = density (g/cm 3 ), m ; = solubility of contaminant
i in water, and P; = vapor pressure of contaminant i.
TABLE 1 Density, Vapor Pressure, and Solubility Values for Common NAPLs
Compound
Density (g/cm 3 )
Vapor Pressure (kPa)
Solubility (mg/L)
U.S. EPA MCL (mg/L)
Toluene
0.86
3.75
500
1
Benzene
0.88
12.4
1800
0.005
o-Xylene
0.88
0.88
175
10
Ethylbenzene
0.90
1.24
150
0.7
Vinyl chloride
0.91
334
2800
0.002
Styrene
0.91
0.62
310
0.1
Aroclor 1242
1.4
5.29 x 10” 5
0.24
0.0005
Trichloroethylene (TCE)
1.5
9.63
1100
0.005
Carbon tetrachloride
1.6
14.9
760
0.005
Tetrachloroethylene (PCE)
1.6
2.48
200
0.005
Lindane
1.9
8.19 X NT 6
7.5
0.0002
Note : Equilibrium aqueous phase solubility concentrations are orders of magnitude larger than the U.S. Environmental Protection Agency's maximum
concentration level (MCL).
has a high vapor pressure, as can be seen when gasoline
fumes rise in the summer, whereas PCBs are almost
wholly semivolatile. Because many NAPLs have a high
vapor pressure, they had sometimes historically been
thought to fully evaporate when poured on the ground.
This belief turned out to be incorrect and, in fact, such
practice significantly contaminated underlying soils
and aquifers. As discussed later, vapor pressure may be
helpful in locating NAPL contaminants in karst.
Viscosity influences the subsurface mobility of non-
aqueous phase liquids in that the less viscous the liq¬
uid, the farther it can migrate into pores and fractures.
Chlorinated solvents such as trichloroethylene
(TCE) and PCE have very low viscosities. Creosote
ENCYCLOPEDIA OF CAVES
168
Water
table
1 TRAPPING IN REGOLITH
1A RESIDUAL LNAPL
IB POOLING ON LOW
PERMEABILITY LAYER
5 TRAPPING IN HORIZONTAL
CONDUITS ABOVE
BEDDING PLANES
Ground-
water
flow
◄-
6 SOIL FILLING OF SOLUTION
2 POOLING AT TOP OF ROCK WIDENED JOINTS
3 AQUIFER ENTRY POINTS
7 TRAPPING IN FRACTURES
4 TRAPPING IN SOLUTION
WIDENED JOINTS
FIGURE 2 Distribution of LNAPLs in a hypothetical karst setting. Source: After Wolfe et al. (1997).
wood-treatment compounds behave as DNAPLs and
are characterized by a high viscosity. Fluids can
increase in viscosity with time, called weathering in the
petroleum industry, as they lose more volatile compo¬
nents. It is easier for flowing water, such as might be
found in a cave stream, to entrain less viscous fluids,
and move them farther from the source.
Although petroleum compounds are formed and
present in the subsurface naturally, they are rarely
found in near-surface aquifers and soils unless they
have been spilled. Other NAPLs, including PCBs, chlo¬
rinated solvents, and agricultural chemicals such as
lindane and atrazine, have been developed in laborato¬
ries and tailored to industrial uses. Freon 11, also
known as trichlorofluoromethane, is one such chemical
that was used extensively until the toxic effects of
spills and leaks was found to exceed its practical use.
The distribution of NAPLs in karst soils and aquifers
is nearly always the result of human misuse, whether
by transportation accidents, leaking tanks, or improper
disposal.
TRANSPORT INTO KARST AQUIFERS
Both the volume and rate of release can play a key role
in the movement and distribution of NAPL contami¬
nants. For example, a gasoline spill from a 10-gallon
bucket will move through and be sorbed to soil and
regolith differently than a slow leak from a 1000-gal-
lon tank that can build up a pressure head of product,
which drives it farther into the surface. On the extreme
ENCYCLOPEDIA OF CAVES
CONTAMINATION OF CAVE WATERS BY NONAQUEOUS PHASE LIQUIDS
169
1 TRAPPING IN REGOLITH
1A RESIDUAL DNAPL
IB POOLING ON LOW
PERMEABILITY LAYER
3 POOLING IN BEDROCK
DIFFUSE-FLOW ZONE
4 POOLING IN CONDUIT
2 POOLING AT TOP OF ROCK 5 POOLING IN FRACTURES
ISOLATED FROM FLOW
FIGURE 3 Distribution of DNAPLs in a hypothetical karst setting. Source: From Wolfe et al. (1997).
end of this scale is the gasoline tanker truck that turns
over and releases thousands of gallons quickly onto
the ground surface, which then flows overland and
into a sinkhole or sinking stream where it can more
quickly be transported into a bedrock aquifer.
Epikarst
Climate can also impact the rate and volume of
NAPL entering the karst subsurface. Karst soils are usu¬
ally thin, and they overlie the regolith and irregular bed¬
rock of the epikarst. In a wet climate three things may
occur to move the NAPLs more quickly through an
unsaturated soil and possibly epikarst zone. First, when
water coats soil, organic, or regolith particles, if NAPL is
spilled, it travels as a nonwetting fluid and will be held
less securely between particle grains because it is, in a
sense, lubricated by the aqueous phase.
Second, and not inconsequentially, soil and regolith
pores may periodically be flushed by precipitation. In
this case, pressure head builds up from the accumula¬
tion of recent precipitation, and even light nonaqueous
phase contaminants can be forced to migrate to deeper
pores. If a less permeable layer is present in the soil,
the influx of rainwater may cause an NAPL that was
temporarily pooled above the layer to migrate horizon¬
tally to an area where the less permeable layer is no
longer present. At that point, the NAPL may resume
its vertical flow toward the water table. While the
NAPL is held in soil or regolith pores, it may dissolve
into the aqueous phase, which may then potentially be
flushed farther through the system during precipita¬
tion events. Thus, NAPLs held in the soil or regolith
can continue to dissolve and be long-term sources of
aqueous phase contamination to an aquifer with which
the NAPL is not in direct contact.
ENCYCLOPEDIA OF CAVES
170
c
A third influence of climate on the transport of
NAPLs into karst aquifers is the impact on sinkhole for¬
mation and collapse. Sinkholes form localized catch¬
ments for storm runoff in karst regions. Some sinkholes
are plugged with soil so that NAPLs held within them
may migrate downward only slightly more quickly
than in nonsinkhole soils, with the only difference being
a comparatively larger volume of precipitation perco¬
lating through. On the other end of the spectrum, some
sinkholes have open drains so that any runoff captured
will be immediately carried into the subsurface. In
some cases, however, a soil plug may itself be subject to
piping failure, especially during large storms. When
this occurs, both the soil and the NAPL are transported
into the subsurface and leave an open hole for further
contaminants to more quickly enter the system.
The top of bedrock in karst is often an irregularly
sculptured surface that may be undulating or may
have deep crevices along joints and fractures separated
by intermediate pinnacles. Water, moving downward
from overlying soil into the epikarst, must often move
laterally for substantial distances before finding an
open fracture or shaft that will permit vertical move¬
ment into the unsaturated zone of the bedrock. Like
groundwater, infiltrating contaminants may be held
for extended periods of time, pooling along the top of
relatively impermeable bedrock. While still in the
vadose zone, pools resting along the top of the bedrock
can be described as microphase pools, which have col¬
lected in pores between regolith and other particles.
The pools should be distinguished from continuous
macrophase pools that may exist floating on water
(LNAPLs) or below water (DNAPLs), but not distrib¬
uted within pore space.
In the event that the top of the limestone or dolo¬
mite bedrock is below the water table, LNAPL pools
will most likely resist resting on the bedrock; rather
they will tend to remain in the soil or regolith directly
above the water table. Wells intersecting a few feet
of LNAPL above the water table are not uncommon
in contaminated areas. DNAPLs, in contrast, are able
to migrate below the water table, and may pool on
less permeable layers, such as the bedrock surface. In
this case, water will flow in the direction of hydrau¬
lic gradient; however, DNAPLs often flow in the
direction of bedrock dip. In one specific case, an
aqueous phase plume developed in one direction,
whereas the DNAPL itself migrated down-dip in a
different direction.
Fractures, Open Drains, and Sinking Streams
Karst bedrock is often dissected by vertical or near¬
vertical fractures and joints, many of which have been
solutionally widened. These provide fast paths for
NAPLs into the subsurface aquifer once the critical
height for entry into the fracture has been exceeded.
The critical height of a DNAPL pool is a function of
the interfacial tension between the NAPL and water,
the wetting angle between the DNAPL and the solid
surface in the presence of pore water, the density dif¬
ference between the NAPL and water, and the fracture
aperture. Both LNAPLs and DNAPLs will migrate in
similar ways through fractures above the water table.
One exception is that the higher density of the
DNAPLs will cause them to move more quickly under
gravity. Viscosity will also play an important role in
the ability to move through a given aperture. Storm
flow can, again, add to the pressure above the non-
aqueous fluid, causing it to migrate farther down the
fracture.
Once at the water table, LNAPLs and DNAPLs will
behave very differently, with fractures acting as tempo¬
rary traps for LNAPLs and preferential flowpaths for
DNAPLs. In a scenario where NAPL is moving down a
fracture, which farther down intersects the water table,
and then joins with a phreatic conduit, LNAPL will
remain in the fracture above the water table, whereas
DNAPL will continue to travel down the fracture into
the conduit, as long as the critical DNAPL pool height is
exceeded (Figs. 2 and 3). If drought conditions lower the
water table so that water flowed like a surface stream in
the conduit, the same LNAPL could potentially travel
down the fracture and float on the water in the free sur¬
face stream down the length of the conduit. When the
water level rises, the LNAPL will again get caught in a
fracture or a sump and again be trapped, possibly even
below the water table.
In this same scenario, there will be much less differ¬
ence in the transport of the DNAPL with a variation in
water level. Because it is denser than water, the
DNAPL, with sufficient pool height, can move through
the fracture and into the conduit. An air-filled fracture
would likely offer less resistance than a water-filled
one, yet in both cases, the DNAPL will not be held in
the fracture indefinitely. In the event that the critical
pool height is not exceeded, the NAPL will diffuse
into the rock matrix over time. Once the NAPL is
depleted, the contaminated aqueous phase will act to
reverse the concentration gradient and dissolve into
freshwater undersaturated with the NAPL, moving
through the fractures.
Open drains, such as those sometimes found in
sinkholes, as well as sinking streams, offer a very
quick path to the underlying karst aquifer. NAPLs can
be poured into the open drains. They can also either
float in the case of LNAPLs, or be pushed along in the
case of DNAPLs in surface streams. Sinking streams
and open drains allow contaminants direct access to
ENCYCLOPEDIA OF CAVES
CONTAMINATION OF CAVE WATERS BY NONAQUEOUS PHASE LIQUIDS
171
quick subsurface flow and can allow NAPLs to move
at a rate of kilometers per hour. This is very important
to take into consideration in an area where a spring
acts as the local water-supply source.
STORAGE IN KARST AQUIFERS
Once NAPLs have entered a karst aquifer, they can
be a source of contamination for thousands of years,
given their low solubility values and the difficulty often
encountered when attempting to retrieve them. Natural
degradation is faster for hydrocarbons than chlorinated
compounds such as chlorinated solvents and PCBs, but
for a large spill may take many decades. Little is known
about the ability of microbes present in karst systems to
degrade any of the NAPL compounds. Hence, it is
important to consider the locations in which NAPL
might be stored in the karst aquifer.
Pools in Conduits
Macrophase NAPL pools—free surface pools not con¬
fined to pore spaces—can be present in conduits.
DNAPL pools could be present at the bottom of a flow¬
ing subsurface stream, whereas LNAPL pools would
float on top of the water. Pools such as these could also
be present in dry cave passages, but in all cases require a
large volume of contaminant, quickly injected to prevent
either (1) seepage into sediments or (2) horizontal move¬
ment to a more permeable material. In turbulent flow,
which characterizes most flow in subsurface streams,
NAPL can be entrained or carried as an emulsion until
the water velocity again decreases. Macrophase pools
have a low surface-area-to-volume ratio, so dissolution
is limited. When exposed to air, NAPL compounds may
volatilize within the cave passages.
Fractures
Fractures may intersect conduits in all directions. As
discussed earlier, fractures extending from the roof of
a conduit may act to indefinitely trap LNAPLs, but
fractures in walls may also act to isolate NAPLs.
DNAPLs are especially likely to migrate into subaque¬
ous fractures in conduit floors (Fig. 3). Once they reach
a depth where the critical pool height is no longer
exceeded, they diffuse into the fracture matrix, from
which they may act as a long-term source for aqueous
phase contamination.
Matrix and Vugs
NAPLs can be stored in the bedrock matrix or in
larger vugs. Diffusion of NAPLs into and out of the
matrix occurs on a relatively slow scale. The quickest
period of activity in the matrix may be the result of
the flushing of pores during storm activity, as when the
potentiometric surface changes slope from down to the
conduit during baseflow, to a mound above the conduit
during intense storm flow. Storage in bedrock matrix
and vugs may represent the portion of a contaminant
spill most resistant to short-term remediation.
Sediment
NAPLs may sorb to or be present in pores between
sediments. In the first case, the NAPL may travel on
the sediment, ready to desorb once it reaches a less
concentrated aqueous environment. In the second case,
once the sediment pile is disrupted, the NAPL is again
free to move with the water, or down-dip in the case
of DNAPLs. NAPL can be held in the sediment pile by
capillary forces or a microphase pool. With storm
movement, a fresh flush of contaminants may be remo¬
bilized when NAPL is associated with sediment piles
in karst conduits.
DETECTION OF NAPL IN KARST
Springs and Caves
In an area where springs and caves are accessible,
monitoring should begin by sampling these locations
during storm and baseflow conditions. Springs are
good monitoring sites as they are often down-gradient
from spills or leaks, and if previous dye traces exist,
they can help to define a groundwater basin and evalu¬
ate the risk to nearby populations. Spring sediments
should be observed. Conductivity and temperature are
inexpensive tools for estimating the timing of a spring's
response. In caves, one must be careful if NAPLs are a
suspected contaminant. In one instance in the 1960s, a
carbide lamp ignited a gasoline spill in a cave and killed
several people, some by flames, but others by asphyxia¬
tion. Less dramatic is the instance in which a volatile
NAPL produces toxic air within a cave.
Wells
As mentioned previously, wells can intersect meters
of LNAPL product above the water table. Fracture
trace analysis can help to site wells in some instances,
and wells can also be used for dye trace studies. In
karst, wells must be carefully considered, especially
with the slow diffusion into and out of the bedrock
matrix as compared to the rate at which other trans¬
port processes may be operating in the aquifer.
DNAPLs are likely more difficult to find using wells.
ENCYCLOPEDIA OF CAVES
172
c
because they may have migrated into deeper fractures
in the aquifer, and they are harder to direct by altering
the hydraulic gradient. With either type of NAPL, but
especially with DNAPLs, one must be very careful to
case wells properly in a contaminated area. An open
borehole is an excellent way to transmit dissolved con¬
tamination or DNAPL to an underlying, formerly
uncontaminated aquifer.
Soils
When an NAPL is spilled, it often passes through
soil, which can be collected for quantitative laboratory
analysis. Soil sampling may be useful for verifying the
type of NAPL and whether any former spills took
place at the site. Soil sampling in a sinkhole is not
recommended, due to possible piping failure. Soil gas
sampling may be very useful, especially for LNAPLs.
Contaminants can volatilize, with the resulting gas
moving up through the soil. Depending on the season,
the contaminant, and the subsurface configuration, soil
gas sampling may be a helpful technique for identify¬
ing and locating an NAPL product.
SUMMARY
Due to the heterogeneous nature of the karst sub¬
surface, the rate of either water or NAPL transport
through karst aquifers is highly variable. Both the
quantity and timing of NAPL releases are important
for understanding how the pollutant might be trapped
in the aquifer. Individual NAPL characteristics such as
density, solubility, vapor pressure, and viscosity are
also key to recovering a contaminant from soils or
groundwater. Over time, NAPL held in the epikarst or
matrix can dissolve into the aqueous phase. Aqueous
concentrations can be toxic and persist for many years,
especially in the case of chlorinated compounds, which
are naturally degraded more slowly than hydrocar¬
bons. However, NAPL from large spills may move
through a conduit on the order of kilometers per hour.
All information about a specific karst system, includ¬
ing dye traces, spring response, depth to and shape of
the bedrock surface, is important for evaluating the
potential of NAPLs to be held in and transported
through the subsurface. The study of NAPL contami¬
nation in karst aquifers is a relatively new aspect of
karst science, and in the future will certainly be
enhanced by additional case studies and research.
Bibliography
Black, D. F. (1966). Howard's cave disaster. National Speleological
Society News, 24, 242—244.
Crawford, N. C., & Ulmer, C. S. (1994). Hydrogeologic investigations
of contaminant movement in karst aquifers in the vicinity of a
train derailment near Lewisburg, Tennessee. Environmental
Geology, 23(1), 41-52.
Ewers, R. O., Duda, A. J., Estes, E. K., Idstein, P. J., Ik Johnson, K. M.
(1991). The transmission of light hydrocarbon contaminants in lime¬
stone (karst) aquifers. Proceedings of the third conference on hydrogeol¬
ogy, ecology, monitoring, and management of ground water in karst
terranes. Dublin, OH: Association of Ground Water Scientists &
Engineers, National Ground Water Association.
Jancin, M., & Ebaugh, W. F. (2002). Shallow lateral DNAPL migration
within slightly dipping limestone, southwestern Kentucky.
Engineering Geology, 65, 141—149.
Krothe, N., Fei, Y., McCann, M. R., & Cepko, R. P. (1999).
Polychlorinated biphenyl (PCB) contamination of a karst aquifer
in an urban environment, central Indiana, USA. In J. Chilton
(Ed.), Groundwater in the urban environment: Selected city profiles.
Rotterdam: A. A. Balkema.
Loop, C. M., & White, W. B. (2001). A conceptual model for DNAPL
transport in karst ground water basins. Ground Water, 39(1),
119-127.
Mercer, J. W., & Cohen, R. M. (1990). A review of immiscible fluids
in the subsurface: Properties, models, characterization, and reme¬
diation. Journal of Contaminant Hydrology, 6, 107—163.
Pankow, J. F., & Cherry, J. A. (1996). Dense chlorinated solvents and
other DNAPLs in groundwater. Waterloo, Ontario, Canada:
Waterloo Press.
Schwarzenbach, R. P., Gschwend, P. M., & Imboden, D. M. (1993).
Environmental organic chemistry. New York: John Wiley & Sons.
Wolfe, W. J., Haugh, C. J., Webbers, A., & Diehl, T. H. (1997).
Preliminary conceptual models of the occurrence, fate, and trans¬
port of chlorinated solvents in karst regions of Tennessee, U.S.
Geological Survey Water-Resources Investigations Report 97-4097.
Reston, VA: U.S. Geological Survey.
COSMOGENIC ISOTOPE DATING
OF CAVE SEDIMENTS
Darryl E. Qranger and Derek Fabel
Purdue University, ^University of Glasgow
INTRODUCTION
Natural curiosity prompts both cave explorers and
first-time visitors to wonder "How old is this cave?"
and "Why is it here?" Scientists have more specific rea¬
sons to study and date cave sediments. For example,
geomorphologists use caves to learn about landscape
evolution, or the sequence of events that shaped the
rivers, hills, and valleys around us. Paleontologists
study fossils in cave sediments to learn about animal
and plant evolution and about the ecological commu¬
nities that lived long ago. Paleoanthropologists study
our ancestors' bones that are found in caves—whether
from cave dwellers, or often those that were eaten in
the caves. These fossils and their dates help teach us
about our own human origins. Archaeologists search
for clues about human use of caves. Some scientists
Encyclopedia of Caves.
© 2012 Elsevier Inc. All rights reserved.
COSMOGENIC ISOTOPE DATING OF CAVE SEDIMENTS
173
also study caves for their own sake, to learn about how
water flows through rock, and how the spectacular and
labyrinthine underground environment is formed. All
of these various fields require information about the
age of the cave and its contents.
Caves are important across so many fields of science
because the conditions underground are so protected
and stable that minerals, rocks, and fossils can be pre¬
served in exquisite condition for millions of years.
Sediments and fossils on the ground surface are gradu¬
ally but constantly weathered and eroded over time.
The landscape on the surface changes slowly but
surely as hillslopes are worn down, rivers incise or fill
their beds with sediment, and forests grow and recede.
By contrast, caves and their sediments can be nearly
pristine, with delicate minerals, fossils, and sediments
still intact. Although the hill or mountain that a cave
is formed in may change over time, the cave itself is
contained in solid rock, so it can maintain its original
shape until the entire mountain itself is eventually
eroded away.
Because knowing the age of a cave or its sedi¬
ments is critical for learning about the past, several
techniques for dating cave sediments and minerals
have been developed. Each dating scheme has its
own advantages and limitations. Some of the dating
techniques such as paleomagnetism, uranium-series
disequilibrium, and radiocarbon dating have become
well established and widely used. This article concerns
another, relatively new, dating technique that employs
radioactive nuclides to date when sediment was
brought into a cave. The method is called cosmogenic
nuclide burial dating (Granger and Muzikar, 2001).
RELATIVE VERSUS ABSOLUTE
DATING TECHNIQUES
When attempting to date a particular cave or its
contents, there are several possible techniques to con¬
sider, depending on the age and the particular fossils
or minerals in question. Some of these dating tech¬
niques are relative, indicating whether one thing is
older or younger than another but not the exact age of
either. For example, characteristic fossils in a cave can
be used to place it relative to other deposits. Other dat¬
ing techniques are absolute, meaning they give a
numeric age that does not depend on correlations with
any other site. Absolute ages are defined using some
sort of "clock" that operates at a known and constant
rate. By far the most widely used and reliable clock is
radioactive decay.
To understand radioactive decay, it is helpful to
first review the basic structure of the atomic nucleus.
A nucleus is made of protons and neutrons. The number
of protons in a nucleus determines to a large degree the
way that an atom behaves; in fact, the elements of the
periodic table are defined by the number of protons
they have. Sometimes two different atoms may have the
same number of protons, but different numbers of neu¬
trons. In this case, the atoms are of the same element,
but they have different masses. These two atoms are
called isotopes. For a given element, some isotopes are
stable, remaining unchanged over time. Other isotopes
are radioactive, in which case the nucleus spontaneously
breaks apart, losing mass and energy in radioactive
decay. Radioactive decay occurs at a constant rate for
any given isotope. If a certain amount of a radioactive
isotope is contained in a rock, then half of that amount
will decay in a characteristic time called the half-life. Half
of the remaining half will decay after another half-life
and so forth ad infinitum. Here is the clock for dating
radioactive materials. If the original amount of a radio¬
active isotope is known, and the remaining amount can
be measured, then the difference indicates the amount
of time that has passed. The trick for cave scientists is
to find a material with a known initial amount of radio¬
active isotopes. Cosmogenic nuclides provide just such
a case.
BURIAL DATING WITH
COSMOGENIC NUCLIDES
One way to date cave sediments is by determining
the radioactive loss of cosmogenic nuclides. (The term
nuclide refers to atoms regardless of their element, as
opposed to isotope, which always refers to atoms of the
same element.) Cosmogenic nuclides are produced by
cosmic rays—energetic particles coming from outer
space that constantly bombard Earth. Most of the cos¬
mic rays are absorbed in the atmosphere, producing
particles such as 14 C that is used for radiocarbon dat¬
ing. Some secondary cosmic rays reach the ground sur¬
face and cause nuclear reactions inside rocks and
minerals found within a few meters of the surface.
During these nuclear reactions, the nuclei inside the
mineral grains are broken apart, forming lighter
nuclides. By chance, some of the products of these
reactions are radioactive. For example, we can consider
reactions in the mineral quartz. Quartz has a chemical
formula Si0 2 . Silicon nuclei each have 14 protons, and
most of them have 14 neutrons. The common silicon
nucleus thus has a mass of 28, written 28 Si. An incom¬
ing cosmic ray particle will occasionally break apart a
silicon nucleus. If a proton and a neutron are lost, then
the 28 Si is converted into 26 Al. Fortunately for dating,
26 A1 is radioactive, with a half-life of 700,000 years.
Another reaction that occurs in quartz is the conver¬
sion of 16 Q to 10 Be through the loss of 4 protons and
ENCYCLOPEDIA OF CAVES
174
c
2 neutrons. Beryllium-10 is also radioactive, with a
half-life of about 1.4 million years. These two different
radioactive nuclides are produced in the same quartz
grain, and are the key to dating sediment burial in
caves. A diagram of the reaction producing 26 A1 is
shown in Figure 1.
Over thousands of years, as rocks are exposed to
cosmic rays they build up an inventory of 26 A1 and
10 Be. Many repeated measurements of quartz grains on
the ground surface have shown that 2(1 A1 is produced
about 6.8 times faster than in Be. Since both the produc¬
tion rates and the half-lives are known, the concentra¬
tions of 26 A1 and 10 Be in rocks exposed to cosmic
rays can be calculated. For most rocks at the ground
surface, the 26 Al: 10 Be ratio is 6.8:1. If quartz is brought
into a cave, though, the grains are shielded from cos¬
mic rays so 26 A1 and 10 Be are no longer produced.
After the quartz-bearing sediment is brought into the
cave, radioactive decay gradually lowers the concentra¬
tions of both 26 A1 and 10 Be. Aluminum-26 decays faster
than Be, so the original Al: Be ratio decreases over
time. After 700,000 years half of the 26 Al is gone, but
only 30% of the 10 Be has decayed. The original ratio of
6.8:1 has thus been lowered to about 4.8:1. The 26 Al: 10 Be
ratio provides the radioactive clock that we can use to
date cave sediments, with an original ratio of 6.8:1 that
decreases exponentially over time. Figure 2 shows the
decay of the two nuclides, and the 26 Al: 10 Be ratio as
a function of time. As the concentrations of 26 A1 and
10 Be get smaller and smaller over time, they become
more difficult to measure. The practical limit to mea¬
surement usually occurs after about 5 million years of
burial for this reason.
Aluminum-26 and 10 Be in sediment can only be mea¬
sured using a very sensitive technique called accelerator
mass spectrometry (AMS). This is because the concentra¬
tions of the cosmogenic nuclides are extremely small.
For example, only 5 atoms of 10 Be may be produced in
a gram of quartz in an entire year, and a sample
28 Si(n, p2n) 26 A!
FIGURE 1 An example of a spallation-type nuclear reaction, in
which an incoming cosmic ray neutron (left) impacts a 28 Si nucleus
(center), to knock off two neutrons and a proton (right), making
radioactive 26 A1. Neutrons are indicated by the darker-colored
spheres, protons by the lighter-colored spheres.
Time (millions of years)
FIGURE 2 A graph of the concentrations of 26 A1 and I0 Be (arbitrary
units), and the 26 Al: 10 Be ratio in quartz grains over time. The grains
are washed into a cave with an original 26 Al: 10 Be ratio of 6.8:1. Because
26 A1 decays faster than 10 Be, the 26 Al: 10 Be ratio decreases over time.
The 26 Al: 10 Be ratio can thus be used to date when the sediment was
deposited in the cave.
may contain less than a million atoms of 10 Be. AMS is
capable of measuring an isotope ratio ( e.g ., 10 Be/ 9 Be
or 26 A1/ 27 A1) as low as 10 -16 . That is, if there are
10 16 atoms of the common isotope 27 A1, then AMS can
detect a single atom of cosmogenic 26 Al.
Requirements for Burial Dating
As with any dating technique, it is important to
consider the circumstances for which dates will be
reliable, and those for which dates will be unreliable
or impossible to obtain. Burial dating has several
rather strict requirements. First, the sediment must
have washed into the cave from outside, that is, it
must be allochthonous. Otherwise, there would be
none of the cosmogenic nuclides to begin with.
Second, the sediment must contain the mineral quartz,
because that is the mineral for which we know the
production rates of 26 Al and 10 Be. Quartz is not always
common in cave-forming bedrock, so even if there is
allochthonous sediment in the cave it may not be dat¬
able. Third, the sample must be buried underground
by at least 20 meters for the technique to be reliable
over millions of years. It is possible to date more
shallowly buried samples, but this introduces higher
uncertainties. Fourth, there are limitations on the
burial times. The sediment cannot have been buried
more than about 5 million years, or the 26 Al and 10 Be
are no longer detectable. Uncertainties in measuring
26 Al and 10 Be are usually 3—5%, making it difficult
to achieve burial dates more precise than about
100,000 years. So the sediment must have been buried
for at least 100,000 years. Finally, the sediment must
ENCYCLOPEDIA OF CAVES
COSMOGENIC ISOTOPE DATING OF CAVE SEDIMENTS
175
come into the cave without a prior history of burial.
If the sediment were, for instance, buried at the bottom
of a doline for a million years and then washed into the
cave, then the burial age would account for the total
time spent buried, not just that in the cave. Although
these uncertainties limit the application of burial dating
somewhat, there are many situations for which the tech¬
nique is ideal. We describe two examples below.
Example 1: The Development of Mammoth
Cave, Kentucky
Mammoth Cave, the longest known cave in the
world, has developed alongside the Green River in
Kentucky. It is an example of a water-table cave, or one
that has developed nearly horizontal passages that are
closely controlled by the level of groundwater flow
(Palmer, 1981). At Mammoth Cave, the water level is in
turn controlled by the elevation of the Green River.
Rainfall on the nearby Pennyroyal plateau quickly infil¬
trates the karst bedrock until it reaches cave passages
that are filled or nearly filled with water. These under¬
ground streams then flow toward the Green River,
passing through a sandstone-capped limestone plateau
to discharge as springs on the Green River. In addition
to the large recharge area that captures abundant rain¬
fall, the sandstone-capped plateau is a major reason
why Mammoth Cave is so long. The sandstone is a rock
made of cemented quartz sand that is very resistant to
erosion. Over time, as the Green River has cut through
the sandstone and into the underlying limestones, cave
passages have formed at successively lower levels. In
many landscapes, the old cave passages above would
be destroyed by erosion as new cave passages were
formed below. However, at Mammoth Cave the sand¬
stone is so resistant to erosion that the older passages
have not eroded away. The sandstone forms ridges
beneath which are preserved stacks of cave passages,
the oldest passages near the top and the youngest
passages at modern river level.
The preservation of old passages at Mammoth Cave
provides a wonderful opportunity for studying how
the cave has developed, and how the Green River
has incised and aggraded over time. Geologist and
hydrologist Art Palmer has spent many years carefully
working out the sequence of events that are encrypted
within Mammoth Cave's passages. It is very difficult,
though, to decipher the history of the cave without
dates that can be used to tie passages together across
the cave system, and to match episodes of cave devel¬
opment with other geologic events. Burial dating with
cosmogenic nuclides has provided a set of dates that
helps show how the development of the Mammoth
Cave System has been strongly influenced by climate
change and the growth of ice sheets across North
America. Although ice sheets did not reach Mammoth
Cave itself, they did impact the Green River, which
alternately incised and aggraded, forming sets of pas¬
sages beneath the ridges at Mammoth Cave.
Burial dating works at Mammoth Cave because
quartz pebbles from conglomerates within the sandstone
upland are carried into the cave by sinking streams.
These streams carry the pebbles through the cave system
and into the Green River. When the river incises deeper,
new passages are formed at lower elevations. The old
passages are no longer occupied by streams, so whatever
quartz pebbles were being carried through the cave
are left in place, to sit within the now-abandoned cave
passages. These packages of quartz-bearing sediment
can be found throughout nearly the entire cave system.
It is important to realize that the sediments indicate not
when the passage formed, but when the passage was
abandoned. It is only through careful analysis of the
cave that the abandonment of one passage can be linked
to the growth of another.
Granger et al. (2001) dated sediments from through¬
out the Mammoth Cave System. These samples reach
ages up to 3.5 million years old in the uppermost
levels of the cave, and tell an interesting story of how
the cave developed over time. First, the dates show
that the Mammoth Cave System is quite old. If the
sediments that fill the cave are up to 3.5 million years
old, then the cave itself must be significantly older
than that! The upper levels of the cave system sub¬
stantially predate the ice ages, which began roughly
2.5 million years ago. The cave, then, reveals how the
landscape of central Kentucky responded to this major
climate change. The initial response to climate change
seems to be that the entire cave system filled up
with sediment. Visitors to Mammoth Cave will notice
telltale signs of sediment everywhere, even in nooks
and crannies on the passage ceilings. These sediments
show that most of the cave was filled up at about
2.4 million years ago, which in turn indicates that
the Green River valley must be filled up as well. The
landscape response to climate change was river aggra¬
dation, perhaps due to increased hillslope erosion
that would have supplied more sediment than the
river could carry. The next chapter in the story of
Mammoth Cave is river incision and cave develop¬
ment at lower levels. Mammoth Cave is developed in
levels, indicating that river incision was episodic. Each
new level represents a pulse of river incision. Granger
et al. (2001) found that these incision pulses correlate
with large glaciations that covered most of eastern
Canada and the northeastern United States, advancing
as far as the northern edge of Kentucky. These large
ice sheets completely reorganized river systems that
were either buried beneath the ice or blocked by great
ENCYCLOPEDIA OF CAVES
176
c
ice dams. In fact, the modern courses of rivers such as
the Ohio River, the Missouri River, and the northern
Mississippi River were shaped along the edges of ice
sheets. Each time that the Ohio River or Mississippi
River incised due to glaciation or sea-level lowering,
the Green River followed closely behind. Mammoth
Cave thus records in its passages a key to the history
of eastern North America over the past 3.5 million
years.
Example 2: Caves and Human Evolution
Caves have always held an important place in
human evolution, because so many fossils have been
found in them. This is not necessarily because early
humans lived in caves, but because the bones are so
well-preserved, and often because predators or scaven¬
gers carried the bones there. Cave deposits, though,
have been difficult to date beyond a few hundred
thousand years. Cosmogenic nuclide burial dating has
provided a new tool for dating these old fossils.
One example is found at the caves of Atapuerca,
Spain. These are multilevel caves formed as the nearby
Arlanzon river incised. A railroad cut through the
limestone breached into the caves, and revealed that at
several places the cave had filled with sediment from
above, where shafts connected the passages to the sur¬
face. These sedimentary infills contain an abundance
of fossils, including human ancestors.
The sedimentary infill at Atapuerca known as Sima
del Elefante fills a vertical canyon 18 meters deep.
The lower part of the infill contains some of the oldest
fauna found at Atapuerca, including weasels, beavers,
cave bears, and European jaguar. In 2007, a partial jaw
and teeth from a human ancestor were discovered,
as were several stone tools. Two samples of sediment
were taken from very near the fossil for cosmogenic
dating. The two samples agree closely, and give an
average burial age of 1.18±0.12 million years. This
makes the bone the oldest dated fossil of a human
ancestor in Europe (Carbonell et al., 2008).
A second example comes from Zhoukoudian,
near Beijing, China. This is a vertical cave that was
completely filled with sediment. The site is famous
for the discovery of "Peking Man" in the late 1920s.
Peking Man was among the first discovered human
ancestors, and played an important role in the early
development of the science of paleoanthropology. The
age of Peking Man has been debated for many years,
with most people believing that the fossil is about
500,000 years old. Cosmogenic nuclide burial dating
provides a way to determine the age absolutely.
Shen et al. (2009) collected six samples of quartz
sediment from the cave infill, and also four stone tools
made of quartz. Out of these ten samples, six were
used to date Peking Man. These samples have an
average age of 0.77±0.08 million years, substantially
older than previously thought. They were able to
correlate sedimentation of the cave with global climate
cycles, and together with the cosmogenic burial ages
suggested that Peking Man occupied northern China
during a cold phase, when Beijing was cooler and drier
than today. This has important implications for the
types of environments that early humans were capable
of colonizing.
SUMMARY
These two examples of burial dating with cosmogenic
nuclides show only the beginning of what the dating
technique can do. The sediments in the Mammoth Cave
System were an integral part of how the cave was
formed. The sediments reveal the evolution of the cave
system, and how cave development is tightly coupled to
river incision and aggradation. In this case. Mammoth
Cave was ideal because it was a water-table cave that
carried quartz from local bedrock. In contrast, Atapuerca
and Zhoukoudian are sedimentary infills where sedi¬
ment (and animals) fell into a preexisting cavity. Such
cave infills are the norm in archaeology and paleoanthro¬
pology because they collect bones and artifacts over long
periods of time. In this case, the cosmogenic nuclides
dated the sedimentary infill rather than the cave itself.
There are many more situations where geomorpholo¬
gists, paleoanthropologists, and other scientists can bene¬
fit from dating cave sediments over the past 5 million
years.
See Also the Following Articles
Clastic Sediments in Caves
Multilevel Caves and Landscape Evolution
Mammoth Cave
Bibliography
Carbonell, E., Burmudez de Castro, J. M., Pares, J. M., Perez-
Gonzalez, A., Cuenca-Bescos, G., Olle, A., et al. (2008). The first
hominin species of Europe. Nature, 452(7 186), 465—470.
Granger, D. E., Fabel, D., & Palmer, A. N. (2001). Pliocene-
Pleistocene incision of the Green River, Kentucky, determined
from radioactive decay of cosmogenic 26 Al and 10 Be in Mammoth
Cave sediments. Geological Society of America Bulletin, 113(7),
825-836.
Granger, D. E., & Muzikar, P. F. (2001). Dating sediment burial with
cosmogenic nuclides: Theory, techniques, and limitations. Earth
and Planetary Science Letters, 188(1—2), 269—281.
Palmer, A. N. (1981). A geological guide to Mammoth Cave Natiotial
Park. Teaneck, NJ: Zephyrus Press.
ENCYCLOPEDIA OF CAVES
CRUSTACEA
177
Shen, G. }., Gao, X., Gao, B., & Granger, D. E. (2009). Age of
Zhoukoudian Homo erectus determined with 26 Al/ 10 Be burial
dating. Nature, 458(7 235), 198-200.
CRUSTACEA
Horton H. Hobbs III
Wittenberg University
INTRODUCTION
Crustaceans are one of the oldest and most diverse
arthropods as well as one of the most successful
groups of invertebrates on Earth, with approximately
40,000 extant species described and some 150,000 spe¬
cies recognized. They have been extremely successful
in aquatic habitats, yet some species have become
adapted on land as well. Their fossil record indicates
that they are an ancient group, having occupied marine
environments since the lower Cambrian period yet
very early in their evolutionary history they invaded
freshwater habitats. Although about 90% of the cur¬
rently recognized taxa are widespread in marine sys¬
tems, the remaining 10% are found in diverse inland
waters and assume important roles in various ecosys¬
tem processes of many surface and subterranean lotic
and lentic habitats (Hobbs 2000). This article focuses on
the hypogean members, specifically on those crustaceans
that are highly adapted to dwelling in groundwater eco¬
systems and generally referred to as stygobionts (obligate
hypogean aquatic forms).
In addition to being taxonomically diverse, crusta¬
ceans are anatomically disparate, having evolved an
assortment of body forms accomplished by developing
highly specialized body segments and appendages as
well as by fusing various segments. As a group, crusta¬
ceans are bilateral, having internal and external
segmentation, and an open hemocoel. They have a
rigid, chitinous exoskeleton composed of a thin pro¬
teinaceous epicuticle and a thick multilayer procuticle
that in many groups is hardened by small inclusions
of calcium carbonate. Their bodies are generally
divided into the cephalon (head), thorax, and abdomen
(with the former two sometimes combined as the
cephalothorax). The many jointed appendages are
biramous (or secondarily uniramous), may occur in all
regions of the body, and these arthropods possess
paired antennules (uniramous in all crustaceans except
malacostracans), antennae, mandibles, and maxillae.
Crustaceans have invaded the hypogean realm,
occupying interstitial and other groundwater habitats,
including anchialine waters (inland groundwater
with subsurface marine connections harboring unique
fauna) in karst (see article entitled "Anchialine Caves"
in this volume) as well as in other landforms (e.g.,
lava). Some species of amphipod and isopod crusta¬
ceans have abandoned the groundwater and have been
successful in the terrestrial hypogean environment.
Most species dwelling in subterranean environments
exhibit a suite of characteristic traits that are adaptive
for life in such extreme ecosystems. Examples include
reduction or loss of eyes and pigments, elongation of
appendages, increased chemical and tactile sensitivity,
degeneration of circadian rhythms, lowered fecundity
and metabolic rates, and increased longevity and
ovum volume. These embody behavioral, ecological,
morphological, and physiological modifications that
include both the reduction or loss of characters (regres¬
sive evolution) as well as the augmentation of others
(constructive evolution). These various adaptations
combine to generate the convergence characteristic of
most obligate, cave-adapted organisms that is referred
to as troglomorphy (see Culver and Wilkens 2000).
Crustacean taxonomy continually undergoes reeval¬
uation and revision and the classification structure
used herein (generally based on Martin and Davis,
2001) reveals five classes having subterranean repre¬
sentatives: Branchiopoda, Remipedia, Maxillopoda,
Ostracoda, and Malacostraca (Table 1); a brief discus¬
sion of the classes and their hypogean representatives
(predominantly stygobionts) is presented next.
SYNOPSIS OF STYGOBIOTIC
CRUSTACEAN TAXA
Class Branchiopoda
Branchiopods are relatively small heterogeneous
crustaceans that share few characteristics, including
small to vestigial head appendages with similar mouth-
parts, flattened leaflike thoracic legs called phyllopods
that usually decrease in size posteriorly, and a pair
of spines or claws on the ultimate body segment.
Classification of branchiopods has undergone numerous
revisions and, of the currently recognized orders, only
Diplostraca has subterranean members. There are about
100 species of the suborder Cladocera (450 total species)
that occupy the subterranean environment (Table 1).
They are known from subterranean waters (espe¬
cially the interstitial/hyporheos) on all continents, but
especially well in Bosnia and Herzegovina, France,
Romania, Slovenia, and Spain. These small transparent
crustaceans have a carapace that is laterally compressed
and attached dorsally to the body around which it is
wrapped, excluding the head. Troglomorphic adapta¬
tions are very minor except in a few species of the
genera Alona and Spinalona, evident by a lack of eyes
Encyclopedia of Caves.
© 2012 Elsevier Inc. All rights reserved.
178
c
TABLE 1 Abbreviated Classification of Crustacea, Inclusive Only of Those Groups Dwelling in Subterranean Habitats 3
Subphylum Crustacea (4570+)
Class Branchiopoda (~100)
Subclass Phyllopoda (—100)
Order Diplostraca (—100)
Suborder Cladocera (79)
Family Daphniidae (20)
Family Moinidae (5)
Family Bosminidae (2)
Family Macrothridicae (8)
Family Eurycercidae [ = Chydoridae] (44)
Infraorder Ctenopoda (2)
Family Sididae (2)
Infraorder Haplopoda (1)
Family Leptodoridae (1)
Class Remipedia (24)
Order Nectiopoda (24)
Family Godzilliidae (5)
Family Micropacteridae (1)
Family Speleonectidae (18)
Class Maxillopoda (1108+)
Subclass Tantulocarida (1)
Subclass Mystacocarida (<20)
Subclass Copepoda (1087+)
Order Platycopioida (11)
Order Calanoida (39)
Order Misophrioida (17)
Order Cyclopoida (231)
Order Gelyelloida (2)
Order Harpacticoida (787)
Class Ostracoda (-1000)
Subclass Myodocopa (235)
Order Myodocopida (8)
Order Halocyprida (29 strictly anchihaline and approximately 200 marine interstitial species)
Subclass Podocopa (—365)
Order Platycopida (1)
Order Podocopida (120+, of which approximately half are anchihaline)
Subclass Mystacocarida (<20 intertidal and subtidal interstices)
Class Malacostraca (2338)
Sublcass Phyllocarida (1)
Order Leptostraca (1)
Subclass Eumalacostraca (2337)
ENCYCLOPEDIA OF CAVES
CRUSTACEA
179
Superorder Syncarida (249)
Order Bathynellacea (228)
Order Anaspidacea (21)
Superorder Peracarida (1866)
Order Spelaeogriphacea (4)
Order Thermosbaenacea (33)
Order Mysida (45)
Order Mictacea (1)
Order Bochusacea (3)
Order Amphipoda (784)
Suborder Gammaridea (784)
Order Isopoda (973)
Suborder Phreatoicidea (12)
Suborder Anthuridea (24)
Suborder Microcerberidea (64)
Suborder Flabellifera (144)
Superfamily Cymothoidea (Cirolanidae: 93)
Superfamily Sphaeromatoidea (Sphaeromatidae: 51)
Suborder Asellota (448)
Suborder Caloabozoidea (1)
Suborder Oniscidea (280)
Order Tanaidacea (7)
Order Cumacea (16)
Superorder Eucarida (222)
Order Decapoda (222)
Suborder Pleocyemata (222)
Infraorder Stenopodidea (1)
Infraorder Caridea (111)
Infraorder Astacidea (41)
Infraorder Anomura (2)
Infraorder Brachyura (67)
a Number = approximate number of described subterranean species.
and a carapace that is translucent and sparsely pig¬
mented. Additionally, some have conserved a suite of
primitive characters ( e.g., setation of valve rims), which
suggests that the protected constancy of the hyporheic
has allowed for the survival of some old taxa.
Class Remipedia
The discovery of remipedes in the subterranean
waters of Lucayan Cavern on Grand Bahama Island in
1979 presented a major surprise. On one hand these
blind crustaceans possess characteristics that are very
primitive (e.g., long homonomous body, paddle-like
antennae, double ventral nerve cord), yet they have
attributes that are traditionally considered advanced
(e.g., maxillipeds and biramous limbs that are not
platelike). These small (<1—4 cm) translucent crusta¬
ceans lack a carapace but a cephalic shield covers the
head, appendages of which include a pair of rodlike
processes anterior to antennules and prehensile
ENCYCLOPEDIA OF CAVES
180
mouthparts; the head and first trunk segment comprise
the cephalothorax; and they have an elongate trunk of
up to 32 unfused segments, each bearing a pair of simi¬
lar, laterally directed biramous limbs (Fig. 1).
They have been observed by divers primarily within
submerged caves below the density interface between
the overlying fresh or slightly brackish water and the
underlying dense saltwater, although one species,
Speleonectes epilimnius Yager and Carpenter, is found
above the density gradient. Most swim ventral side
up as the result of synchronized beating of the trunk
appendages and apparently stay below this interface
in the oxygen-poor (as low as 0.08 mg/L) saltwater
layer. These predators likely feed in the overlying,
well-oxygenated freshwater lens and locate their food
by chemosensory means. Remipedes are small (up to
45 mm total length) and are commonly associated with
other stygobiotic crustaceans such as caridean shrimps,
cirolanid isopods, haziid amphipods, mysids, ostra-
cods, and thermosbaenaceans. Twenty-four species
FIGURE f Dorsal view of the remipede, Micropacter yagerae
Koenemann, Iliffe, and van der Ham, from Old Blue Hill Cave,
Caicos Island, Bahamas. Photo courtesy of Stefan Koenemann. Used ivith
permission.
reside in three families containing eight genera
(Tables 1 and 2) and are known only from anchialine
caves in the Cape Range Peninsula in Western
Australia, Bahamas, Canary Islands, Cuba, Turks and
Caicos, and the Yucatan Peninsula (Mexico).
Class Maxillopoda
Maxillopods are mostly small crustaceans although
barnacles (Thecostraca: Cirripedia) are conspicuous
deviations. They have a reduced abdomen and lack
a full complement of legs. Treated below are those sub¬
classes having subterranean representatives: Tantulocarida,
Mystacocarida, and Copepoda.
Subclass Tantulocarida
These small (<0.5 mm long), ectoparasitic crusta¬
ceans are restricted to crustacean hosts. The only sty¬
gobiotic tantulocarid (Stygotantulus stocki Boxshall and
Huys 1989) is known from an anchialine lava pool on
Lanzarote in the Canary Islands where it is parasitic
on two families of harpacticoid copepods.
Subclass Mystacocarida
This small group (most species <0.5 mm in total
length; two genera with <20 described species and
subspecies; Table 1) of marine interstitial crustaceans is
characterized by an elongate body that is divided into
a head and a 10-segmented trunk. The carapace and
compound eyes are lacking; the first trunk segment
has one pair of maxillipeds not fused to the head; and
it has a telson bearing large, pincer-like furca. Because
of the retention of primitive head segmentation, the
lack of fusion of the cephalon and maxillipedal trunk
segment, simplicity of mouth appendages, and absence
of trunk compartmentalization, this is likely one of the
most primitive of all crustaceans. (Others argue that
these features may be related entirely to pedomorpho-
sis and adaptation for interstitial habitats.)
TABLE 2 Occurrence of Remipedia
Family
Genus
Species
Distribution
Godzilliidae
Godziliiognomus
2
Bahamas
Godziliius
1
Turks and Caicos
Pleomothra
2
Bahamas
Micropacteridae
Micropacter
1
Turks and Caicos
Speleonectidae
Cn/ptocorynetes
3
Bahamas, Dominican Republic
Kaloketos
1
Turks and Caicos
Lasionectes
2
Australia, Turks and Caicos
Speleonectes
12
Bahamas, Canary Islands, Cuba, Dominican Republic, Mexico (Yucatan)
ENCYCLOPEDIA OF CAVES
CRUSTACEA
Little is known concerning the life history of this
group. Apparently eggs are laid free and likely they
have up to six naupliar stages. Their small size and
worm-like body are adaptations for interstitial life, using
head appendages to aid in crawling among sand grains
where they glean detritus and microorganisms from the
surfaces of sediment particles (see Zinn 1986). Although
probably more widespread, they demonstrate a patchy
distribution within the littoral and sublittoral sands in
southern and western Africa, Australia, Brazil, Chile,
southern Europe, the Gulf of Mexico, the Mediterranean
Sea, and the east coast of the United States. Species of the
genus Ctenocheilocaris are restricted to the Neotropical
region, whereas Derocheilocaris spp. inhabit the Nearctic,
Palearctic, and Ethiopian regions.
Subclass Copepoda
The subclass Copepoda is a very large and diverse
group of crustaceans (approximately 220 families, 2300
genera, and 14,000 species) and, because they can attain
incredibly high densities, are considered to be the most
abundant metazoans on Earth (more individuals but
fewer species than the insects). Lacking compound
eyes and a carapace, the basic body plan consists of a
head with well-developed mouthparts and antennae; a
six-segmented thorax bearing swimming appendages,
with the first segment fused to the head with maxilli-
peds; and a five-segmented abdomen, lacking appen¬
dages but including a telson. Development occurs
within a few days to three weeks from fertilized eggs
that hatch into a larval stage called a nauplius. Six nau¬
pliar stages are followed by six copepodid stages, the
last of which is the adult (no additional molts). Six of
the 10 orders have subterranean representatives, most
of which demonstrate varying degrees of troglomor-
phy, and are treated next briefly (Table 1).
The Platycopioida order consists of one family
(Platycopiidae), four genera, and 11 species that have
retained numerous primitive characters. They are consid¬
ered to be the first order to diverge from the main lineage
181
of copepods. The eight species of the genus Phtycopia
are known from the benthos in sea and coastal waters
(Africa, Bahamas, northern Europe, Japan, United States).
Sarsicopia polaris Martinez-Arbizu is found in muddy
sediments of the Arctic Ocean, and two anchialine genera
and species are endemic to a single cave in Bermuda:
Antrisocopia prehmsilis Fosshagen and Nanocopia minuta
Fosshagen, one of the smallest known copepods.
The Calanoida order contains approximately 2400
species in about 250 genera, yet only 39 species
assigned to 23 genera and seven families are described
from subterranean waters (Tables 1 and 3). Calanoids
have biramous antennae and antennules that are
greatly elongated, and the point of major body articu¬
lation occurs between the thorax and abdomen, which
is marked by a distinct narrowing of the body. This
primarily filter-feeding, planktonic group is geographi¬
cally widespread, having been found in hypogean set¬
tings in Australia, Bahamas, Balearic Islands, Barbuda,
Belize, Bermuda, Canary Islands, Caroline Islands,
southeast China, Cuba, Dinaric Alps, Fiji, southern
France, Galapagos Islands, Herzegovina, Istria, Italy,
Madagascar, Mediterranean, Mexico, Philippines, and
Russia.
The Misophrioida order is represented by 16 genera
and 34 widely distributed copepods that are plank¬
tonic, hyperbenthic, and deep-sea, open-ocean, and
hypogean water-dwellers. The subterranean species
demonstrate particularly disjunct distributions, which
are summarized in Table 4.
Members of the Cyclopoida order are free-living,
planktonic, and associated with various substrates in
benthic or littoral habitats. They detect their prey with
the aid of mechanoreceptors on their first antennae
and grasp their food with precision with their first
maxillae or, occasionally, with their second maxillae
and maxillipeds. Most are voracious predators although
some are parasitic. Antennules and uniramous anten¬
nae are moderately long (never as long as antennules
of harpacticoids; see below); a fifth pair of legs is
TABLE 3 Occurrence of Described Stygobiotic Calanoid Copepods in Anchihaline, Cave, or Well Habitats
Family
Genus
Species/ Subspecies
Occurrence
Boholiniidae
2
2
Anchihaline
Diaptomidae
8
11
Freshwater
Epacteriscidae
3
14
Anchihaline; freshwater
Fosshageniidae
1
1
Anchihaline
Pseudocyclopiidae
3
4
Anchihaline; marine
Ridgewayiidae
5
5
Anchihaline
Stephidae
1
2
Anchihaline
Total
23
39
ENCYCLOPEDIA OF CAVES
182 c
TABLE 4 Subterranean Misophrioida Copepoda Distributions and Habitats
Genus
Species
Distribution
Habitat
Boxshallia
bulboantennulata
Canary Islands
Anchihaline
Dimisophria
cavernicola
Canary Islands
Anchihaline
Expansophria
4
Canary Islands; Galapagos Islands; Italy; Sardinia; Palau
Anchihaline
Huysia
bahamensis
Bahamas
Anchihaline
Misophria
kororiensis
Atlantic
Anchihaline
Palpophria
aestheta
Canary Islands
Anchihaline
Protospeleophria
lucayae
Bahamas
Anchihaline
Speleophria
4
Angaur Islands, Palau; Balearic Islands; Bermuda
Anchihaline, cave
Speleophriopsis
2
Balearic Islands; Canary Islands
Anchihaline
Stygomisophria
kororiensis
Koror Island, Palau
Anchihaline
highly reduced, the sixth pair vestigial. Cyclopoid
copepods have not been well studied in subterranean
environments. Nearly 750 species are recognized, but
only about 231 species and subspecies are reported
from freshwater caves, interstitial habitats, and wells
(Table 1) (e.g., Acanthocyclops, Diacyclops, Eucyclops,
Halicyclops, Idiocyclops, Kieferiella, Metacyclops, Neocyclops,
Speleoithona, and Speocyclops) from Africa, Asia, Australia,
Bahamas, Cuba, Europe, Madagascar, and North and
South America. A diverse and abundant copepod fauna
is actively being studied inhabiting epikarstic drip
pools (Pipan 2005). About 60 additional species and
subspecies are known from marine and brackish water
interstices.
The Gelyelloida order is represented by a single
genus and two species inhabiting European freshwater
karst systems. Gelyella droguei Rouche and Lescher-
Moutoue and G. monardi Moeschler and Rouch are
restricted to karst waters in Montpellier, France, and in
the Swiss Jura, respectively. This copepod order is
characterized by a distinct combination of gnathosto-
mous mouthparts and unusual derived features.
The Harpacticoida order is characterized by a body
that is generally worm-like and cylindrical, with anterior
segments not much larger than posterior ones; anten-
nules and biramous antennae are quite short and the
point of major body articulation occurs between the
fifth and sixth thoracic segments; locomotory pereiopods
are reduced, consistent with loss of swimming ability
and use of these appendages as levers against sand
grains. This group of copepods occurs virtually in all
aquatic environments; most are benthic and are well
suited to move through interstices feeding on detritus
and on microorganisms (e.g., bacteria, diatoms, and pro¬
tozoa). The harpacticoid copepods comprise 42 families
with approximately 375 genera containing about 3000
species. At least 12 families (about 445 species) are
FIGURE 2 A candonid ostracod (subclass Podocopa). Photo
courtesy of Stuart Halse. Used with permission.
known from freshwater caves, interstitial groundwaters,
and wells (e.g., Chappuisius, Elaphoidella, Neochinophora,
Parastenocaris, Spelaeocamptus, Stygonitocrella). Some 342
species and subspecies in 63 genera and 13 families are
found in the marine interstitial habitat (e.g., Leptocaris,
Nitocra, Novocrinia, Psammotopa, Stygolaophonte).
Class Ostracoda
Ostracods are small (usually <1 mm, rarely 2 cm)
crustaceans, with short, oval bodies encased within
bivalved shells hinged dorsally (Fig. 2). This is a very
diverse group (5700 species) and although most are
found in the marine environment (to depths of 7000 m)
they are abundant worldwide in all aquatic systems.
Most species are benthic, many are planktonic, some
are commensal on echinoderms or various malacostra-
can Crustacea (e.g., Sphaeromicolinae on Isopoda,
Entocytherinae on cambarid Decapoda), and a few are
terrestrial in moist habitats.
ENCYCLOPEDIA OF CAVES
CRUSTACEA
183
Although some species are known only from caves,
most hypogean ostracods are found in anchialine habi¬
tats, interstitial waters, springs, and wells. Examples of
ostracods occupying various subterranean waters are
Cavernocypris lindbergi Hartman (Afghanistan cave),
Pseudocandona jeanneli Klie (cave in Indiana, United
States), Mixtacandona juberthieae Danielopol (cave in
southern France), and Pseudolimnocythere hartmanni
Danielopol (well in Greece). These stygobionts display
various troglomorphic adaptations including reduction
or loss of eye structure, pigments, number of setae,
and some have a very elongate or trapezoidal-shaped
carapace. Most of the approximately 1000 subterranean
species are assigned to the order Podocopida (Fig. 2)
although the order Myodocopida is dominated by ostra¬
cods living in marine interstitial habitats (Bermuda,
Galapagos Islands, Jamaica) and a coastal sea cave on
Niue in the central Pacific ( Dantya ferox Kornicker and
Iliffe) (Table 1). The order Halocyprida is represented
by 29 troglomorphic species dwelling in anchialine
habitats, blue holes, and caves: DanielopoUna (11 species
in the Bahamas, Canary Islands, Cuba, Galapagos
Islands, Jamaica, Yucatan Peninsula of Mexico, and the
Cape Range Peninsula of Western Australia), Deeveya
(7 species in the Bahamas and the Turks and Caicos
islands), Euconchoecia (single species in Palau), and
Spelaeoecia (11 anchialine species in the Bahamas,
Bermuda, Cuba, Jamaica, and the Yucatan Peninsula);
approximately 200 are marine interstitial species.
Of interest, the wide, irregular distribution, primitive
nature, and troglomorphic adaptations of these taxa
point to an extended history in a suitable cave environ¬
ment. However, during the most recent period of
Pleistocene glaciation, the sea level was lowered at least
100 m, resulting in coastal, anchialine caves becoming
dry and then reinundated (substantiated today by the
presence in these submerged caves of stalactites and
stalagmites and other speleothems that were formed
only in air by dripping or flowing water). This suggests
that present-day amchialine ostracod fauna are recent
invaders (within the past 15,000—18,000 years) and that
they likely used an alternate, deeper habitat as refuge
for considerable periods of time.
Class Malacostraca
This diverse group (about 29,000 species) far exceeds
the species richness of any other crustacean and is
divided into three subclasses, of which two have species
dwelling in hypogean environments: Phyllocarida and
Eumalacostraca. The body fundamentally comprises a
five-segmented cephalon, eight-segmented thorax, six-
segmented abdomen (seven in leptostracans), and telson;
the carapace may be absent, reduced, or may cover part
or all of the thorax and even several abdominal
segments; many have none to three pairs of maxillipeds;
antennules and antennae are usually biramous; the
abdomen generally bears five pairs of biramous pleo-
pods and one pair of biramous uropods; and they are
mostly gonochoristic.
Subclass Phyllocarida
The subclass Phyllocarida is represented by the
single order Leptostraca, which is the most primitive
member of Malacostracans and is characterized by a
head with movable, articulated rostrum; biramous
antennules and uniramous antennae; absent maxilli¬
peds; phyllopodous thoracopods; a laterally com¬
pressed bivalved carapace (lacking hinge) covering the
thorax; and an elongate abdomen consisting of seven
free pleomeres plus telson. The order is represented
by fewer than 20 species assigned to six genera, most
of which are small (5—15 mm long), but one species
is nearly 4 cm in length. Most are epibenthic, are suspen¬
sion feeders, and occur in low-oxygen marine environ¬
ments. A single stygobiont, Speonebalia cannoni Bowman,
Yager, and Iliffe (1985) (Nebaliidae), is known from two
caves on Providenciales, Caicos Islands (Fig. 3).
Subclass Eumalacostraca
Members of this subclass have head, thorax, and
abdomen; up to three thoracomeres are fused with the
head, appendages of which are usually modified as
maxillipeds; most groups have a well-developed cara¬
pace; and they possess a telson and paired uropods.
There are three of four superorders with subterranean
representatives.
SUPERORDER SYNC ARID A
This freshwater group, derived from marine stock,
demonstrates the most primitive living body
FIGURE 3 The stygobiotic Ieptostracan, Speonebalia cannoni
Bowman, Yager, and Iliffe from Providenciales, Caicos Islands. Photo
courtesy of Tom Iliffe. Used with permission.
ENCYCLOPEDIA OF CAVES
184
c
FIGURE 4 The parabathynellid syncarid, Hexabathynella spp.
Photo courtesy of Jane McRae. Used with permission.
architecture of any eumalacostracan and many of the
rather uniform trunk segments lack appendages. They
do not have a carapace; telson with or without furcal
lobes; some pereiopods are biramous; and the pleopods
are variable. They either crawl or swim and, although
little is known about most species, some are likely
omnivorous. Unlike most other crustaceans, which
transport eggs and thus carry early embryos, syncarids
lay their eggs or release them into the water subse¬
quent to copulation. Approximately 260 species have
been described that are placed into two orders
(Table 1): Bathynellacea and Anaspidacea; 95% of these
are stygobionts.
The primitive Bathynellacea order lacks maxilli-
peds, is by far the more diverse group of syncarids,
and is worldwide in distribution. Approximately
228 species are assigned to 60 genera that are placed
within three families: Bathynellidae (e.g., Bathynella
chappuisi Delachaux, Switzerland; Bathynella primaus-
traliensis Schminke, Australia), Leptobathnelidae
(e.g., Acanthobathynella knoepffleri Coineau, Ivory Coast,
Africa), and Parabathynellidae (e.g., Hexabathynella spp.,
Africa, Australia, New Zealand, North and South
America (Fig. 4); Notobathynella ivilliamsi Schminke,
Australia; Parabathynella stygia Chappuis, Croatia,
Slovakia). They typically inhabit freshwater interstitial
media in epigean, cave, and well hypogean habitats in
Africa, Asia, Australia, Europe, Japan, Madagascar,
Malaysia, New Zealand, and North and South America.
Order Anaspidacea (five families) has one pair of
maxillipeds, is diverse in Australia, and species richness
is particularly high on the island of Tasmania (numerous
species awaiting formal descriptions). The Anaspididae
family (e.g., Anaspides tasmaniae (Thomson) (Fig. 5))
inhabits various freshwater environments in Tasmania,
including caves. However, most other syncarids are
interstitial dwellers or live strictly in subsurface ground-
water, including caves and springs (e.g., the exclusively
stygobiotic family Psammaspididae— Psammaspides spp.—
particularly diverse in New South Wales, Australia, and
Eucrenomspides spp. in Tasmania). Psammaspidae is repre¬
sented only by Eucrenonaspides oinotheke, a spring-dwelling
syncarid endemic to Tasmania. Family Koonungidae is
FIGURE 5 The anaspidacean syncarid, Anaspides tasmaniae
Thomson, from Tasmania. Photo courtesy of Dr. Stefan Eberhard
(www.subterraneanecology.com.au). Used with permission.
found in sediment interstices, open water, and sinkholes
(e.g., Koonunga crenarum Zeidler in southeast Australia and
Tasmania). Family Stygocarididae (historically considered
a separate order) is represented by fewer than 10 species
assigned to four genera: Oncostygocaris, Parastygocaris,
Stygocarella, and Stygocaris. These tiny species inhabit inter¬
stitial waters in epigean and hypogean environments in
New Zealand, Tasmania, and South America.
SUPERORDER PERACARIDA
Crustaceans placed into this superorder demon¬
strate a trend toward reduction of the carapace. They
possess one (rarely two to three) maxillipeds; gills are
thoracic or abdominal; unique thoracic coxal endites
(oostegites) form a ventral brood pouch or marsupium
in females; they lack true larval stages; the young
hatch as mancas, a prejuvenile stage lacking the last
pair of thoracopods; pleopods lack appendix interna;
and the telson is without caudal rami.
This highly successful group of malacostracans
(approximately 11,000 species) is divided into 10 orders;
all but the Lophogastrida have known subterranean
species (Table 1). Most are marine but many occupy
freshwater and terrestrial habitats. Peracarids are
diverse in their habits and size, ranging from a few
millimeters to 44 cm in length and some are symbionts
as well as stygobionts.
The Spelaeogriphacea order is limited to fresh ground-
water habitats in South Africa, South America, and
Western Australia (Gondwana plates). This group was
initially represented by a single species, Spelaegriphus
lepidops Gordon, which was described from pools and
a stream in Bats Cave, Table Mountain (now known
from a second cave). South Africa, where it was observed
swimming swiftly using rapid undulations of the body.
Evidence suggested that it feeds largely on detritus.
ENCYCLOPEDIA OF CAVES
CRUSTACEA
In 1987 and again in 1998 two genera were described,
one from a lake in a freshwater cave in Brazil (Potiicoara
brasiliensis Pires—now known from two other caves) and
the other from the Millstream aquifer in arid northwest¬
ern Australia (Mangkurtu mityula Poore and Humphreys;
a second species, M. kutjarra Poore and Humphreys, has
been added to this genus (Fig. 6) (Poore and Humphreys
1998, Humphreys and Harvey 2001).
These small, blind, unpigmented stygobionts pos¬
sess a short, saddle-like carapace fused with the first
thoracomere and, anteriorly, produced into a broadly
triangular rostrum; they have one pair of maxillipeds;
pereiopods 1 through 7 are simple, biramous, with
shortened exopods; the exopods on pereiopods 1, 2,
and 3 are modified for producing currents, and on per¬
eiopods 4 through 7 they are modified as gills; the
abdomen is elongated, often exceeding half of the total
body length; and pleopods 1 through 4 are biramous
and natatory, and pleopod 5 is reduced.
Order Thermosbaenacea is a group of small
(2—5 mm), aquatic crustaceans that have a short cara¬
pace fused with the first thoracic somite (remaining
seven thoracic segments free) extending posteriorly
over two to three additional segments. In females, the
carapace provides a brood pouch; there is a single pair
of maxillipeds; pereiopods are biramous, simple, and
lack epipods; there are two pairs of uniramous pleo¬
pods; the uropods are biramous; and the telson is free
or forms a pleotelson with the last pleonite.
At least 33 stygobiotic species have been assigned to
seven genera and four families (one additional species
FIGURE 6 The spelaeogriphacean, Mangkurtu katjarra Poore and
Humphreys, from groundwater in limestone in the upper Fortescue
River in Western Australia. Photo courtesy of Jane McRae. Used with
permission.
185
of Limnosbaena not formally described) (Table 5; see
Wagner, 1994, for a revision of the order). They are
known from anchialine habitats, caves, cenotes, vari¬
ous interstices, cold springs, thermal springs (45° C),
and wells in fresh, to oligohaline, to hypersaline waters.
These tethyan relicts demonstrate a very large geographic
range in those areas once covered by the shallow Tethys
Sea or along its former coastlines.
Order Mysida is represented by more than 1000 spe¬
cies that are widespread over all continents where they
inhabit coastal and open ocean waters as well as conti¬
nental fresh waters and various groundwater habitats.
Some species are intertidal and burrow into the sand
during periods of low tides; most of these shrimp-like
crustaceans swim with the aid of thoracic exopods and
are omnivorous suspension feeders that eat algae,
zooplankton, and suspended detritus. They range in
length from approximately 2 mm to 8 cm and display a
well-developed carapace covering most of the thorax.
Compound eyes are stalked, sometimes reduced; they
have one to two pairs of maxillipeds not associated with
cephalic appendages; the abdomen is elongated; pereio¬
pods are biramous although the last pair is sometimes
reduced; pleopods are reduced or modified; and the
statocyst is usually located in each uropodal endopod.
Of the two suborders, only the Mysida has stygo¬
biotic representatives and is divided into four fami¬
lies, three of which have stygobiotic and stygophilic
species (Table 6). At least 45 stygobiotic species are
recognized, most of which are endemics. The current
distribution of the majority of these suggests that
most colonized ground waters as a consequence of
uplifting and stranding of their marine ancestors,
which resulted from regressions of the Tethys and
Mediterranean seas. Clearly other taxa have invaded
ground waters more recently, most being stygophiles.
Adaptations for subterranean existence seem to be
limited to reduction or loss of body pigments (red pig¬
ments often retained) as well as reduction or loss of
eyes but not eyes talks that contain endocrine organs.
Mictaceans are stygobiotic crustaceans, lacking body
pigmentation and without visual elements in reduced
eyestalks. The head narrows anteriorly into a triangular
TABLE 5 Occurrence of Subterranean Thermosbaenaceans
Family
Genus
Species
Occurrence
Halosbaenidae
3
5
Caves, marine interstitial
Monodellidae
2
24
Anchihaline, artesian wells, brackish wells, caves, coastal springs, interstitial, phreatic, thermal springs,
wells
Thermosbaenidae
1
1
Thermal springs
Tulumellidae
1
3
Anchihaline, caves, cenotes
Total
7
33
ENCYCLOPEDIA OF CAVES
186 c
TABLE 6 Occurrence of Described Subterranean Mysids (Order Mysida)
Family
Genus
Species
Occurrence
Lepidomysidae
2
10
Anchihaline, caves, crab burrows, interstitial, phreatic, prawn culture field
Mysidae
16
28
Anchihaline, caves, coral reef caves, land crab burrows, marine caves
Stygiomysidae
1
7
Anchihaline, caves, phreatic
Total
19
45
rostrum and is fused posteriorly with the first thoraco-
mere. The carapace is not developed but small lateral
carapace folds produce a head shield laterally covering
bases of mouthparts; gills are lacking; pereiopods are
simple; pleopods are reduced and uniramous; and uro-
pods are biramous, with two to five segmented rami.
A recent revision of the order Mictacea has resulted
in its being monotypic (Table 1). The single species,
Mictocaris halope Bowman and Iliffe, assigned to the
family Mictocarididae, is known from four marine
caves on Bermuda (Fig. 7). This small species (up to
3.5 mm in total length) swims, using exopods of thora-
copods 2 through 6, and rarely rests or walks on the
substrate. Mouthparts suggest that it is not a predator
but probably procures food by scraping and/or filtering
fine particulates from the water column and benthos.
Order Bochusacea is the most recently established
peracarid order, which was erected to accommodate
three species placed in two genera belonging to the
family Hirsutiidae (initially placed in the order
Mictacea; two deep-sea benthic species: Hirsutia bathylis
Sanders, Hessler, and Garner off Surinam in northeast
South America, and Hirsutia sandersetalia Just and Poore
from Bass Strait, Australia). With additional material it
became clear that a new order, Bochusacea, should be
erected to house the hirsutiids (Gutu and Iliffe 1998).
Additionally, a third species, Thetispelecaris remex Gutu
and Iliffe, from three marine/anchialine caves in the
Bahamas, was placed in the order. Although similar
to the mictaceans, the bochusaceans possess different
pereiopod forms, among other features. They too are
swimmers and filter feeders; pereiopod 1 is specialized
for feeding and not for locomotion.
The Amphipoda order is characterized by the
absence of a carapace; the body is divided into head,
thorax, and abdomen, each bearing appendages; the
head bears two pairs of antennae; it has a single pair
of maxillipeds, seven pairs of uniramous pereiopods,
with the first, second, and often others modified as
chelae or subchelae; periopodal coxae are expanded as
lateral side plates; gills are thoracic; the abdomen con¬
sists of two regions of three segments each, an anterior
pleon with pleopods, and posterior urosome with
appendages modified as uropods; and the telson is
FIGURE 7 The stygobiotic mictacean, Mictocaris halope Bowman
and Iliffe, from a cave in Bermuda. Photo courtesy of Tom Iliffe. Used
with permission.
free or fused with the last urosomite. This group of
crustaceans is generally slender and laterally com¬
pressed although some are dorsoventrally flattened
(e.g., Heterophlias). They range in length from 1 mm to
giant deep ocean benthic forms reaching 25 cm.
Amphipod crustaceans are commonly found in
numerous aquatic ecosystems around the globe
where they have invaded freshwater, brackish, and
marine environments, often comprising a large por¬
tion of the biomass in a habitat. A few species also
dwell in various terrestrial ecosystems (Talitridae;
e.g., supralittoral sandy beaches, moist forest litter)
and they exhibit a great diversity of feeding strate¬
gies including carnivory, herbivory, parasitism, scav¬
enging, and suspension feeding. Approximately 7000
species are assigned to three suborders: Gammaridea,
Caprellidea, and Hyperiidea. Stygobiont amphipods
are known only from the very large (>5700 species)
Gammaridea suborder where some 784 troglo-
morphic species are assigned to approximately 159
genera and 29 families (Table 7). The following fami¬
lies demonstrate the richest subterranean biodiversity:
Bogidiellidae, Crangonyctidae, Hadziidae, Melitidae,
and Niphargidae. The amphipod suborder is
ENCYCLOPEDIA OF CAVES
CRUSTACEA 187
TABLE 7 Occurrence of Subterranean Amphipods (Suborder Gammaridae)
Family
Genus
Species
Occurrence
Allocrangonyctidae
1
2
Caves: United States (Missouri, Oklahoma)
Aoridae
1
1
Marine interstitial: lie de Batz (Atlantic Ocean)
Bogidiellidae (including
Artesiidae)
33
105
Caves, hyporheos, interstices of sandy beaches and sublittoral sands, cold mountain springs,
wells: nearly worldwide (not from continental Africa south of equator, or continental Australia)
Calliopiidae
3
3
Groundwater, springs: New Zealand (North and South islands)
Crangonyctidae
6
158
Caves, springs, wells: Holarctic region (mostly North America)
Gammaridae
17
38
Caves, springs, wells: Canary Islands, Eurasia, Mediterranean region. North America, Solomon
Islands
Hadziidae
26
78
Caves: Australia, southern Europe, Bahamas, Caribbean, Fiji, Hawaiian Islands, Mexico, United
States
Hyalellidae
2
2
Caves: Australia, Brazil, Venezuela
Hyalidae
1
3
Subterranean freshwater: Comores and Zanzibar islands (off coast of Eastern Africa)
Ingofiellidae
3
37
Caves, interstitial (marine), springs, wells: Africa, Caribbean, Mediterranean regions. South
America
Liljeborgiidae
1
1
Marine cave: Bermuda
Lysiannasidae
1
1
Anchihaline cave: Galapagos Islands
Melitidae
23
54
Anchihaline, caves, freshwater and marine interstitial, and wells: Australia, Balearic, Canary,
Caribbean, Galapagos, Hawaiian, Pacific, and Philippine islands
Metacreangonyctidae
2
15
Hyporheos, springs, wells: Canary Islands, Mediterranean region
Metaingolfiellidae
1
1
Well: southern Italy
Neoniphargidae
2
2
Shallow groundwaters: Victoria and Tasmania (?), Australia
Niphargidae
8
221
Caves, hyporheos, interstitial, wells: Asia, Europe
Paracrangonyctidae
2
3
Interstices, wells: Chile, Kerguelen Island (in Indian Ocean), New Zealand
Paramelitidae
9
10
Caves, groundwater habitats: South Africa, Western Australia, southern Victoria, and Tasmania
Pardaliscidae
2
3
Anchihaline caves: Bahamas, Galapagos, Lanzarote, Turks and Caicos
Perthiidae
1
1
Caves: southwestern Australia
Phreatogammaridae
2
2
Caves, wells: south Australia, New Zealand (South Island), Spain
Pontogeneiidae
1
2
Cave: Japan
Plustidae
2
3
Freshwater caves and wells: Japan
Pseudocrangonyctidae
2
11
Caves, springs, wells: northeastern Asia
Salentinellidae
2
14
Freshwater and brackish water caves, hyporheos, wells: Mediterranean region of northern Africa
and southern Europe
Sebidae
1
2
Artesian well: Texas (Hays County)
Sternophysingidae
1
8
Caves, springs: southern Africa
Talitridae
3
3
Caves: Corsica, Isla de la Palma (Canary Islands), Kauai (Hawaiian Islands), Sardinia
Total
159
784
widespread globally with the greatest taxonomic diver¬
sity of troglomorphic species occurring in eastern and
southern North America, the Caribbean/West Indian
region, central and southern Europe, and the
Mediterranean region. These stygobionts evolved from
surface ancestors, moving into hypogean ground waters
from both freshwater and marine environments involv¬
ing both active and passive dispersal.
Order Isopoda is quite a diverse group with more
than 11,000 described aquatic and terrestrial species.
ENCYCLOPEDIA OF CAVES
188
c
They inhabit nearly all environments where most are
free living yet some are partly (Flabellifera) and others
are exclusively (Epicaridea) parasitic. Their feeding
habits extend from herbivores to omnivorous scaven¬
gers, detritivores, predators, and parasites. They range
in length from about 0.5 mm to 4.4 cm (the largest is
the benthic Bathynomus). This diverse group of crusta¬
ceans is dorsoventrally flattened; they lack a carapace;
the first thoracomere is fused with the head; they
have a single pair of maxillipeds and seven pairs of
uniramous pereiopods which are modified as ambula¬
tory, natatory, or prehensile limbs; pleopods are
biramous and modified as natatory or for gas exchange
(gills in aquatic taxa, pseudotracheae in terrestrial
Oniscidea); and the telson is usually fused with pleo-
nites 1 through 6.
Representatives of subterranean isopods are known
from seven suborders: Phreatoicidea, Anthuridea,
Microcerberidea, Flabellifera, Asellota, Calabozoidea,
and Oniscidea (Tables 1, 8) and are global in
distribution.
Suborder Phreatoicidea is the most ancient group of
isopods and is derived from marine ancestors. These
isopods are currently restricted mostly to freshwaters
of South Africa, Australia, India, and New Zealand.
Most are epigean; however, stygobiotic species (at least
12 species in nine genera), known from caves and
wells, are extant on all of these continental Gondwana
fragments except South Africa.
Anthuridea is principally a marine group consisting
of four families; however, at least 24 troglomorphic
species assigned to two families and three genera dwell
in the interstitial sediments of anchialine habitats,
bays, beaches, caves, and wells in the Canary Islands,
the Caribbean and Indian Ocean islands, Indonesia,
Mexico, New Zealand, and South America.
The Microcerberidea consists of two families.
The Atlantasellidae is represented by a single species,
Atlantasellus cavernicolous Sket, which is restricted to
anchialine caves in Bermuda. Clearly its recent origin
is from marine stock. The Microcerberidae are slender,
stygobionts of which approximately 64 species in six
genera inhabit interstices of caves, marine beaches,
and wells in the coastal regions of southern and
western Africa, Southeast Asia, the Caribbean, Indian
Ocean islands, Japan, the Mediterranean, and western
North America.
Suborder Flabellifera is represented by 18 fami¬
lies, only 2 of which have subterranean members:
Cirolanidae and Sphaeromatidae. The predominantly
marine family Cirolanidae has species dwelling in sub¬
terranean environments ranging from fresh water to
salinities near that of seawater. Approximately 93 spe¬
cies assigned to 29 genera are currently recognized
and range from less than 3 mm in length ( Arubolana
parvioculata Notenboom) to around 33 mm ( Speocirolana
bolivari (Rioja)). Most are blind, lack body pigments,
are somewhat convex dorsally, and are opportunistic
benthic scavengers and predators. Their geographical
distribution includes East Africa, the western Atlantic,
and North America. (Cave and phreatic habitats are
occupied by Speocirolana in Mexico and Texas, and
by Antrolana and Cirolanides in Virginia and West
Virginia, and Texas, respectively, and by Kagalana and
Haptolana in Western Australia.)
The Sphaeromatidae is mostly a marine group with
four genera of stygobionts represented by 43 species
inhabiting subterranean waters in southern Europe and
8 species in hot springs in the southwestern United
States and north-central Mexico. The very diverse
genus Monolistra (at least 35 species) occupies karst
waters along the Dinaride and Italo-Dinaride systems.
The Asellota are divided into a number of super¬
families (not treated herein) and nine families have
subterranean representatives. The family Asellidae is a
large benthic group (about 18 genera and 266 species)
that is found primarily in caves, springs, and wells in
northwestern Africa, Europe, Japan, and Central and
North America. The Stenasellidae represent an ancient
group that demonstrates primitive characters and that
are blind and often orange or pink in body color.
At least 71 species are assigned to 10 genera and
these crustaceans inhabit caves, the hyporheos, cool
and thermal springs, and wells in much of Africa,
Southeast Asia, eastern, southern, and western Europe,
Indonesia, Malaysia, Mexico, and the Edwards Plateau
in North America. The small Stenetriidae isopods are
known only from mixohaline waters in a cave on
Curasao and from a blue hole on Andros Island,
Bahamas. The Janiridae is represented by 13 species
assigned to five genera of marine origin. They are
known from caves, springs, and wells in east Asia,
western Europe, Italy, Japan, and North America
(California). Microparasellidae are small, quite diverse
(79 species in four genera), and widespread. They are
known from eastern Asia, Australia, the Caribbean,
Indian Ocean islands, Japan, and the Mediterranean.
The Paramunnidae is represented by a single species
inhabiting subterranean waters. Although the genus
Munnogonium currently contains seven species, only
M. somersensis Kensley inhabits a marine cave in
Bermuda and shows no obvious troglomorphic adap¬
tations. The gnathostenetroidid isopods are small crus¬
taceans of marine origins and the seven species
distributed evenly over three genera demonstrate
a disjunct distribution in anchialine caves in the
Turks and Caicos islands and Bermuda (the eyed
Stenobermuda), an anchialine cave in the Bahamas
(Neostenetroides ), and the intertidal zone in Japan and a
thermal spring on Ischia, a small coastal island
ENCYCLOPEDIA OF CAVES
CRUSTACEA
TABLE 8 Taxonomic Treatment of Subterranean Isopods
189
Suborder
Family
Genus
Species/Subspecies
Phreatoicidea
Amphisopidae
3
5
Amphisopodidae
1?
1?
Nichollsidae
1
2
Phreatoicidae
3
4
Total
4
8
12
Anthuridea
Anthuridae
1
20
Paranthuridae
2
4
Total
2
3
24
Microcerberidea
Atlantasellidae
1
1
Microcerberidae
6
63
Total
2
7
64
Flabellifera
Cirolanidae
29
93
Sphaeromatidae
4
51
Total
2
33
144
Asellota
Asellidae
18
266
Stenasellidae
10
71
Stenetriidae
1
2
Janiridae
5
13
Microparasellidae
4
79
Paramunnidae
1
1
Gnathostenetroididae
3
7
Protojaniridae
4
9
Total
8
46
448
Calabozoidea
Calabozoidae
1
1
Total
1
1
1
Oniscidea
Armadillidae
-
-
Armadillidiidae
-
-
Berytoniscidae
-
-
Cylisticidae
-
-
Eubelidae
-
-
Ligiidae
-
-
Mesoniscidae
-
-
Oniscidae
-
-
Philosciidae
-
-
Plathyarthridae
-
-
Porcellionidae
-
-
Scleropactidae
-
-
Scyphacidae
-
-
(Continued)
ENCYCLOPEDIA OF CAVES
190
TABLE 8 ( Continued )
C
Suborder
Family
Genus
Species/Subspecies
Spelaeoniscidae
-
-
Styloniscidae
-
-
Trachelipidae
-
-
Trichoniscidae
54
-
Total
17
99
280
GRAND TOTAL
35
197
973
TABLE 9 Occurrence of Tanaidaceans in Subterranean
Habitats
Family
Genus
Species
Occurrence
Apseudidae
Apseudes
bermudeus
Anchihaline: Bermuda
bowmani
Marine: Koror Island (Palau)
orghidani
Anchihaline: Bermuda
Pugiodactylus
agartthus
Marine cave: Niue Island (South Pacific)
Metapseudidae
Calozodion
propinquus
Anchihaline, open ocean: Bermuda
Nototanaidae
Nesotanciis
maclaughlime
Anchihaline: Eli Malk Island (Palau)
Parapseudidae
Swireapseudes
birdi
Anchihaline: Bahamas
southwest of Naples, Italy (Caecostenetroid.es). Nine spe¬
cies belonging to four genera are assigned to the
Protojaniridae family and are quite separate from other
members of the Asellota suborder. They are found in
subterranean freshwaters in South Africa, Argentina
(South America), southern India, islands of the Indian
Ocean, and Namibia.
The monotypic suborder Calabozoidea (Calabozoidae
(Calabozoa pellucida Van Lieshout)) is known only from
wells in the environs of the small town of Calabozo
in northern Venezuela. This transparent, blind isopod
appears to be restricted to phreatic waters.
The suborder Oniscidea is generally a terrestrial
group (e.g., Sinoniscus cavernicolous Schultz, China)
characterized by extreme reduction of the first pair of
antennae and by biramous pleopods modified into
pseudotracheae. Most members of the suborder are
troglobionts yet some are stygobionts; some 276
aquatic and terrestrial species assigned to 99 genera
and 17 families are recognized. Examples of aquatic
species are Abebaioscia troglodytes Vandel from the
Nullarbor Plain, Australia; Haloniscus searlei Chilton, a
scyphacid from Australia in saline lakes up to 159
parts per thousand; a scyphacid isopod, Haloniscus sp.
from Australia; the styloniscid Thailandoniscus annae
Dalens, from Thailand; the trichoniscids Mexiconiscus
laevis (Rioja), which is terrestrial as a juvenile and
aquatic as an adult, and Titanethes albus (Koch), an
amphibious species from Italy and Slovenia.
The order Tanaidacea is known worldwide from
benthic marine habitats and only a few species live
in brackish or nearly fresh water. They often dwell in
burrows or tubes and are known from all ocean
depths. Approximately 850 species are recognized and,
being small (0.5—2 cm long), many are suspension fee¬
ders, some detritivores, and others are predators. The
carapace is fused with the first two thoracic segments;
the first and second thoracopods are maxillipeds,
with the second one being chelate; and thoracopods
3 through 7 are simple, ambulatory pereiopods.
Very little is known about subterranean species,
mostly due to very minimal sampling. Currently
seven species, assigned to the families Apseudidae,
Netapseudidae, Nototanaidae, and Parapseudidae, are
known from anchialine/marine caves in the Bahamas,
Bermuda, Eli Malk and Koror islands (Palau) (Gutu and
Iliffe 1989), and Niue Island (South Pacific) (Table 9).
The order Cumacea is represented by rather
strange-looking crustaceans having a large, bulbous
anterior end and elongated posterior, a carapace that is
fused to and covers the first thoracic segments, and
three pairs of maxillipeds. Pereiopods 1 through 5 are
simple and ambulatory, and pleopods are present in
males, but usually absent in females.
ENCYCLOPEDIA OF CAVES
TABLE 10
Subterranean Cumaceans
CRUSTACEA
191
Family
Genus
Species
Occurrence
Bodotriidae
Cyclaspis
2
Submarine caves: Bahamas, Jamaica
Diastylidae
Oxyurostylis
antipai
Submarine cave: Jamaica
Nannastacidae
Campylaspis
2
Blue hole: Andros Island, Bahamas; submarine cave: Bermuda
Cumella
9
Anchihaline: Bermuda; blue hole: Andros Island, Bahamas; submarine cave: Jamaica
Schizotrema
2
Anchihaline, caves: Bermuda
Total
5
16
They are distributed worldwide and about 850 spe¬
cies are recognized. These small peracarids (0.5—2 cm
in length) are mostly marine (a few brackish and fresh¬
water species) and are generally benthic. Subterranean
species are not well studied and only 16 stygobionts
placed in five genera and three families are known
from blue holes and submarine caves on Andros
Island (Bahamas), Bermuda, and Jamaica (see Petrescu
et al. 1994) (Table 10).
SUPERORDER EUCARIDA
Eucarida is the final superorder treated herein and is
characterized by crustaceans having a carapace covering
and fused dorsally with the head and all thoracomeres
(cephalothorax); they usually have stalked compound
eyes; gills are thoracic; and the telson lacks caudal rami.
Order Decapoda is the only eucarid found in subterra¬
nean environments. This order is amazingly diverse with
nearly 15,000 species known from marine, brackish, fresh¬
water, and terrestrial habitats. Although widespread in
distribution, these crustaceans are significant subterranean
players in temperate and tropical regions of the world
dominated by karst or volcanic terrains and are well repre¬
sented in the western Atlantic, Caribbean region. Central
and North America, and numerous Pacific Ocean islands.
They are characterized by a well-developed carapace
enclosing the branchial chamber, are unique among
decapods in possessing three pairs of maxillipeds, and
have five pairs of functional uniramous or weakly bira-
mous pereiopods (thus the derivation of the order
name). The order is split into two suborders, the
Dendrobranchiata (about 450 species, mostly penaeid
and sergestid shrimps) and the Pleocyemata (all
remaining decapods as well as all stygobionts). The
only dendrobranchiate decapod that occupies subterra¬
nean waters is a stygoxenic penaeid shrimp, Penaeus
indicus H. Milne Edwards, found in Mangapwani Cave
in East Africa. The suborder Pleocyemata contains all of
the remaining decapods and is geographically wide¬
spread globally. It is commonly divided into 7 infraor¬
ders, only 5 of which have stygobiotic representatives:
Stenopodidea (Fig. 8), Caridea, Astacidea, Anomura,
FIGURE 8 The stygobiotic stenopodid, Macromnxillocaris baha-
maensis Alvarez, Iliffe, and Villalobos, from an anchihaline cave in
the Bahamas. Photo courtesy of Tom Iliffe. Used with permission.
and Brachyura. The two remaining pleocyemate infra¬
orders are represented in subterranean environments
by the stygophilic Jasus edwardsii (Hutton) from an
intertidal marine cave in southeastern New Zealand
(Palinura), by the stygoxenic Stenopus spinosns in a
lava tube in the Canary Islands, by Odontozona addaia
from a marine cave on the island of Minorca, Spain
(Stenopodidea), and by a probable stygophile Naushonia
manningi Alvarez, Villalobos, and Iliffe from an anchialine
cave on Acklins Island, Bahamas (Thalassinidea).
Approximately 2800 species of natant decapods are
assigned to the infraorder Caridea. These caridean
shrimps have chelate first or second pereiopods (no
chelation in the genus Procaris ); first pleopods are
somewhat reduced; and the pleuron of the second
abdominal tergite is enlarged, overlapping that of the
first and third. Global in distribution, only some 111
species in 40 genera and six families are known from
anchialine habitats, blue holes, caves, cenotes, springs,
and wells in some temperate but mostly tropical lati¬
tudes (Table 11).
ENCYCLOPEDIA OF CAVES
192
C
TABLE 11 Occurrence of Stygobiotic Caridean Shrimps
Family
Genus/
Subgenus
Species
Occurrence
Procarididae
1
3
Anchihaline
Atyidae
15
55
Anchihaline; marine caves; cenote; coral limestone pools; saline caves, pools, and wells; wells (see
Sket and Zaksek 2009)
Agostocarididae
1
2
Anchihaline, blue hole, cenote
Palaemonidae
11
30
Anchihaline, caves, coral rock pools, fissures, submarine caves, subterranean water, wells
Alpheidae
6
12
Anchihaline, caves, coral rock pools, lava pools, lava tubes (marine), submarine caves (see Anker
2008)
Hippolytidae
6
9
Anchihaline, blue holes, caves, cenotes, pools in coral or lava, sea caves
Total
40
111
TABLE 12 Stygobiotic Crayfishes (Astacoidea: Cambaridae) in the United States, Cuba, and Mexico
Genus
Subgenus
Species/Subspecies
Occurrence
Cambarus
Aviticambarus
3
United States: Alabama, Tennessee
Erebicambarus
hubrichti
United States: Missouri
Jugicambarus
6
United States: Arkansas, Florida, Georgia, Missouri, Oklahoma
Puncticambarus
nerterius
United States: West Virginia
Orconectes
Orconectes
9
United States: Alabama, Indiana, Kentucky, Missouri, Tennessee
Procambarus
Austrocambarus
5
Cuba; Mexico: Oaxaca, Veracruz
Leconticambarus
1
United States: Florida
Lonnberguis
2
United States: Florida
Ortmannicus
11
Mexico: San Luis Potosi; United States: Florida
Remoticambarus
pecki
United States: Alabama
Troglocambarus
maclanei
United States: Florida
The infraorder Astacidea, which includes about 900
species of crayfishes and chelate lobsters (>500 of
those are crayfishes), is global in distribution (all conti¬
nents except Africa), and the first three pairs of pereio-
pods are chelate. Crayfishes are the only astacideans
that are stygobiotic and, of these, only the superfamily
Astacoidea is represented in the subterranean environ¬
ment. Some 41 species from a single family (Cambaridae)
have been described as obligate cave-dwelling species
and subspecies; about 50 additional pigmented, stygoxe-
nic, or stygophilic crayfishes commonly invade subterra¬
nean waters around the globe (Astacidae, Cambaridae,
Parastacidae). The following four genera are assigned
stygobiotic species and subspecies from Cuba, Mexico,
and North America (primarily in the Appalachian,
Florida Lime Sink, Interior Low Plateau, and Ozark karst
regions): Cambarus (11 species), Orconectes (9 species and
subspecies), Procambarus (20 species and subspecies).
and the monotypic Troglocambarus (Table 12). Most of
these are troglomorphic and are opportunistic omnivores
(see Hobbs et al. 1977).
Although diverse, the infraorder Anomura has only
two stygobiotic species assigned to two families
(Table 1). They possess a variably shaped carapace; the
first pair of pereiopods is chelate, the third pair never,
and the fifth pair reduced (never used for walking)
and function as gill cleaners. The pleopods are reduced
or absent. The superfamily Galatheoidea is represented
by two families having subterranean species: Aeglidae
and Galatheidae. A eg la cavernicola Tiirkay is a poorly
studied aeglid, known only from a cave in Sao Paulo,
Brazil. The aggressive galatheid Munidopsis polymorpha
Koelbel has received much attention (life history and
agonistic, feeding, and reproductive behaviors) and is
restricted to a single cave (lava tube) and wells (marine
groundwater) on Lazarote, Canary Islands.
ENCYCLOPEDIA OF CAVES
CRUSTACEA
193
TABLE 13
Occurrence of Stygobiotic Brachyurans
Family
Genus Species Occurrence
Hymenosomatidae
3
3
Caves: Indonesia
T richodacty lidae
2
2
Caves: Chiapas and Tabasco, Mexico
Goneplacidae
1
1
Caves: New Britain, Indonesia
Xanthidae
1
1
Anchihaline cave: Ecuador
Potamidae
5
7
Caves: Sarawak, Borneo; Laos; Palau Tioman, Malaysia; Thailand; Ryukyu Islands
Pseudothelphusidae
5
20
Caves: Belize; Chiapas, Mexico; Colombia and Venezuela, South America; Guatemala; Cuba
Gecarcinucidae
5
6
Caves: Sarawak, Borneo; Thailand; Ryukyu Islands
Parathelphusidae
2
10
Caves: Irian Jaya, Indonesia (stygophile?); Philippines
Sundathelphusidae
4
7
Caves: New Britain, Indonesia; Papua New Guinea; Philippine Islands
Hydrothelphusidae
1
1
Cave: Madagascar
Grapsidae
2
7
Anchihaline, caves: Guam; Jamaica; Java, Nusa Lain Island, Indonesia; Niue Island, Polynesia; Papua
New Guinea: Solomon Islands
Sesarmidae
1
1
Cave: Jamaica
Varunidae
1
1
Anchihaline cave: Ryukyu Islands
33
67
The cephalothorax of the infraorder Brachyura
(true crabs) is dorsoventally flattened and commonly
expanded laterally; the abdomen is reduced, symmetri¬
cal, and flexed beneath the thorax; uropods are usually
absent; first pereiopods are chelate; and the males lack
third, fourth, and fifth pairs of pleopods.
Approximately 7000 crabs are recognized, most of
which are marine; however, freshwater and tropical
terrestrial species occur in epigean and hypogean
habitats. At least 67 crabs assigned to 33 genera and
13 families are categorized as stygobionts (Table 13).
Additionally, the following families are represented
as facultative cavernicoles in primarily tropical
subterranean settings (mostly caves and anchialine
habitats; numbers are approximate): Trichodactylidae
(1 genus, 2 species), Xanthidae (2, 2), Potamidae (7, 9),
Potamonautidae (2, 2), Pseudothelphusidae (10, 21),
Gecarcinucidae (2, 2), Parathelphusidae (1, 5),
Sundathelphusidae (5, 5), Hydrothelphusidae (1, 1),
Gecarcinidae (3, 5), and Grapsidae (5, 12).
See Also the Following Articles
Asellus aquaticus: A Model System for Historical
Biogeography
Gammarus minus: A Model System for the Study of
Adaptation
Bibliography
Anker, A. (2008). A worldwide review of stygobiotic and stygophilic
shrimp of the family Alpheidae (Crustacea, Decapoda, Caridea).
Subterranean Biology, 6, 1 — 16.
Botosaneanu, L. (Ed.), (1986). Stygofauna mundi, a faunistic,
distributional, and ecological synthesis of the world fauna inhabiting
subterranean waters (including the marine interstitial). Leiden: E. J.
Brill.
Boxshall, G. A., & Huys, R. (1989). New tantulocarid, Stygotantulus
stocki, parasitic on harpacticoid copepods, with an analysis of the
phylogenetic relationships within the Maxillopoda. Journal of
Crustacean Biology, 9(1), 126—140.
Culver, D. C., & Wilkens, H. (2000). Critical review of the relevant
theories of the evolution of subterranean animals. In H. Wilkens,
D. C. Culver, & W. F. Humphreys (Eds.), Ecosystems of the world
30: Subterranean ecosystems (pp. 381—398). Amsterdam: Elsevier
Press.
Gutu, M., & Iliffe, T. M. (1989). Description of two new species of
Tanaidacea (Crustacea) from the marine water caves of the Palau
Islands (Pacific Ocean). Travaux du Museum d'Histoire Naturelle
Grigore Antipa, 30, 169—180.
Gutu, M., & Iliffe, T. M. (1998). Description of a new hirsutiid (n.g.,
n.sp.) and reassignment of this family from order Mictacea to
the new order, Bochusacea (Crustacea, Peracarida). Travaux du
Museum d'Histoire Naturelle Grigore Antipa, 40, 93—120.
Hobbs, H. H., Jr., Hobbs, H. H., Ill, & Daniel, M. A. (1977). A review
of the troglobitic decapod crustaceans of the Americas. Smithsonian
Contributions to Zoology, 244, 1—183.
Hobbs, H. H., Ill (2000). Crustacea. In H. Wilkens, D. C. Culver, &
W. F. Humphreys (Eds.), Ecosystems of the world 30: Subterranean
ecosystems (pp. 95—107). Amsterdam: Elsevier Press.
Humphreys, W. F., & Harvey, M. S. (Eds.), (2001). Subterranean biology
in Australia 2000 Records of the Western Australian Museum
(Supplement No. 64).
Martin, J. W., & Davis, G. E. (2001). An updated classification of the
recent Crustacea (Science Series 39). Los Angeles, CA: Natural
History Museum of Los Angeles County, 1—24.
Petrescu, I., Iliffe, T. M., & Sarbu, S. M. (1994). Contributions to the
knowledge of cumaceans (Crustacea) from Jamaica. II. Five new
species of the genus Cumella. Travaux du Museum d'Histoire
Naturelle Grigore Antipa, 34, 347—367.
ENCYCLOPEDIA OF CAVES
194
c
Pipan, T. (2005). Epikarst —A promising habitat: Copepod fauna, its diver¬
sity and ecology: A case study from Slovenia (Europe). Carsologica 5,
1 — 101 .
Poore, G. C. B., & Humphreys, W. F. (1998). First record of
Spelaeogriphacea from Australasia: A new genus and species
from an aquifer in the arid Pilbara of Western Australia.
Crustaceana, 71(7), 721—742.
Sket, B., & Zaksek, V. (2009). European cave shrimp species
(Decapoda: Caridea: Atyidae), redefined after a phylogenetic
study: Redefinition of some taxa, a new genus and four new
Troglocaris species. Zoological Journal of the Linnean Society, 155(4),
786-818.
Wagner, H. P. (1994). A monographic review of the
Thermosbaenacea (Crustacea: Peracarida): A study on their
morphology, taxonomy, phylogeny and biogeography. Zoologische
Verhandelingen (Leiden), 291, 3—338.
Zinn, D. J. (1986). Mystacocarida. In L. Botosaneanu (Ed.),
Stygofauna mundi, a faunistic, distributional, and ecological synthesis
of the world fauna inhabiting subterranean waters (including the
marine interstitial) (pp. 385—388) Leiden: E. J. Brill.
ENCYCLOPEDIA OF CAVES
DINARIC KARST: GEOGRAPHY
AND GEOLOGY
Nadja Zupan Hajna
ZRC SAZU Karst Research Institute
TRAITS OF THE DINARIC KARST
There are many important karst regions around the
world, but the Dinaric karst remains the "classical"
karst for many reasons. A large limestone region
with similar style of landscape, it is the type-site for
many features and phenomena. The term karst (kras) is
derived from the Kras plateau (the northwest part
of the Dinaric karst). From the region originates inter¬
national terms such as polje, uvala, doline, kamenitza,
and potior. The Dinaric karst is also the landscape
where karstology and speleology as sciences were
born. Already three World Heritage properties from the
region, Plitvice Lakes (Croatia), Skocjanske Jame
(Slovenia), and Durmitor National Park (Montenegro),
are inscribed on the UNESCO list.
The Dinaric karst is geographically and geologically
the carbonate part of the Dinaric Mountains (Dinarides),
which form the western part of the Balkan Peninsula
and the entire Adriatic Sea littoral belt, between
42—46°N and 14—21°E. The name of the mountain belt
derives from Dinara Mountain, extending in a north¬
west-southeast direction for a length of 84 km between
Croatia and Bosnia, with Troglav (1913 m) as the highest
peak. The Dinarides (Dinaric Alps) from a tectonic
view include all of the External (Outer) Dinarides and
Internal (Central) Dinarides. According to some authors
the Dinarides also include the Southern Limestone Alps
in Slovenia, but for various reasons they are considered
part of the Alps.
The boundary line of the Dinaric karst differs
depending on whom you consult (e.g., Roglic, 1965;
Gams, 1974) and more or less follows the boundary of
carbonate rocks toward the Pannonian basin (Fig. 1).
Present in the region are some noncarbonate rocks, such
as the Eocene flysch rocks in the External Dinarides and
clastic sediments and igneous rocks in the Internal
Dinarides. The western border of the Dinaric karst is
represented by the Adriatic Sea. The northern border is
the border between the Alps and the Dinarides and fol¬
lows the Soca, Idrijca, and Sava rivers. The eastern bor¬
der is not very clear, because there carbonate rocks
alternate with noncarbonates of the Internal Dinarides,
the Dinaridic ophiolite zone, and the Panonian Basin
(Fig. 2). The eastern border is more or less a line from
Samobor and Karlovac in Croatia, Banja Luka in Bosnia,
and Pec in Kosovo. According to Roglic (1965), carbon¬
ate areas between Sarajevo, Uzice, and Durmitor are
also included in the Dinaric karst because, from a geo¬
logical point of view, these are Adria-derived thrust
sheets (e.g., Durmitor, Sandzak plateau). The southern
border is also not very distinctive; it more or less follows
the last carbonates of Prokletije mountain, northwest of
the river Drim/Drin in Albania.
From the northwest to the southeast, the Dinaric
karst is about 650 km long and up to 150 km wide.
Covering more than 60,000 km 2 this is the paramount
karst region in Europe. The highest peaks are more
than 2000 m high and some of the karst features (e.g.,
caves with speleothems) are now below the present
sea level. According to different geological, hydrologi¬
cal, climate, and geomorphic characteristics, the whole
Dinaric karst can be divided into three belts parallel
to the Adriatic Sea: low coastal Adriatic karst, high
mountain karst, and low continental interior karst.
The Dinaric karst is situated in a humid temperate
climate zone, with prevailing west winds. Along the
Adriatic Sea and along river valleys toward Dinaric
Mountains there is a Mediterranean climate with
dry and hot summers and wet and fresh winters. The
average precipitation there is about 800 mm per year.
On the south sides of the Dinaric Mountains a strong
downslope northeast wind burja (bora) is typical. The
Encyclopedia of Caves.
195
© 2012 Elsevier Inc. All rights reserved.
196
D
FIGURE 1
FIGURE 2
Approximate borders of the Dinaric karst. After Roglic (1965) and Gams (1974).
Simplified tectonic map of the Dinaric Mountains. After Schmid et al. (2004) and Sumanovac et al. (2009).
ENCYCLOPEDIA OF CAVES
DINARIC KARST: GEOGRAPHY AND GEOLOGY
197
inner parts of the Dinaric karst have a mountainous
climate, especially on higher elevations of karst pla¬
teaus; a moderate continental climate is present in the
intermountain basins and in transition areas toward
the Pannonian basin. The Dinaric Mountains act as
an orographic barrier where the highest amount of
precipitation falls (an average of about 3000 mm per
year). The highest annual precipitation is above Boka
Kotorska bay (about 5000 mm per year).
The vegetation corresponds to the climate and the
soil cover, which on the karst is not very abundant.
The Dinaric karst is also well known as limestone des¬
ert; the bare rocky landscape was and is the result of
climate conditions, and intense land use over the
centuries. There is almost no vegetation on the islands
(Fig. 3) nor along the coast where salt water sprayed
by bora impedes all growth. There are large grasslands
with plant species that vary from sub-Mediterranean
to alpine floral elements. Forest is present on higher
parts of the plateaus and mountains (e.g., Trnovski
gozd, Sneznik, Velebit, Kapela). From the coast toward
the high peaks of the mountains, the typical vegetation
is as follows. From the coast toward the mountains
are degradated forests of pubescent oak and horn¬
beam; they are gradually followed by forests of pubes¬
cent oak and flowering ash. Vertically, between 700
and 1300 m, are coastal beech forests with the indige¬
nous forests of the black pine. The continental slopes
of mountains at elevations between 600 and 900 m are
covered with alpine beech forests. They extend quite
low on the continental side, sometimes descending
to the edges of karst plains. Elevations between
1200 and 1400 m are inhabited by beech and fir forests
(Abieti-Fagetum dinnaricum ) which are also the most
common forests of the Dinaric karst. Higher up
(above 1600 meters), there is normally a subalpine
beech forest. In different areas are well-preserved
FIGURE 3 Bare rocky landscape on Pag Island in Croatia. In the
foreground is thick-bedded, well-fissured Cretaceous limestone, and
in the distance are thin-bedded Paleogene limestones on which stone
walls were made on the boundaries of plots. Photo by the author.
spruce forests, which inhabit altitudes over 1400 m.
The highest parts of the mountains are covered with
the indigenous virgin forest of dwarf mountain pine.
The hydrological characteristics of the Dinaric karst
are the result of lithology, tectonic structures, climate,
and geomorphic evolution. Water in the Dinaric
karst is drained toward the Adriatic and Black Sea.
There are only a few surface streams in spite of high
amounts of precipitation. In the areas with prevailing
dolomites, fluviokartic drainage is developed. On karst
poljes are sinking rivers, usually flowing underground
from one polje to another. Underground connections
between different karst poljes of the Dinaric karst are
known, such as the underground water connections
between Gatacko, Nevesinjsko, Fatnicko, Dabarsko,
and Popovo poljes (Bosnia and Herzegovina) and
Ombla spring at Dubrovnik (at sea level, Croatia).
Another example is the river Ljubljanica in Slovenia,
draining 1100 km 2 (a mean discharge of 56 m 3 s -1 ),
which crosses four karst poljes. A few large rivers
cross the carbonate belts along the Dinaric karst:
Kolpa, Zrmanja, Krka, Cetina, Neretva, Moraca, Una,
Vrbas, Bosna, and Drina are the largest. Most of the
rivers form deep canyons (e.g., Cetina, Pliva, Tara). On
some of the rivers tufa deposits are abundant (e.g.,
Krka, Zrmanja, Una, and Pliva). Some of the more
famous bodies of water are the Plitvice lakes with
dams and waterfalls on the Korana river (Croatia);
they were added to UNESCO's World Heritage List in
1979. Huge karst springs are typical of the region, with
up to a few hundred m 3 s -1 maximum discharge (e.g.,
Buna (Fig. 4), Bunica, Trebisnjica, and Ljuta). Some of
the springs along the Adriatic Sea are located under
present sea level (e.g., Donjobrelska Vrulja below
Biokovo, and Zecica below Velebit); locally they are
called vrulje.
FIGURE 4 Efficacious karst spring of Buna in Bosnia. Photo by the
author.
ENCYCLOPEDIA OF CAVES
198
D
STRUCTURAL GEOLOGY OF
THE DINARIDES
The geological evolution of the Dinarides is closely
associated with the history of the Tethys Ocean, which
closed during the Mesozoic and Cenozoic as a result
of the convergence of the African and Eurasian plates.
Intermediate microplates played an important role in
ocean shortening. The present geological structures
resulted from post-collision processes in the Alpine
orogenic system, which started before 35 Ma (Vrabec
and Fodor, 2006). The main period of thrusting and
folding of the area is post-Eocene—the result of post¬
collision processes between the Africa and Europe
plates. The latest tectonic phase in the region started
by counterclockwise rotation of the Adria microplate
approximately 6 Ma BP. The rotation caused reactiva¬
tion of already existing Dinaric faults, which were
the consequence of the aforementioned thrusting, as
dextral strike-slip faults.
Different opinions about the geologic evolution
of the Dinaric system in the northeastern Adriatic
region reflect its complexity. The main tectonostrati-
graphic units of the Dinarides (Tari, 2002) are related to
Early and Middle Triassic rifting and Late Jurassic-
to-present-day compression.
Early and Middle Triassic rifting was marked by
strong magmatism and horst and graben-related depo¬
sition overlying the Variscian basement. Along the
"eastern" Apulian margin toward the Sava-Vardar,
ocean rifting was followed by subduction-generated
extension from the Early Triassic until the Late
Jurassic. Subduction-related attenuation of the conti¬
nental crust along the eastern margin of Apulia
caused the formation of a back-arc basin in the oceanic
crust. The remnants of these active continental margin
lithologies are found within the eastern thrust belt
as the ophiolite melange of the Internal Dinarides
Ophiolite Belt.
Late Jurassic-to-present-day compression generated
(1) the eastern thrust belt, foredeep and foreland;
(2) the northern Dinarides accretionary wedge; (3) the
western thrust belt, foredeep and foreland; (4) the
eastern Adria imbricated structures; and (5) wrenching
and tectonic inversion. The Dinaridic carbonate plat¬
form toward the west presented the foreland of the
generally west-directed thrusting. During the Early
Cretaceous, compressional stresses began to be trans¬
mitted westward through the Dinarides, causing the
migration of the foredeep basin and regional uplift
of the eastern thrust belt. Subduction of the oceanic
plate along the northern margin of the Dinarides
resulted in the accumulation of this accretionary
wedge from the Maastrichtian to the Eocene. From the
end of the Cretaceous until the Early Eocene the entire
carbonate platform was uplifted. During the Eocene,
the Dinaridic carbonate platform was finally buried
under the flysch deposits in the broad foredeep basin
of the western thrust belt. At the beginning of the
Oligocene, collision and progressive subduction of
the Adria below the Dinarides created the imbricate
structures of the Adria provenance in front of the western
thrust belt. The structural style of the Dinaridic thrust
belt is a result of the polyphase tectonic compression
and the competence of the sedimentary units involved.
The competent carbonate rocks are the strongest influ¬
encing factor on the structural style of the thrust belt.
The compression started with ramping along the
deep decollement from the root zone with a south¬
western tectonic transport. In this way, by progressive
overstepping of the thrust faults, various structural
forms were created along the eastern and western
thrust belt: fault bend folds, tear fault-related folds,
and folded thrust structures reworked by footwall
deformations. The northeast—southwest striking system
of the dextral strike-slip faults during the Oligocene
to Miocene was followed by northwest—southeast
striking and wrenching in the Early and Middle
Miocene, affecting the South Pannonian Basin, the
western thrust belt, and the Adriatic foreland. This
activity is reflected in the large flower structures of the
Dinaridic thrust belt.
The main structures of the Dinaric karst relief have a
northwest—southeast direction, in the so-called Dinaric
direction. This structural predisposition is visually
reflected in the direction of high plateaus and karst
poljes from south Slovenia to Montenegro. From a dis¬
tant view, the Dinaric karst has expressive structural
relief dependent on overthrust structures and on dextral
strike-slip faults (both in Dinaric direction).
LITHOLOGY OF THE DINARIC KARST
The Dinaric karst denotes an area confined mostly to
Mesozoic and Cenozoic carbonates of External Dinarides
and Mesozoic carbonates of Internal Dinarides. Car¬
bonate sedimentary rocks along the Dinarides were
formed on carbonate platform from the Upper Paleo¬
zoic to the Paleocene. The youngest limestones are
present in the belt along the Adriatic coast, from
Eocene, Paleocene, to Cretaceous in age; more or less
Cretaceous and Jurassic limestone and dolomites
are present in the middle (External Dinarides) and
northeastern part (Internal Dinarides); Triassic lime¬
stones and dolomites are mainly parts of the middle
(External Dinarides) and northeastern outer parts
(Internal Dinarides) of the Dinaric karst. Geological
data ( e.g ., Vlahovic et al., 2002) indicate that the
External Dinarides were formed by the destruction of
ENCYCLOPEDIA OF CAVES
DINARIC KARST: GEOGRAPHY AND GEOLOGY
199
a single, yet in morphological terms highly variable,
shallow water carbonate platform: the Adriatic Carbonate
Platform. There is confusion in the scientific literature due
to the different names used for the same shallow-water
carbonate platform and ideas of two existing carbonate
platforms at the same time. The platform was very
dynamic in some periods because of its paleogeographic
position during the Mesozoic, especially during the Late
Cretaceous. The final disintegration of the platform area
culminated in the formation of flysch trough(s) in the Late
Cretaceous and Paleogene and the subsequent uplift of
the Dinarides.
Along the Dinaric karst, karst features are favored
especially in massive and thick bedded Cretaceous
limestones and Jelar Breccia of the Oligocene or youn¬
ger age. Upper Cretaceous massive limestones are
exceptionally well karstified because of their properties
(very pure). In Jelar Breccia solutional karst features
such as karren, dolines, conical hills, and caves are
also well developed. Along the south and southwest
parts of Velebit, where breccia is located, ridges of con¬
ical hills (kuk in the local language) are numerous.
Between these hills, big and deep karst depressions
(duliba in the local language) are developed. Among
the most well known are Hajducki and Rozanski kukovi
near Zavizan (Northern Velebit National Park); Dabarski
kukovi near Baske Ostarije; Bojinac (Fig. 5) above
Paklenica Canyon (National Park Paklenica); and Tulove
grede on the south edge of the Velebit. The thick-bedded
Jelar Breccia, comprised predominantly of angular,
poorly sorted clasts of Cretaceous limestones and dolo¬
mites, Triassic carbonates, and Paleogene limestones,
covers large areas along the northeastern Adriatic coast
on the southwest rim of Velebit Mountain (western
Croatia). The largest outcrop is >100 km long, 2—10 km
wide, and thicker than 500 m in places. The Jelar Breccia
FIGURE 5 Karst features in Jelar Breccia on Bojinac, South
Velebit in Croatia. Photo by the author.
is usually considered a result of the disintegration of
frontal parts of the major thrusts. The large leveled sur¬
faces and caves are developed also in Eocene—Oligocene
Promina Beds (Promina Formation) which consist of
noncarbonate and carbonate parts. Carbonate conglom¬
erates may be seen, for instance, on mountain Promina,
at Sjevernodalmatinska zaravan (northern Dalmatian
Plain; the area between Velebit, Novigradsko more and
river Krka).
During the Pleistocene, Dinaric Mountains with eleva¬
tion above 1100 m were glaciated. Stagnant glaciers were
covering the top of the plateaus with small tongues
flowing over plateau edges. Valley glaciers reshaped the
original fluvial valleys into typical U-shaped valleys and
deposited vast amounts of sediments. There is a number
of characteristic glacial features such as hanging valleys,
remains of moraines, fluvioglacial sediments, erratic
blocks, and striated cobbles and boulders. The remains
of Pleistocene glaciations are visible on mountains and
plateaus from Slovenia to Montenegro; for example,
Sneznik, Risnjak, Velebit, Orjen, Lovcen, and Durmitor.
KARST FEATURES OF THE
DINARIC KARST
The most characteristic relief forms of the Dinaric
karst are high karst plateaus, numerous poljes in the
central part stretched in northwest—southeast direc¬
tion, enormous leveled surfaces, various dolines, large
and deep caves, sinking rivers, and abundant springs.
Karst forms are surface and subsurface karst features
formed mainly by solutional processes of carbonate
rocks of the region which are extremely well folded
and fractured due to tectonic events. Linear faulting is
rare in carbonate rocks; usually carbonates are frac¬
tured to different degrees. According to the degree
of deformities of the bedrock, we distinguish in carbo¬
nates fissured (fissures), broken (blocks of rock), and
crushed zone (clay, silt, breccia). The presence of
fractured zones is very important for karstification and
formation of karst features usually favor defined ones;
for example, fissure zones are important for formation
and location of karren fields, giant grikelands, and
strings of dolines.
On various limestones, dolomites, and so on, several
kinds of karren are developed, especially on massive
and thick-bedded limestone. Due to its specific mech¬
anical and petrographic properties, the Jelar Breccia is
characterized by an abundance of dissolution features
on exposed surfaces ( e.g ., solutional runnels, heel-
prints, corrosional steps, kamenitzas).
The most typical karst features of the Dinaric
karst are depressions of various sizes: dolines, col¬
lapse dolines, uvalas, and poljes. Dolines are closed
ENCYCLOPEDIA OF CAVES
200
D
karst depressions of different sizes and genesis
(Ford and Williams, 2007). Solution dolines form
where vertical karst drainage is present and sup¬
ported with suitable lithology, geological structures,
slope inclination, and long enough time. In some
areas their distribution is very dense, up to 200
dolines per km 2 . Where dolines totally pock some
parts of the surface and occupy almost all the space,
we are dealing with typical polygonal karst (egg-box¬
like topography). Landscapes displaying polygonal
karst can be found in many places in the Dinaric
karst; for example, on Biokovo above Makarska, on
plateaus above Boka Kotorska bay, and on high pla¬
teaus in Herzegovina. Where the principal process of
the formation of a doline is the collapse above under¬
lying caves, it is called a collapse doline. Collapse
dolines are common features of the Dinaric karst and
are especially numerous above known large caves.
More than 15 large collapse dolines are located above
the underground stream of river Reka which sinks to
Skocjanske Jame in southwest Slovenia. The biggest
collapse dolines are near Imotski (Croatia). With a
depth of 528 m, Crveno Jezero is the deepest of
them; its bottom water is about 280 m deep. On the
Dinaric karst many dolines are inherited forms of
primary caves and shafts (Mihevc, 2007), which, due to
surface denudation, become part of the karst relief.
Many large and deep closed depressions are located
along the high Dinaric plateaus (e.g., Sneznik, Velebit).
They are named konta, draga, or duliba and their origin
as solutional or collapse dolines is not always clear.
During the Pleistocene they were significantly remo¬
deled by glaciation and frost action and in some of
them moraines are found.
Uvalas are polygenetic-formed karst closed depres¬
sions, which are connected to the geological structure
(Calic, 2009), usually in broken (well-fractured) zones
of regional extension. They are larger than dolines
(km-scaled), they are mainly disconnected with the
karst water table, and they can look like a small polje
without a surface stream. Uvalas are frequent relief
features of the Dinaric karst but they are not devel¬
oped at low elevations and on leveled surfaces.
Karst poljes are large karst depressions with flat bot¬
toms, which are the result of solutional processes and
karst water drainage. Due to different geological and
hydrological factors Gams (1978) distinguished five
types of karst poljes: border polje, piedmont polje in alluvial
valley, peripheral polje, overflow polje, and baselevel polje.
Most of large karst poljes of the Dinaric karst are basele¬
vel poljes in their origin even if they have some noncar¬
bonate rocks at their bottom. Water on the surface
causes base-leveled corrosion, lateral corrosion (mar¬
ginal corrosion), and consecutive leveling of the rocky
bottom at the water table. Oscillation of the water
table causes hydrological phenomena such as springs,
ponors, estavelas, floods, and intermittent lakes.
The karst poljes of the Dinaric karst are structur¬
ally controlled and almost all of them are stretched
along overthrust edges and fault zones of regional
importance in so-called Dinaric direction (north¬
west—southeast). Between parallel dextral strike-slip
faults or between normal faults, separate blocks are
sunk to different elevations (e.g., pull-apart basin
formation; graben formation). Some of them are just
on the water table level, or they were in the past.
These blocks of carbonates at the water table are, or
were, leveled and sometimes also laterally enlarged
by corrosion and karst poljes were formed. Some
parts of the poljes sank, or they are still sinking, and
these immersed parts are usually filled up with
sediments. Many of the karst poljes along the Dinaric
karst contain fluvial, lacustrine, or glacial Neogene
and Quaternary sediments up to few hundred meters
deep (e.g., Nevesinjsko, Gacko, Glamocko, and Livanjsko
poljes). The proofs for existing lakes in sinking basins
during the Miocene and Plio-Pleistocene are also fossil
fauna and coal-bearing deposits in some of the karst
poljes (e.g., Kocevsko, Sinjsko, Duvanjsko, Livansko, and
Gacko poljes) from Slovenia to Montenegro. Older sedi¬
ments are in karst poljes in the northeast and central
area of the Dinaric karst, while in the bottoms of the
poljes on the southwest side of the highest karst plateaus
(e.g., the southwest part of Slovenia, Istria, and Velebit,
and along the Adriatic Sea) sediments are in thinner
layers or they do not exist (e.g., Planinsko polje in
Slovenia). It is also seen that in some of the Dinaric fault
zones in the area, karst poljes are not developed yet
(blocks are not developed or sink to the water table; e.g.,
along the Rasa fault in southwest Slovenia). The age of
the sediments in the karst poljes (whether Miocene,
Pliocene, Pleistocene, or Holocene) decline in age from
east to west and the presence of the developed karst
poljes along fault zones is in connection with the youn¬
gest dextral strike-slip tectonic influence by recent Adria
microplate movements. Below the sediments, more
or less leveled rocky bottoms of karst poljes exist. At dif¬
ferent elevations of karst polje bottoms or at their edges
(e.g., polje of Lika, Nevesinjsko polje, and Fatnicko polje)
or even between poljes (e.g., between Fatnicko and
Dabarsko poljes) are leveled surfaces, which are rem¬
nants of a time when polje bottoms were corroded at the
water table as one single unit and now they are as sepa¬
rate blocks left behind (they are no longer sinking). On
some of these older leveled blocks (surfaces) dolines
are already formed (e.g., the north central part of
Nevesinjsko polje and the south edge and southeast part
of Fatnicko polje).
Leveled surfaces are often connected with baselevel
corrosion on karst poljes, but they can be also produced
ENCYCLOPEDIA OF CAVES
DINARIC KARST: GEOGRAPHY AND GEOLOGY
201
after a long period of denudation on the input margin
(Ford and Williams, 2007). The complex of processes
producing corrosional plains in karst has been called
lateral solution planation or corrosional planation and
involves a combination of vertical dissolution, lateral
undercutting of hillsides, and spring head sapping.
Larger leveled plains are found along the Dinaric karst
in south Slovenia, in Istria, between Karlovac (Croatia)
and Bihac (Bosnia), Dalmatia (Fig. 6), and in Bosnia and
Herzegovina. Corrosional plains cut different geological
structures ( e.g ., folds, lithology) in the larger area. In
local languages, leveled surfaces are named ravnik,
podolje, and zaravan. In some of them rivers formed can¬
yons, for example, at Krka in the Sjevernodalmatinska
zaravan (north Dalmatian karstic plain) or at river
Cetina in the Zadvarska zaravan; some of old ones are
already dissected by dolines and/or conical hills (e.g.,
Kras plateau in Slovenia).
Contact karst develops on the contact between per¬
meable (karst) and nonpermeable (nonkarst) rocks,
for example, on the contact between carbonates and
Paleozoic and Mesozoic clastic rocks or Eocene flysch
rocks on the outer borders of the Dinaric karst or
within its borders. Contact karst is a landform associ¬
ated with allogeneic input. Different karst features are
limited with the border belt of carbonate areas and
they are influenced by sinking rivers. There are many
big sinking rivers along Dinaric karst (e.g., Reka sinks
in Skocjanske Jama, river Pivka sinks in Postojnska
Jama, and Zalomka River sinks to the Biogradski
ponori). Remnants of old river streams are dry valleys,
which lack an active watercourse over all or parts
of their length for at least part of the year, because
water is lost in karst or is in a much lower position.
For instance, there are two dry valleys crossing the
Kras plateau in southwest Slovenia: the largest one in
Slovenia is the 400-m-deep Cepovanski dol between
Trnovski gozd and Banjscice plateaus, and another big
dry valley is situated in upper part of river Bregava
between Dabarsko polje and Stolac in Herzegovina.
A contact karst feature is also the blind valley, which is
a valley cut by a stream flowing from insoluble rocks
onto soluble rocks and sinking in a ponor. A series of
blind valleys is developed in Matarsko podolje in
southwest Slovenia, where along geological contact
between flysch and limestones 17 brooks sink into the
karst edge. Blind valley bottoms are widened by basele-
vel and lateral corrosion (Mihevc, 2007).
More than 20,000 caves are known in the Dinaric
karst; they are located from the high mountains to
below present sea level in the form of vertical shafts or
horizontal passages. Some of the caves are active and
some of them are relict parts of karst drainage systems.
Caves are developed in various carbonate rocks of
the region and their formation was influenced by the
local presence of different geological structures and
position of water gradient (base level). The deepest
explored caves are known from the North Velebit in
Croatia; for example, shafts Lukina Jama-Trojama
(—1392 m), Slovacka Jama ( — 1320 m), Velebita cave
system (—1026 m) with a 513-m underground vertical
shaft, and Meduza (—679 m). Several significant caves
were discovered in the Jelar Breccia and Liassic lime¬
stone of the Crnopac area (South Velebit). Their main
morphological characteristic is a network of multi¬
phase cave passages, some of them with very large
cross-sectional dimensions. The most important caves
of Crnopac massif are Munizaba (5993 m long,
—437 m) and Kita Gacesina (10,603 m long, —456 m).
There are many horizontal caves linked to the ponors
such as Postojnska Jama (a 20,570-m-long system of 5
caves) and Skocjanske Jama (6200 m long) in Slovenia,
and Dulin ponor-Medvedica (16,396 m long) in
Croatia. Some of the horizontal caves are managed for
cave tourism; notably Postojnska Jama, Cerovacke
pecine (3800 m long) in Croatia, Vjetrenica (6700 m
long) on Popovo polje in Bosnia and Herzegovina, and
Lipska pecina (3500 m) in Montenegro.
Twenty years ago the prevailing thinking was that
the surface of the Dinaric karst was formed in the
Middle Pliocene (Habic, 1991). Before then karst was
covered and contained by impermeable rocks, and
waters flowed superficially over impounded carbonate
rocks. Later tectonic movements contributed to surface
transformation and to karst dissection and surface
streams disappeared into karstic underground. The
evidence of the surface water flows were clastic
(fluvial) sediments found on the karst surface. With
FIGURE 6 Sjevernodalmatinska zaravan (north Dalmatian karstic plain in Croatia) cuts different geological elements; in the distance is
Promina Mountain. Photo by the author.
ENCYCLOPEDIA OF CAVES
202
D
FIGURE 7 Unroofed cave near Grotta Gigante at the Italian part
of Kras plateau. Photo by the author.
different methods the denudation rate in humid tem¬
perate zone was calculated to about 20—50 m per mil¬
lion years (Gams, 1974). With this rate of surface
denudation many caves were already breached and
their ceilings were removed—this is how roofless caves
were formed (Mihevc, 2007; Fig. 7)—and from some
caves only their insoluble clastic sediments remain on
the karst surface. Unroofed caves are calculated to be
older than 5 Ma BP. The paleomagnetic and magnetos-
tratigraphic research of karst sediments in Slovenia
(Zupan Hajna et al., 2008) has been carried out since
1997. More than 2500 samples were taken and analyzed
in 36 different profiles at 21 locations in caves and on
the surface. The results were in some sites calibrated by
the thorium-uranium (Th-U) method and by paleonto¬
logical and geomorphological analyses. Calibrated data
contributed to reconstruction of speleogenesis, deposi¬
tion in caves, and indirectly to evolution of karst sur¬
faces and succession of tectonic movements. The
evolution of caves in Slovenia took part within one
post-Eocene karstification period. The period contains
distinct phases of massive deposition in now-existing
caves with still preserved sediments dated to about
5.4—4.1 Ma BP (Miocene—Pliocene), 3.6—1.8 Ma BP
(Pliocene—Quaternary), and Quaternary. From the age
of the cave sediments it follows that the caves have to
be even older.
LAND USE
The Dinaric karst has a long history of human
impact such as deforestation and transformation into
stony semidesert, and also a long history of reforesta¬
tion (Gams and Gabrovec, 1999). Originally the karst
was completely forested. When the human economy
changed to stockbreeding and farming, the process of
anthropogenic deforestation began. In the history there
have been two main reasons for deforestation (Kranjc,
2008): (1) economic (the requirements of new land,
pastures, timber use, and trade) and (2) social (local
increases in population, mass migration, wars, and
raids). Nowadays dense natural forests, extensive forest
plantations, dry karst shrublands, and also completely
barren karst areas can all be found on the Dinaric karst.
The main regions of the Dinaric karst, the Kras
plateau, the islands, and the low plateaus of the
Dalmatian coast, present the types of land use typical of
Mediterranean countries which produce changes in the
surface (Nicod, 2003): the deforestation and adaptation
of the karst surface for agriculture; the stone clearing
effects; the use of extracted stones for dry-stone walls
and hillslope terraces; and the land reclamation and
management in the dolines, uvalas, and poljes. Dry
stone wall terraces were built around hillsides in order
to limit soil erosion and to provide small fields for
agriculture. Stones removed from the soil have mostly
been accumulated in dry walls. In some areas the only
agriculture surfaces were cultivated dolines. With the
declining role of agriculture, increased social mobility,
and an aging and insufficient agricultural workforce,
terraces and small gardens in the dolines have lost their
former role and now they have been almost entirely
replaced by meadows and bushes.
The environmental problems, on which many fac¬
tors interfere, were of interest to various and multidis¬
ciplinary researchers. During the Austro-Hungarian
Empire and Yugoslavia kingdom, surface streams
were regulated, ponors were widened, and screens
were installed in front of them to hold back deposits,
all to shorten the flood time in the karst poljes. After
World War II the construction of dams on rivers and
streams for water reservoirs started. The managements
in the large Dinaric poljes pose very intricate and diffi¬
cult problems ( e.g ., water tightness of artificial lakes
and draining of natural habitats) depending on the
structural and hydrogeological conditions. They have
been studied in many works (e.g., Bonacci, 1987;
Milanovic, 1981).
See Also the Following Article
Diversity Patterns in the Dinaric Karst
Bibliography
Bonacci, O. (1987). Karst hydrology: With special reference to the Dinaric
karst. Berlin: Springer-Verlag.
Calic, J. (2009). Uvala: Contribution to the study of karst depressions, with
selected examples from Dinarides and Carpatho-Balkanides (Doctor
thesis). Nova Gorica, Slovenia: University of Nova Gorica.
Gams, I. (1974). Kras. Ljubljana, Slovenia: Slovenska matica.
Gams, I. (1978). The polje: The problem of definition. Zeitschrift fur
Geomorphologie, 22(2), 170—181.
ENCYCLOPEDIA OF CAVES
DIVERSITY PATTERNS IN AUSTRALIA
203
Gams, I., & Gabrovec, M. (1999). Land use and human impact in the
Dinaric karst. International Journal of Speleology, 28B( 1—4), 55—77.
Habic, P. (1991). Geomorphological classification of NW Dinaric
karst. Acta Carsologica, 20, 133—164.
Kranjc, A. (2008). History of deforestation and reforestation in the
Dinaric karst. Geographical Research, 47(1), 15—23.
Mihevc, A. (2007). The age of karst relief in West Slovenia. Acta
Carsologica, 36(1), 35—44.
Milanovic, P. T. (1981). Karst hydrology. Littleton, CO: Water Resources
Publications.
Nicod, J. (2003). Understanding environmental problems in Dinaric
karst. Dela, 20, 27—41.
Roglic, J. (1965). The delimitations and morphological types of the
Dinaric karst. Nose fame, 7(1—2), 79—87.
Schmid, S. M., Fiigenschuh, B., Kissling, E., & Schuster, R. (2004).
Tectonic map and overall architecture of the Alpine orogen.
Eclogae geologicae Helvetiae, 97, 93—117.
Sumanovac, F., Oreskovic, J., Grad, M., & ALP 2002 Working Group
(2009). Crustal structure at the contact of the Dinarides and
Pannonian basin based on 2-D seismic and gravity interpretation
of the Alp07 profile in the ALP 2002 experiment. Geophysical
Journal International, 179, 615—633.
Tari, V. (2002). Evolution of the northern and western Dinarides:
A tectonostratigraphic approach. EGU Stephan Mueller Special
Publication Series, 1, 223—236.
Vlahovic, I., Tisljar, J., Velic, I., & Maticec, D. (2002). The karst
Dinarides are composed of relics of a single Mesozoic platform:
Facts and consequences. Geologia Croatica, 55(2), 171 — 183.
Vrabec, M., & Fodor, L. (2006). Late Cenozoic tectonics of Slovenia:
Structural styles at the northeastern corner of the Adriatic micro¬
plate. In N. Pinter, G. Grenerczy, J. Weber, S. Stein & D. Medak
(Eds.), The Adria microplate: GPS geodesy, tectonics and hazards
(NATO Science Series, IV, Earth and Environmental Sciences, 61)
(pp. 151—168). Dordrecht, Germany: Springer.
Zupan Hajna, N., Mihevc, A., Pruner, P., & Bosak, P. (2008).
Paleomagnetism and magnetostratigraphy of karst sediments in Slovenia.
(Carsologica, 8, Zalozba ZRC) (p. 266). Ljubljana, Slovenija. Slovenia:
Zalozba ZRC, Postojna.
DIVERSITY PATTERNS
IN AUSTRALIA
William F. Humphreys
Western Australian Museum, University of Adelaide and
University of Western Australia
INTRODUCTION
The diversity of subterranean fauna in Australia,
and tropical areas worldwide, has not long been recog¬
nized. Until recent decades, Australia was thought to
be deficient in overtly cave-adapted (troglomorphic)
animals. This circumstance was considered to have
resulted from a number of causes: (1) the relative
sparseness of carbonate rocks in Australia, as found in
other Gondwanan fragments, compared with the
world average (Fig. 1); (2) the general aridity of the
FIGURE 1 Karst areas of Australia and the bioclimatic zones:
II, tropical; III, subtropical dry; IV, transitional zone with winter rain;
V, warm temperate; II—IV warm temperate/tropical transition zone.
The two pink areas within the common outline represent the Pilbara
(north) and Yilgarn (south) cratons and their associated orogens that
together comprise the western shield. After Hamilton-Smith and
Eberhard, 2000. Graphic by K. G. Grimes, modified.
continent—it is the most arid inhabited continent, two-
thirds of which receives less than 500 mm of rain
annually—generally resulting in both dry caves and
low input of food energy into the underground voids;
(3) the global lack of cave-adapted animals in tropical
areas; and (4) the lack of widespread and repeated gla¬
ciations, which was perceived to be the main factor
driving the evolution of troglobites in the Northern
Hemisphere, then the focus of research on subterra¬
nean animals. Concomitantly, in Australia there was
perceived to be a high proportion of animals found
only in caves but not specialized for cave life, that is,
lacking overt troglomorphisms. Although not articu¬
lated, these arguments would have applied also to sty-
gofauna, the inhabitants of underground waters in
both karstic and alluvial aquifers.
Understanding the biogeography of an area is reliant
on having a broad spatial and taxonomic sample of the
biota, a comprehensive taxonomy, a well-developed
systematic and paleoclimate framework, and a fully
developed geographical understanding (especially of
paleodrainage and plate tectonics). There are serious
deficiencies in information on most of these fields of
endeavor in Australia. The taxonomic and systematic
framework is very patchy and many groups of interest
to hypogean questions remain largely unstudied
(e.g., Thysanura, Collembola, Diplura, Oligochaeta), or
are just beginning to be studied, so it is still too early
for them to contribute in detail to biogeographical
understanding (e.g., many higher taxa in Oligochaeta,
Encyclopedia of Caves.
© 2012 Elsevier Inc. All rights reserved.
204
D
Copepoda, Ostracoda, Amphipoda, Diplura, Gastropoda).
Hence, the focus here will be on some higher taxa for
which there is more adequate information, and on
some systems, such as the groundwater inhabitants of
the smaller voids (mesovoids), for which there is a use¬
ful body of data.
During the last two decades of the twentieth century,
more focused, as well as more widespread, exploration
of caves (Humphreys 2000; Eberhard and Humphreys
2003) and later, in the third millennium, groundwater
and nonkarst substrates, has shown that the Australian
tropics and arid zones contain especially rich subterra¬
nean faunas (Humphreys and Harvey 2001; Austin et ah,
2008). However, no area of Australia has been well stud¬
ied for its hypogean life, the distribution of effort has
been very uneven across the country, and many areas
remain effectively unexplored for subterranean fauna.
Detailed examination of subterranean biology in
Australia is sparse and studies have been largely
restricted to faunal surveys. Prominent karst areas, such
as the Barkley and Wiso regions, have barely been
examined because of their remoteness from population
centers. Other remote areas, such as the Nullarbor, in
which there has been a long history of cave research,
have proved to have sparse hypogean assemblages,
especially among the stygofauna, but interesting occur¬
rences of the misophrioid copepod Speleophria, and
heavily sclerotized troglobitic species such as the cock¬
roach Trogloblatella and the mygalomorph spider
Troglodiphim which has South American affinities, and
very diverse chemoautotrophic microbial communities
form mantles in the saline groundwater (Holmes et ah,
2001). Even within those relatively well-surveyed areas,
the taxonomic effort is seriously underdeveloped. For
example, in one compilation, 63% of the stygofauna
from New South Wales was undescribed. Where species
have been described, there are many oddities, not yet
well placed within their lineage and thus contributing
poorly to understanding the biogeography of the
Australian hypogean biota. The main knowledge base
and most active research are from the western shield
and South Australia.
GEOGRAPHIC FACTORS^
In contrast to the widespread glaciation that directly
influenced many of the classical karst areas in the
Northern Hemisphere, Australia has not been sub¬
jected to extensive glaciation since the Permian. The
biogeography of the hypogean fauna of Australia has
been influenced by the continent's past connections
with Pangaea and Gondwana, as well as having
formed the eastern seaboard of the Tethys Ocean
during the Mesozoic. Australia is a fragment of
Gondwana together with Africa, India, Madagascar,
South America, and Antarctica. Gondwana itself frag¬
mented and Eastern Gondwana (India, Antarctica, and
Australasia) became isolated from South America and
Africa by 133 million years ago. By the Upper
Cretaceous ( ca. 80 million years ago), Australia was
joined only to Antarctica and it formed the eastern sea¬
board of Tethys. These lands shared a Gondwanan
flora and fauna, and when the final separation
between them occurred (45 million years ago), both
lands were well watered and supported cool temperate
and subtropical forests.
The separation of Australia from Antarctica, and its
subsequent rapid northward drift toward Southeast
Asia, has been the most significant factor that has shaped
the Australian subterranean fauna in the Tertiary. It
resulted in the formation of the Southern Ocean seaway
that led to the development of the circum-Antarctic
Ocean winds and currents that markedly altered the cli¬
mate of the Southern Hemisphere, causing Australia to
become much drier. The formation of the Antarctic ice
cap 15 million years ago saw the beginning of a series of
marked climatic fluctuations that have greatly stressed
the Australian (and other Gondwanan) flora and fauna.
Warm and wet interglacial periods alternated with very
dry, cool, and windy glacial stages, but only a small area
of the Eastern Highlands and Tasmania were subject to
extensive ice cover. These cyclic fluctuations, superim¬
posed on a generally increasing and spreading aridity,
provided conditions under which subterranean refugia
played an important role. Most of the detailed molecular
phylogenetic studies of Australian subterranean fauna
indicate an origin of subterranean species associated
with this developing aridity.
Shield Regions and the Cretaceous Marine
Transgressions
Australia has several major shield regions—parts of
the Earth's crust little deformed for a prolonged
period—that have been emergent since the Paleozoic.
The largest is the western shield, which includes the
Pilbara and Yilgarn cratons and associated orogens
(Fig. 1). These stable, truly continental areas of
Australia have a nonmarine, presumably freshwater
history extending through several geological eras. The
Cretaceous marine inundation, at circa 120 Ma BP,
would have eliminated nonmarine life in the sub¬
merged areas (Fig. 2) and only 56% of the current land
area of the continent remained above sea level. This has
important implications for lineages with poor dispersal
ability, as is typical of subterranean fauna. The distribu¬
tion of ancient lineages, both epigean and subterranean,
may be expected to reflect this marine incursion in two
ENCYCLOPEDIA OF CAVES
DIVERSITY PATTERNS IN AUSTRALIA
205
FIGURE 2 Deep history events that have influenced the biogeog¬
raphy of Australian subterranean faunas (see text). The shaded conti¬
nental areas have not been covered by the oceans since the
Paleozoic.
ways. First, ancient terrestrial and freshwater lineages
may have survived on these continually emergent land-
masses. Second, marine ancestors may have become
stranded along the shores as the Cretaceous seas
retreated and today may be represented as relictual
marine lineages now far inland.
Caves and other subterranean habitats can remain
as relatively stable environments over long periods of
time because they are well insulated from the climatic
perturbations that profoundly affect surface environ¬
ments and surface animals. There, a number of ancient
geographical and phylogenetic relictual groups have
survived (Spelaeogriphacea, Remipedia, Thermosbae-
nacea, etc.) (Fig. 3). Owing to their limited potential for
dispersal, their present distributions may contain a
great deal of information about past geography and
climates. The ghost of Cretaceous and earlier marine
transgressions is probably reflected in the distribution
of Phreatoicidea, an ancient group of isopods, in both
their epigean and subterranean forms, the latter being
restricted to the tropics, and it has been well docu¬
mented in the crangonyctoid amphipods (see Box 1).
In this respect aquatic subterranean faunas hold a
special significance because, unlike terrestrial troglo-
bionts (troglobites), the aquatic troglobiont (stygo-
bites) fauna contain many relict species that are only
distantly related to surface forms. These lineages pro¬
vide the most compelling evidence that the distribu¬
tion of some relict fauna occurred through rafting on
tectonic plates moved by seafloor spreading. Recently
a number of notable discoveries of such relict fauna
have been made in Australia whose geographical
distribution and lifestyles suggest origins variously
in Pangaea, Gondwana, Eastern Gondwana, and
Tethys.
Cave Atmosphere
The latitudinal position and general aridity of
Australia make cave atmosphere a significant biogeo¬
graphic determinant in Australia. Cave environments
have traditionally been separated into different
zones—the entrance, twilight, transition, and deep
zones—with characteristics related to the remoteness
from the surface environment, such as more stable
temperature and humidity and reduced light and food
energy input. On the basis of research in the Undara
lava tube, Howarth and Stone (1990) developed the
concept of a fifth zone, the stagnant-air zone, which is
characterized by elevated carbon dioxide and
depressed oxygen levels. Only in such areas were
highly troglomorphic species found in cave passages.
However, in other tropical areas, such as arid Cape
Range, highly troglomorphic species occur in caves
that have unremarkable concentrations of oxygen and
carbon dioxide, some even occurring in sunlight near
cave entrances, but only where the air is saturated, or
nearly saturated, with water vapor.
Howarth (1987) also addressed the importance of
water content in the cave atmosphere, largely from his
Australian studies. Both tropical and temperate cave
systems lose water when the outside air temperature
(strictly, the outside water vapor pressure) drops
below that in the cave. In the tropics, where average
seasonal temperature differences are less than in tem¬
perate regions, caves tend to be warmer than the sur¬
face air at night and cooler during the day. Even if
both air masses are saturated with water, the cave will
tend to dry out as water vapor leaves the cave along
the vapor pressure gradient—the so-called tropical win¬
ter effect.
Owing to widespread aridity, this concept has par¬
ticular relevance to Australia and also in tropical areas
where the general form of many caves (giant grikes,
small and shallow caves) and/or low subterranean
water supply make them vulnerable to drying. Within
this context the extent of the deeper cave zones (transi¬
tion and deep) will fluctuate as the boundary of
threshold humidity levels migrates with the changing
atmospheric conditions farther into or out of the cave.
Such changes occur in ecological time, associated with
daily and seasonal fluctuations in air density and
humidity, and through evolutionary time, in response
to climatic cycles and long-term climatic trends. Such
changes should have little effect on groundwater or on
troglobites in deep caves, which are extensive enough
ENCYCLOPEDIA OF CAVES
206
FIGURE 3 Subterranean animals: (A) Limbodessus eberhardi (Dytiscidae), one of 100 species of blind diving beetles from calcrete aquifers in
the Australian arid zone; (B) unnamed blind philosciid isopod; (C) Ngamarlanguia luisae (Gryllidae: Nemobiinae) from Cape Range, the only
troglobitic cricket in Australia; (D) head of Opliisternon candidum (Synbranchiformes), one of two Australian cavefish; (E) Draculoides vinei
(Schizomida), one of seven species of microwhip scorpions known from Cape Range; (F) the phreatoicidean isopod Phreatoicoides gracilis; (G)
Mangkurtu mityula (Spelaeogriphacea), a subterranean family that is known from only four species, two in Australia, and Africa and Brazil;
(H) Pygolabis humphreysi, from Ethel Gorge calcrete, belongs to a family of flabelliferan isopods, the Tainisopidae, known only from groundwa¬
ter in the Kimberley and Pilbara regions of Western Australia. Photographs (B—E) Douglas Elford, Western Australian Museum; (A and G) William
Humphreys, former from a painting by Elyse O'Grady; and (F and H) GDF Wilson, Australian Museum.
ENCYCLOPEDIA OF CAVES
DIVERSITY PATTERNS IN AUSTRALIA
207
BOX 1
AMPHIPODS
Australia is a major center of amphipod diversity and
much of this diversity is represented by stygobitic spe¬
cies. In a checklist of the Australian inland aquatic
amphipods, Bradbury and Williams (1999) reported 74
species belonging to nine different families, of which 29
species in five families were reported from subterranean
waters. Ten years later the number of stygobitic amphi¬
pod species has significantly increased, particularly
because of systematic surveys in the Pilbara and Yilgarn
region of Western Australia and various regions in
South Australia. The data provided here are based on
taxonomic descriptions and on (partly unpublished)
molecular assessments. At present, there is evidence for
more than 144 stygobitic amphipod species: more than
37 species (8 genera) in the Paramelitidae; more than 24
species (9 genera of which 3 are undescribed) in the
Neoniphargidae; more than 51 species (several unde¬
scribed genera) in the Chiltoniidae; more than 30 species
(4 genera) in the Melitidae; and 4 species (2 genera) in
the Bogidiellidae. An interesting pattern of amphipod
distribution and diversity is emerging. The majority of
the species have distributions that are restricted to indi¬
vidual aquifers or tributaries, as is expected for taxa that
are confined to groundwater and that have low dispersal
abilities. Much of the diversity is occurring in ground-
waters of the arid region. Whereas some families appear
to be restricted to the temperate southeast and southwest
of the continent (Neoniphargidae), others are much
more widespread and encompass parts of the arid zone
(Chiltoniidae, Melitidae, Paramelitidae) and tropical
areas (Melitidae, Paramelitidae). Some families are more
restricted, such as the Bogidiellidae, which are known
only from coastal (anchihaline) habitats in the Cape
Range and Barrow Island of Western Australian and
Eyre Peninsula of South Australia. Notably, while taxa
in southern areas comprise both stygobitic and epigean
species, northern taxa, in the arid tropics and subtropics,
comprise only stygobitic taxa.
Speciation and adaptation to subterranean habitats
probably involved marine transgressions in the context of
the Melitidae and Bogidiellidae because species in these
families are all found near present or historical shorelines.
The evolution of the stygobitic species in the Paramelitidae,
Neoniphargidae, and Chiltoniidae, which occur predomi¬
nantly in the areas that were not inundated by the high sea
level stand in the Cretaceous, probably was driven by ari-
dification, forcing epigean freshwater species to seek refuge
in subterranean habitats.
to contain the entire change. But, in the shallow caves
common in the Australian tropics, such changes are
likely to cause large areas of cave systems to dry out.
Such processes may lead to the extinction of certain
cave fauna, or impede movement through the epikarst
and thus could promote speciation between different
karst areas. The high diversity of Schizomida in the
Pilbara and Cape Range, and of the paradoxosomatid
millipede Stygiocliiropus in arid Cape Range, are candi¬
dates for such analysis.
Humid caves within the arid zone have permitted
the survival of a diverse troglobitic fauna in arid Cape
Range, the affinities of which lie with the inhabitants
on the floor of the rainforest, both temperate and tropi¬
cal, habitats now thousands of kilometers distant.
While the fauna is now geographically relict, the driv¬
ing force resulting in the initial invasion of the caves is
unknown—species may have established in caves coin¬
cident with the onset of aridity to escape the surface
drying, or they may have established in caves seeking
resources unrelated to the onset of aridity. This ques¬
tion cannot be resolved for Cape Range because the
aridity has been sufficiently intense to extirpate
entirely close relatives at the surface. Other tropical
areas, such as North Queensland, offer greater pros¬
pect of resolving such issues because contemporary
lineages occur with surface and cavernicolous species
exhibiting various degrees of troglomorphy. However,
the recent finding of diverse troglobiont communities
also throughout the arid western shield supports the
hypothesis that aridity has driven the fauna to seek
more humid habitat underground and where it has
evolved troglomorphic traits, because the lineages to
which they belong are not commonly subterranean in
more humid regions with suitable subterranean voids.
Another area where a resolution of the causes of col¬
onization of the hypogean environment may be
resolved is in the groundwater calcrete deposits (see
Box 2) in the arid zone. There, many different lineages
of diving beetles (Dytiscidae) have invaded the ground-
water and become eyeless and flightless (Fig. 3). Each
calcrete body has a unique dytiscid assemblage, there is
no overlap in species between different calcretes, and
speciation appeared to have taken place in situ because
multiple species (1 to 4 of different sizes in many cases
being sister species) are common among the 100 stygal
ENCYCLOPEDIA OF CAVES
208
BOX 2
GROUNDWATER CALCRETES
The long period of emergence and the ensuing ero¬
sion down to the Archaean basement has resulted in
classical karst terrain being absent from the western
shield. However, thin carbonate deposits are widespread
throughout the arid zone and are well developed as
groundwater (valley) calcretes (hereafter termed calcrete)
which occur widely in Australia but in isolated, though
sometimes extensive, pockets usually associated with
paleodrainage lines (Fig. 5). Calcretes are carbonate
deposits forming from groundwater near the water
table in arid lands as a result of concentration processes
by near-surface evaporation. They form immediately
upstream of salt lakes (playas), chains of which form a
prominent part of the landscape in the more arid parts
of Australia. The playas, which reflect the groundwater
base level, are the surface manifestation of paleodrainage
channels incised into Precambrian basement rocks by
rivers that largely stopped flowing when the climate
changed from humid to arid in the Paleocene. Hence, the
paleovalleys predate the fragmentation of Gondwana.
Calcretes are especially important in the Australian
context as they form in arid climates (annual rainfall
<200 mm) with high potential evaporation (>3000 mm
per year). Although quite thin (10—20 m thick) the
groundwater calcretes often develop typical karst fea¬
tures and within them easy movement of groundwater
reflecting open conduits. Groundwater salinity may vary
markedly owing to the episodic recharge characteristic
of the arid zone (Humphreys 1999, 2001) and the calcrete
aquifers have been described as subterranean estuaries
(Humphreys et al., 2009).
Because they are deposited at intervals from the
groundwater flow, the calcrete masses are separated by
habitat that is unsuitable for stygofauna, namely,
Tertiary valley-fills, largely clays, and salt lakes.
Consequently, they form isolated karst areas along the
numerous major paleovalleys, some of which may date
from the Permian. The sediments filling the paleochan-
nels are mostly Eocene or later but the age of the cal¬
cretes is poorly defined. The extensive alluvial fan
calcretes and some of the river valley calcretes formed in
the Oligocene (37—30 Ma BP) may have followed the
onset of the continental aridity. Many of the calcrete
areas, especially those north of 31 °S, are being actively
deposited and the others have probably been remobi¬
lized and redeposited, attributes that make the dating of
calcrete deposits using standard radiometric methods
problematic. However, molecular phylogenies of the
diverse diving beetle, oniscidean isopod, parabathynel-
lid, and amphipod faunas, the numerous species of
which are each restricted to a single major calcrete area
or paleodrainage tributary (e.g.. Leys et al., 2003; Finston
et al., 2007; Guzik et al., 2009), indicate that the calcretes
have been present for at least 5—8 million years.
species in the arid zone. Molecular studies suggest that
numerous lineages invaded the calcrete aquifers during
the constrained time period (8—4 Ma BP), which sug¬
gests that it occurred in response to a widespread factor,
such as might be expected from spreading aridity.
Within the third millennium there has been a recog¬
nition that nonkarst and nonalluvial substrates, such as
pisolites and fractured rocks, support extremely diverse
subterranean faunas in both aquatic and terrestrial
habitats, particularly in the western shield region of
Western Australia. This recognition has resulted from
the requirement in Western Australia to include subter¬
ranean fauna as a component of environmental assess¬
ment for major mining projects in this mineraliferous
region. At a local scale this troglobiont diversity is com¬
parable to that found in the richest karst areas, such as
Bayliss Cave in Queensland. Regionally, consultancy
companies report between 150 and 300 troglobiont spe¬
cies from their projects within the 220,000 km 2 Pilbara
region, about the same size as the massive Nullarbor
karst of southern Australia where 27 troglobiont spe¬
cies are known. The proportion of species in common
to the consultancy companies is currently unknown but
is likely to be low, as many troglobiont species occupy
small geographic areas.
Energy Supply
Energy enters subterranean systems largely medi¬
ated by water, but also by animals and plant roots.
Because these elements themselves are not uniformly
distributed across Australia they have the potential to
influence Australian cave biogeography. The carriage
of organic matter in surface water is strongly affected
by seasonal rainfall and plant growth. The episodic
rainfall, characteristic of the arid zone, means that
some areas potentially have unpredictable energy sup¬
plies. Evaporation greatly exceeds rainfall in arid
zones and groundwater recharge—which carries
ENCYCLOPEDIA OF CAVES
DIVERSITY PATTERNS IN AUSTRALIA
209
dissolved organic matter, the predominant energy sup¬
ply in subterranean waters—occurs only following epi¬
sodic large rainfalls, sometimes years apart.
Plants provide the raw material that is transported
by water into the subterranean realm, but they also
directly transport energy into hypogean habitats by
means of sap transport within the roots and by root
growth to depths of up to 76 m. Roots, especially tree
roots, are an important and reliable source of energy
for troglobitic cixiid and meenoplid fulgoroid
Homoptera. These occur in the lava tubes of tropical
North Queensland and similar fauna are found in
karst across the tropics, into the Kimberley and
down the arid west coast, to the south of Cape
Range. They are also found on roots in calcrete, piso¬
lite, and fractured rock substrates on the western
shield. Tree roots are also utilized by cockroaches
throughout the country (e.g., in the Nullarbor,
Trogloblatella nullarborensis).
Tree root mats also represent a reliable food supply
for elements of the rich communities of aquatic inverte¬
brates, including some exhibiting troglomorphisms,
occurring in some shallow stream caves of Western
Australia. They provide habitat, and probably food, for
stygofauna in the Nullarbor, in calcrete aquifers of the
western shield, and in anchihaline caves in Cape Range
and Christmas Island where they are associated with a
diverse fauna largely comprising crustaceans.
Roots, like guano, often provide copious quantities
of energy to cave communities, which may be quite
diverse. Roots in the Undara lava tube, Queensland,
and the Tamala Limestone of Western Australia, sup¬
port diverse cave communities. However, whereas the
former contain numerous highly troglomorphic spe¬
cies, the latter has few stygomorphic species, many
being indistinguishable for surface species. Superficial
subterranean systems are often supported by roots but
to what extent is unknown, as, for example, the cave
and groundwater faunas of groundwater calcretes in
the arid zone.
Animals may transport energy into cave systems
and deposit it there as excreta, exuvia, carcasses, and
eggs. In Australia such troloxenic agents exhibit
marked latitudinal differences. In the south, rhaphido-
phorid crickets are the most conspicuous trogloxenic
agents, whereas bats, while not diverse, are locally
abundant where they form breeding colonies. In the
tropics bats are widespread, diverse, and important
producers of guano, as, to a lesser degree, are swiftlets
in more humid areas.
Guano is usually intermittently distributed in both
space and time because it is dependent on the sea¬
sonal biology of the birds and bats. In consequence,
the cave communities associated with guano are
highly specialized and differ markedly from the cave
fauna not dependent on guano. Markedly troglo¬
morphic species are not commonly found in the
energy-rich, but temporally unstable, guano commu¬
nities. However, immense populations of mites
(>100 cm 2 ) are seasonally present and numerous,
mostly rare, species (50 +) may occur, occupying dis¬
tinct parts of the cave system.
STYGOFAUNA AND CRUSTACEA
Stygofauna are discussed in the context of
Crustacea, which comprise the overwhelming majority
of stygofauna, but the Dytiscidae example above intro¬
duced the insect component.
The magnitude of the biodiversity present in subter¬
ranean waters globally has only recently been given
prominence. Australia, especially the northwestern
and southeastern parts, has unexpectedly come to
the attention of stygobiologists and systematists on
account of its diverse regional groundwater fauna
(stygofauna). Recently, these have been determined to
include a number of higher order taxa variously
new to science (i.e., the flabelliferan isopod family
Tainisopidae from Kimberley and Pilbara), new to the
Southern Hemisphere (Thermosbaenacea, Remipedia,
Epacteriscidae), or new to Australia (Spelaeogriphacea,
Pseudocyclopiidae). Many of these taxa occur near
coastal and anchihaline waters and are widely inter¬
preted as comprising a relictual tethyan fauna,
although molecular phylogeographic support has yet
to be demonstrated. Several of these lineages have con¬
generic species, which are known elsewhere only from
subterranean waters on either side of the North
Atlantic—the northern Caribbean region and the
Balearic and Canary archipelagos (see Box 3).
Syncarida
The Syncarida are crustaceans now entirely of
inland waters. The Anaspidacea are confined to
Australia, New Zealand, and southern South America.
In southeastern Australia they are often large and
mostly surface living, although several stygomorphic
species occur in cave streams and groundwater, and
an undescribed family has been reported that is
restricted to caves. In contrast, both families of
Bathynellacea have a global distribution, often even at
the generic level, and are widespread in Australia.
Bathynella (Bathynellidae) is found from Victoria to
the Kimberley and elsewhere the genus occurs glob¬
ally. Genera within the Parabathynellidae known
from Australia exhibit different regional affinities.
Chilibathynella and Atopobathynella are known from
ENCYCLOPEDIA OF CAVES
210
BOX 3
ANCHIHALINE HABITATS —TETHYAN RELICTS?
Anchihaline (or anchialine) habitats comprise near¬
coastal mixohaline waters, usually with little or no expo¬
sure to open air and always with more or less extensive
subterranean connections to the sea. They typically show
salinity stratification and may usefully be considered
groundwater estuaries. They typically occur in volcanic
or limestone bedrock and show noticeable marine as
well as terrestrial influences. The water column is per¬
manently stratified with a sharp thermohalocline sepa¬
rating a surface layer of fresh or brackish water from a
warmer marine, oligoxic water mass occupying the
deeper reaches. They have a significant amount of
autochthonous primary production, via a sulfide-based
chemoautotrophic bacterial flora, as well as receiving
advected organic matter from adjacent marine or terres¬
trial epigean ecosystems. Anchihaline habitats are
mostly found in arid coastal areas and are circumglob-
ally distributed in tropical/subtropical latitudes.
Anchihaline habitats support specialized subterra¬
nean fauna (Fig. 6), predominantly crustaceans repre¬
senting biogeographic and/or phylogenetic relicts. These
specialized anchihaline endemics are largely restricted to
the oligoxic reaches of the water column below the ther¬
mohalocline. The structure of these assemblages is
predictable, and, remarkably, however remote an anchi¬
haline habitat, this predictability frequently extends to
the generic composition.
In continental Australia, anchihaline systems occur
adjacent to the North West Shelf (Cape Range and
Barrow Island) and on Christmas Island (Indian Ocean),
an isolated seamount 360 km south of Java but separated
from it by the Java Trench.
Cape Range supports a fauna comprising atyids, ther-
mosbaenaceans, hadziid amphipods, cirolanid isopods,
remipeds, thaumatocypridid ostracods ( Danielopolina ),
and an array of copepods such as epacteriscid and pseu-
docyclopiid calanoids, and speleophriid misophrioids.
Some are the only known representatives of higher taxa
in the Southern Hemisphere (Class Remipedia; Orders
Thermosbaenacea, Misophrioida), and several genera are
known elsewhere from anchihaline systems on either
side of the North Atlantic ( Lasionectes , Halosbaena,
Speleophria, Danielopolina (Fig. 6)) and the atyid
Stygiocaris (Fig. 6) is closely related to the amphi-Atlantic
genus Typhlatya. These obligate stygal lineages are
thought to have poor capability or opportunities for dis¬
persal, and this attribute, combined with their distribu¬
tions which closely match areas covered by the sea in
the late Mesozoic, suggest that their present distribu¬
tions could have resulted by vicarianace as a result of
the movement of tectonic plates (Fig. 2).
Anchihaline systems on oceanic islands support a
different group of fauna, but the structure of these
assemblages is similarly predictable, even between
oceans. Christmas Island is a seamount and supports an
anchihaline fauna characterized by the stygobitic
shrimp Procaris (Decapoda), which belongs to the primi¬
tive, highly aberrant family, Procarididae which appears
globally to be restricted to anchihaline caves. This fam¬
ily has been reported elsewhere only from other isolated
seamounts, namely, Bermuda and Ascension Island in
the Atlantic Ocean, and Hawaii in the Pacific. In each
case, as with Christmas Island, the procaridids are asso¬
ciated with alpheid, hippolytid, and atyid shrimp.
These co-occurrences of two primitive and presumably
ancient caridean families support the contention that
crevicular habitats have served as faunal refuges for
long periods of time. Recently, an endemic species of
Danielopolina has been described from Christmas Island,
the first juxtaposition of a member of the anchihaline
faunas characteristic of epicontinental and seamount
island anchihaline faunas. There is no coherent theory
as to their distribution to remote seamounts such as
Christmas Island.
Chile and southeastern Australia, while tire latter is
also found in India and throughout northwestern
Australia, including Barrow Island and Cape Range,
and the paleodrainage channels of the arid center.
Notobathynella is found across Australia and New
Zealand, while Hexabathynella, from the eastern
Australian seaboard, has a more global distribution,
being found in New Zealand, southern Europe,
Madagascar, and South America. Endemic genera
include Kimberleybathynella from the Kimberley,
Brevisomabathynella and Billibathynellla from the western
shield, and Octobathynella from New South Wales. As
30 described species are endemic to single calcrete
masses the overall diversity is expected to be very high.
Bathynellacea are small stygobites, mostly inhabitants
of interstitial freshwater environments, although an
undescribed genus of large, free-swimming para-
bathynellid occurs in water close to marine salinity
(>30 000 mg L 1 TDS) in the Carey paleodrainage sys¬
tems of the arid zone, where it is associated with a
ENCYCLOPEDIA OF CAVES
DIVERSITY PATTERNS IN AUSTRALIA
211
number of maritime copepod lineages such as Ameiridae
(Harpacticoida) and Halicyclops (Cyclopoida).
Copepoda
Remarkably little work had been conducted on non¬
marine copepods in Australia until this millennium.
Recent work on groundwater copepods, largely from
groundwater calcretes of the western shield, and the
near coastal, especially anchihaline systems of the
northwest, has revealed higher taxa not previously
described from Australia, in some cases even from the
Southern Hemisphere.
Numerous species of copepods are being described
from Australian groundwaters, largely from the western
shield, including ten new genera of Cyclopoida and
Harpacticoida, and several genera are reported for the
first time from Australia [Nitocrella Ameiridae (Eurasia),
Parapseudoleptomesochra Ameiridae (global), Nitocrellopsis
Ameiridae (Mediterranean) Haifameira Ameiridae (depth
of Mediterranean Sea), Pseudectinosoma Ectinosomatidae
(Europe) and the family Parastenocarididae (Pangaea,
freshwater)]. The broader distribution of these lineages
within Australia awaits investigation but some of them
have been found in tropical Queensland.
The occurrence of about 17 near-marine lineages
(e.g., Halicyclops) in the center of the western shield
alongside lineages considered to be ancient freshwater
lineages ( Parastenocaris: Parastenocarididae) is notable.
It may reflect both the salinity stratified, often hypersa¬
line groundwater in the more southerly paleodrainage
systems, as well as the ancient origins of the fauna.
Interestingly, Halicyclops is almost completely absent
from marine interstitial in Australia, although it is the
world's most speciose element in this habitat. It was
probably replaced here by the genus Neocyclops, which
is very diverse, testifying to an ancient invasion of
inland waters by Halicyclops. Mesocyclops has a mostly
tropical distribution; Metacylcops (trispinosus group) and
Goniocyclops have an Eastern Gondwanan distribution;
and the limits to the distributions of newly described
genera of Ameiridae, Canthocamptidae, and Cyclopinae
await confirmation.
Ostracoda
Ostracods recorded from Australian inland waters
are mainly from the families Limnocytheridae,
Ilyocypridae, and Cyprididae. In the Murchison, ostra¬
cods from the families Candonidae, Cyprididae, and
Limnocytheridae have been recorded in open ground-
water but stygophilic species occur only in the
Limnocytheridae and Candonidae, the latter including
the globally widespread genus Candonopsis (subfamily
Candoninae), which occurs widely and in a wide vari¬
ety and age of substrates. Species are known from
Pleistocene syngenetic dune karst (Tamala Limestone),
several species from Tertiary (probably Miocene)
groundwater calcretes on the western shield, and from
the Kimberley (Devonian Reef Limestone). In Europe
there are only a few, mostly hypogean species that are
considered to be Tertiary relicts with surface relatives
today occurring in tropical and subtropical surface
waters; they are especially diverse in Africa. The sub¬
family Candoninae (family) Candonidae are common
elements of stygofauna globally but recent finds from
the Pilbara describe about 25% of the world's genera
but these are more closely related to the South
American and African Candoninae than to those of
Europe.
The thaumatocypridid genus Danielopolina, previ¬
ously unreported in the Southern Hemisphere, occurs
as a tethyan element in the anchihaline system at Cape
Range and on Christmas Island, Indian Ocean, an iso¬
lated seamount island where it occurs with the first
living representative of Microceratina (Cytheruridae).
Fossils in marine cave facies in the Czech Republic
suggest that the former lineage already inhabited
marine caves in the Jurassic.
In the five years to 2010 there was major progress in
the describing and understanding of the subterranean
ostracods from Australia. One of the hotspots of their
diversity is the Pilbara region from where 86 species
have been described belonging to 12 genera in the sub¬
family Candoninae. Only one of the species and one
genus is not endemic to this region, while all the other
species have a very limited distribution, comprising
short-range endemics. The tribe Humphreyscandonini
was described to accommodate eight Candoninae gen¬
era endemic to the Pilbara. This tribe shows some close
similarities with the fossil. Tertiary fauna of the
Northern Hemisphere. Other genera from this region
have a clear Gondwanan connection, and even the
genus Candonopsis, of the tribe Candonopsini, is here
represented by a distinct lineage, and a separate sub¬
genus. This subgenus is the only taxon connecting the
Pilbara's candonins with those from the Kimberley
and Murchison regions. Candoninae fauna from the
Murchison is less rich, but also comprises only
endemic species. Even though it has been quite thor¬
oughly explored, the Murchison Candoninae com¬
prises only five described species. East Australia also
has a unique Candoninae fauna, but only one mono-
typic genus have been described from Queensland,
while many more await description. The east
Australian Candoninae belongs to the tribe
Candonopsini and they are more closely connected
with the South and Central American Candoninae
than to the rest of Australia. Few taxa other than
ENCYCLOPEDIA OF CAVES
212
D
Candoninae have been described from Australian sub¬
terranean waters, and most belong to the genus
Gomphodella (family Limnocytheridae) and are endemic
to the Pilbara and Murchison regions, and represent a
very old lineage of freshwater ostracods. Most of the
species are known from fossils, and some close rela¬
tives today live in the subterranean waters of the
Balkan Peninsula, and ancient African Lakes. One spe¬
cies of the family Cyprididae has been described from
the Kimberley region, but, as all the other Cyprididae
species found in the wells and occasionally bores
around Australia, probably represents only a stygo-
phile, rather than a true stygobiont.
The western shield is considered to have been a sin¬
gle landmass continuously emergent from the sea since
at least the Paleozoic. If this is the case then there is no
clear hypothesis to account for the extraordinary dis¬
parity (Table 1) in the gross composition of the subter¬
ranean fauna between contiguous parts of the western
shield, loosely termed the Pilbara (to the north) and
the Yilgarn (to the south) and the associated orogens.
Isopoda
Phreatoicidean isopods (Fig. 3) have a Gondwanan
distribution and occur widely across southern
Australia (and in tropical Arnhem Land and the
Kimberley) in surface habitats that have permanent
water—surface expressions of groundwater—usually as
cryptic epigean species. Their distribution is strongly
associated with the areas of the continent not sub¬
merged by Cretaceous seas (see also Fig. 4). About 76
species in 30 genera are described from Australia, of
which 10 species in 8 genera are hypogean (cavernico-
lous or spring emergents). They are under active
TABLE 1 The Distribution of the better known higher taxa of
subterranean invertebrates between the adjacent Pilbara and
Yilgarn regions on the Australian Western Shield that has been
continuously emergent since the Paleozoic
Taxon
Pilbara
Yilgarn
% in
Common
Candonine ostracod genera
13
1
8
Candonine ostracod species
58
5
0
Copepoda genera
43
30
4
Copepoda species
25
15
21
Spelaeogriphacea species
2
0
0
Tainisopidae species
5
0
0
Schizomida species
26
0
0
Dytiscidae species
0
89
0
Nocticolidae species
9
0
0
revision and numerous taxa are being described. Five
hypogean species occur on the Precambrian western
shield. The family Hypsimetopidae is represented by
the genera Pilbarophreatoicus in the Pilbara and
Hyperoedesipus in the Yilgarn regions (separate cratons
of the western shield). These genera are closely related
to the hypogean genus Nichollsia found in the Ganges
Valley of India and a new genus from caves of Andrah
Pradesh (East-Central India). These occurrences suggest
that Hypsimetopidae were hypogean prior to the sepa¬
ration of Greater Northern India from the western
shore of Australia ( ca. 130 My BP). Crenisopus, a stygo-
bitic genus occurring in a sandstone aquifer in the
Kimberley and another new species on Koolan Island
off Western Australia, is a link between African and
Australasian lineages of phreatoicideans. The genus is
basal to most families in the Phreatoicidea, suggesting
divergence after they entered freshwater but prior
to the fragmentation of East Gondwana during the
Mesozoic era.
The family Cirolanidae with three species and two
genera occurs on Cape Range, Barrow Island, and in the
Pilbara of Western Australia. The family Tainisopidae,
endemic to northwestern Australia, occurs in the exposed
and greatly fragmented Devonian Reef system through¬
out the western Kimberley as well as in remote outcrops
of this fossil reef in northeastern Kimberley. A second
clade of this family (Fig. 3) with five described species
inhabits groundwater calcretes in the Pilbara, from which
the Kimberley was separated by the Cretaceous marine
incursions. The location and distribution of this family is
indicative of ancient marine origins. Recent cladistic anal¬
ysis suggests that this family is related to the cosmopoli¬
tan marine Limnoriidae but at a basal level, suggesting
FIGURE 4 The distribution of subterranean amphipod families in
relation to long emergent land areas (deeper shading) in Australia.
Map by Remko Leys.
ENCYCLOPEDIA OF CAVES
DIVERSITY PATTERNS IN AUSTRALIA
213
the family is older than the more derived families like
the Cirolanidae, which have an extensive Cretaceous fos¬
sil record. Among Asellota, the Janiridae occur widely
across southern Australia and Tasmania and the genus
Heterias, which also occurs in New Zealand, is likely to
be highly speciose. Protojaniridae occur in the Northern
Territory and this family has a Gondwanan distribution
similar to the Phreatoicidea.
Terrestrial isopods (Oniscidea) are a prominent com¬
ponent of cave fauna throughout Australia, as elsewhere
in the world, yet there are few described highly troglo-
morphic Oniscidea (Fig. 3) and these are placed in the
genera Abebaioscia, Andricophiloscia, and Haloniscus
(Philosciidae), and Troglarmadillo (Armadillidae) that
occur widely on the western shield. Numerous other
troglobitic species are recognized belonging to the
genera Styloniscus (Stylonisciidae), an undescribed
genus (Stenoniscidae), Laevophiloscia and Haloniscus
(Philosciidae), Hanoniscus (Oniscidae), and Buddelundia
and Troglarmadillo (Armadillidae). Their distribution
FIGURE 5 The distribution of groundwater calcretes in Western
Australia, which occur throughout the arid land north of 29°S. Most
occur immediately upstream of salt lakes (playas) within paleo-
drainage channels (dotted lines).
seems to reflect the general aridity that developed fol¬
lowing the separation of Australia from Antarctica,
rather than to suggest more ancient relictual distribu¬
tions. So, in the more humid southern regions
Styloniscidae are a prominent component of cave fauna,
as they are in the wet forest of the surface, but they are
troglophilic. Armadillidae, Ligiidae, and Scyphacidae
are also common in Tasmanian caves but none is troglo-
morphic. On the mainland, Olibrinidae, Philosciidae,
and Armadillidae are prominent among subterranean
fauna. In the drier areas of Australia, where armadillids
are such a prominent part of the surface fauna, they
appear in caves more frequently, and many have overt
troglomorphies. These troglobites are known from the
Nullarbor, North Queensland (Chillagoe), Cape Range,
and Kimberley and the western shield. The troglobitic
Philosciidae, Platyarthridae, and Oniscidae from
Western Australia are undescribed.
A single epigean species of Haloniscus searlei, an
aquatic oniscidean isopod, is known from salt lakes
(playas) across southern Australia. Numerous stygobi-
tic species of Haloniscus occur in groundwater calcrete
deposits of the Yilgarn region of the western shield
sometimes in saline waters, and from Ngalia Basin
northwest of Alice Springs in Central Australia.
Spelaeogriphacea
In Australia, this order of stygal crustaceans is
known from two species, each in a separate lacustrine
calcrete deposit in the Fortescue Valley, a major ancient
paleovalley of the northern Pilbara region. The four
extant species of spelaeogriphaceans occur with very
circumscribed distributions in subterranean freshwater
habitats of Africa (Table Mountain, South Africa),
South America (western Mato Grosso, Brazil), and
Australia, all fragments of Gondwana. The supposition
of a Gondwanan origin is possibly refuted by a marine
fossil, Acadiocaris novascotia, from shallow marine sedi¬
ment of Carboniferous age in Canada, which has been
attributed to this order. All living spelaeogriphaceans
occur in or above geological contexts that are earliest
Cretaceous or older. The colonization of Gondwanan
freshwater is likely to have occurred after the retreat of
the Gondwanan ice sheet (after 320 Ma BP) and prior
to the dissolution of Gondwana (142—127 Ma BP).
Decapoda
Atyid shrimps (Decapoda) are widespread in surface
waters throughout the tropics. They may have colo¬
nized Australia from Asia via the Indonesian archipel¬
ago, but their presence as stygobitic species in caves
and groundwaters of the Canning Basin suggest a more
ENCYCLOPEDIA OF CAVES
214
D
FIGURE 6 Stygal animals from Australian anchihaline waters. (A) Milyeringa veritas (Eleotridae); (B) Halosbaenci tulki (order
Thermosbaenacea); (C) Stygicicaris stylifera (Decapoda: Atyidae); (D) Danielopolina kornickeri (Ostracoda: Thaumatocyprididae). Photographs (A)
and (B) by Douglas Elford; (C) and (D) by William Humphreys; both of the Western Australian Museum.
ancient origin. The genera Parisia, Pycnisia, Pycneus, and
Caridina, the first three being endemics, form a paraphy-
letic clade and occur as species with small ranges in
northern and central Australia. A second clade includes
Antecaridina and the endemic genus Stygiocaris (Fig. 6)
from anchihaline systems of an isolated seamount
(Christmas Island, Indian Ocean) and the epicontinental
system of the northwest, respectively. Stygiocaris is most
closely related to Typhlatya found in subterranean
waters on either side of the North Atlantic.
CHELICERATA
The chelicerates globally comprise a biodiverse com¬
ponent of cave communities and the class is represented
in the Australian cave fauna by the orders Acarina,
Palpigradi, Amblypygi, Opiliones, Pseudoscorpiones,
Schizomida, Scorpiones, and Araneae.
Acarina
The mite family Pediculochelidae (Acariformes)
was first recorded in Australia from a dry cave in
Cape Range where a specimen was attributed to
Paralycus lavoipierrei that is described from California.
Tiramideopsis (Mideopsidae) occurs in the Millstream
aquifer, a genus previously known from similar habi¬
tats of India and suggesting ancient links ( cf.
Phreatoicidea). Generally, the poorly known mesostig-
matid mite fauna of Australian caves does not appear
to constitute a distinctive cave fauna or exhibit any
of the morphological characteristics of deep-cave
arthropods.
Palpigradi
The first native species, Eukoenenia guzikae, was
described in 2008, collected from a borehole in calcrete
substrate in the southern western shield. Since then
other species have been recovered from the Pilbara,
Barrow Island, and Kimberley of Western Australia, all
from boreholes in karst and fractured rock substrates,
or caves.
Amblypygi
Troglophilic species of Charon are found on
Christmas Island and in the Northern Territory.
ENCYCLOPEDIA OF CAVES
DIVERSITY PATTERNS IN AUSTRALIA
215
Opiliones
Cavernicolous species of Triaenonychidae in
Tasmania and New South Wales often show depig¬
mentation, attenuation of pedipalps and legs, a reduc¬
tion in size (but not loss of) eyes, reduced
sclerotization, and other troglomorphic features. The
cave fauna of Tasmania, unlike continental Australia,
have distribution patterns more like those of other
periglacial areas of the world in which profound envi¬
ronmental changes were associated with Cainozoic gla¬
cial cycles. The distribution of the opilionid genus
Hickmanoxyomma, which is exclusively cavernicolous,
appears to have resulted from the ablation of surface
forms in the south and east of Tasmania, where the
effect of glaciation was most intense, and the occur¬
rence of some sympatric species suggests that there
may have been multiple phases of cave invasion. In
contrast, in the coastal lowlands to the north and north¬
east, where periglacial conditions were less extreme,
surface-dwelling species of Hickmanoxyomma are pres¬
ent. A cavernicolous assamiid with reduced eyes, but
not strongly troglomorphic, and the strongly troglo¬
morphic Glennhuntia glennhunti ("Phalangodidae")
from arid Cape Range, are both probably rainforest
relicts although the wider affinities of both families are
unknown. Several species of unknown affinity have
recently been collected from boreholes in the arid
Pilbara region.
Pseudoscorpiones
The worldwide family Chthoniidae is most commonly
represented among troglomorphic species in Australia.
The genera Tyrannochthonius, Pseudotyrannochthonins, and
Austroclithonius are widespread with cave populations in
eastern and western parts of the continent. Syarinidae,
which occur in the rainforests of Africa, Asia, and the
Americas, occur widely in Australia and as a troglophile
in Cape Range and the Pilbara. The Hyidae, known from
India, Madagascar, and southeast Australia, are repre¬
sented in Australia in the Kimberley and by markedly
troglomorphic species of Indohya from arid ecosystems.
Schizomida
Schizomids are essentially a tropical forest element
that occurs across the top of the continent, as far south
as the humid caves in the arid Cape Range and subter¬
ranean systems throughout the Pilbara. Troglobitic
species are known from various genera, of which the
most troglomorphic belong to the genera Dracidoides
(Fig. 3) and Paradracidoides.
Scorpiones
A troglobitic scorpion ( Liocheles : Liochelidae) occurs
on rainforest-covered Christmas Island, Indian Ocean.
Aops oncodactylns, a genus endemic to arid Western
Australia, is the first troglobitic urodacid and the
first troglobitic scorpion recorded from continental
Australia. Only one other troglobitic scorpion is known
outside the Americas, from Sarawak (Malaysia).
Araneae
Troglodiplura, which has South American affinities,
is the only troglobitic mygalomorph spider in
Australia, and occurs in caves in the arid Nullarbor
region. Like the sympatric cockroach Trogloblatella, it is
heavily sclerotized, suggesting a more drying atmo¬
sphere than is generally associated with troglomorphic
animals found elsewhere, such as in the arid zone
caves of Cape Range, which have affinities with rain¬
forest floor communities. The primitive araneomorph
(true) spider Hickmania troglodytes from Tasmania is a
troglophilic member of Austrochilidae, a family that
also occurs in Chile and Argentina.
Large lycosoid spiders occur widely in the arid
areas, one of which, Bengalla bertmaini (Tengellidae),
from Cape Range, is highly troglomorphic, lacking
eyes and pigment.
Symphytognathidae occur as troglobitic elements in
the tropical caves of arid Cape Range and monsoonal
Northern Territory (Katherine) as Anapistula, found as
epigean elements in the wet tropics of Australia,
Malaysia, and Indonesia.
Cavernicolous Filistatidae occur throughout Australia
(Wandella ) but the monotypic Yardiella from Cape Range
has relatives in northeast India and the family has a gen¬
erally Gondwanan distribution.
A blind troglobitic Tetrablemma (Tetrablemmidae)
occurs in the Pilbara, presumably a relict from past
humid environments. Likewise Desognanops (Tro-
chanteriidae), from the Pilbara, is a blind rainforest
relict.
Among the Pholcidae, Wugigarra occurs along the
eastern seaboard and to the southeast of the continent
while the western three-quarters of the continent con¬
tains old elements of the pholcid fauna. If the distribu¬
tion of the genus were restricted by current ecological
conditions, then the genus would be expected to be
found in the west and other refugia, but this apparently
is not the case. This distribution may be due to the
marine subdivision of the continent by the Cretaceous
marine transgression. The genus Trichocyclus occurs as
a cavernicolous element throughout much of the rest of
the continent from the Nullarbor to the tropics. Both
genera are considered to be Gondwanan relicts.
ENCYCLOPEDIA OF CAVES
216
MYRIAPODA
Diplopoda
Troglobitic millipedes have rarely been reported in
Australia. The Cape Range karst hosts a number of
paradoxosomatid millipedes of the genus Stygiochiropus,
three of which are declared rare fauna being known
from single small caves, while S. communis comprises a
species swarm occurring through the range in three
regions each with a major genetic cluster. Caves
between regions differ by a mean of 25.6% fixed allelic
differences (allozymes), while caves within regions
show 0.92—3.06% fixed differences. The first recorded
troglobitic spirobolid millipede is Speleostropus nesiotes,
a genus endemic to Barrow Island.
Chilopoda
The scolopendrid centipede Cryptops ( Trigonocryptops)
roeplainsensis occurs along the coastal boundary of the
Nullarbor karst. Large scutigerid centipedes of the
genus Allothereura are seen in the more superficial parts
of both humid and dry caves in the arid areas but are
not troglobitic.
INSECTS
Diplura
Numerous undescribed species belonging to Campo-
deidae, Japygidae, Projapygidae, and Parajapygidae
have been collected from limestone caves and boreholes
in karst and fractured rock aquifers.
Zygentoma
Among the Nicoletiidae, Trinemura is represented in
caves in the west, while Metrinura is found in the caves
of the northeast. Numerous species of Nicoletiidae and
Ateluridae have recently been collected from boreholes
in the western shield, especially the Pilbara region.
Collembola
The composition of collembolan fauna changes
between the south and north of Australia. Caves in the
south of the continent contain up to five genera of
troglobitic Collembola, while those in tropical areas
have only two genera. The genera Adelphoderia and
Arrhopalites are not recorded as troglobites in tropical
caves, but because the former is known from both tem¬
perate and tropical rainforest litter it seems likely to
occur in tropical caves. This apparent trend in
diversity may well reflect the greater sampling effort
in southeast Australia. Oncopodura occurs in southeast
Australia and in the Northern Hemisphere.
Planthoppers: Relicts or Invaders?
There is continuing debate as to whether cave fauna
result from active colonization or occur as relicts as a
result of the extirpation of surface populations by
adverse conditions (e.g., glaciation, aridity). The cave
fauna on arid Cape Range are clearly relictual in that
they are now remote from the humid forest from
which the fauna were sourced. However, the aridity is
sufficiently intense to have obliterated all close surface
relatives and so the process by which it became relic¬
tual cannot be resolved. By contrast, in a grossly simi¬
lar fauna in Far North Queensland, the troglobitic
cixiid and meenoplid planthoppers have some mem¬
bers with surface relatives and many intermediate
forms. These lineages show many reductive, but no
progressive trends, and this has been interpreted as
support for the active colonization of the subterranean
realm, rather than as a process of relictualization
(Hoch and Howarth, 1989).
In North Queensland seven evolutionary lines of
planthoppers (Fulgoroidea) of the families Cixiidae
(genera Solonaima, Undarana, Oliarus ) and Meenoplidae
(Phaconeura , a continent-wide genus) are found.
Solonaima (Cixiidae) exhibits four independent inva¬
sions of the caves and shows a full range of adapta¬
tions to cave life, from epigean to troglobitic, together
with intermediate stages. This lineage provides an
excellent model for the stepwise evolution of cave
forms and the reconstruction of the historic process of
cave adaptation—the loss of eyes and pigmentation,
reduction of wings and tegmina, and increased pheno¬
typic variation, such as wing venation, even within
same species, suggesting a relaxation of selection pres¬
sure. To Hoch and Howarth (1989) this suggested that
there had been fragmentation of the rainforest owing
to the drying climate during the Miocene. This model,
argued on other evidence, has also been suggested for
the arid Cape Range region on the west coast of the
continent.
Blattodea
Cockroaches represent a widespread and common
element of many Australian caves, particularly those
where the predominant energy source is guano from
bats or swiftlets where. Paratemnopteryx and related gen¬
era ( Gislenia , Shazvella) are prominent. Paratemnopteryx
stonei exhibits significant morphological variation in
seven tropical caves spread over a 150-km distance in
ENCYCLOPEDIA OF CAVES
DIVERSITY PATTERNS IN AUSTRALIA
217
North Queensland, such variation being consistent with
molecular variation (Slaney and Weinstein, 1996). The
genus Neotemnopteryx is widespread on the east coast
and is represented by 14 species, of which five species
are cavernicolous, but troglobitic species occur in the
Nullarbor and the southwest coast: N. wynnei and N.
douglasi, respectively. In the Nullarbor, where caves are
relatively dry, the large, eyeless but highly sclerotized
Trogloblatella nullarborensis is found. In contrast, the
Nocticolidae occur widely in the Old World tropics
and a number of cave species occur throughout the
Australian tropics, down to arid Cape Range where
Nocticola flabella is found, the world's most troglo-
morphic cockroach, which is distinguished by its pale,
fragile, translucent appearance. In contrast, a more
robust monotypic troglobite, Metanocticola, is found on
Christmas Island. Numerous species of troglobitic
Nocticola (and Blattidae) are being collected from baited
traps inserted down holes bored in pisolites and frac¬
tured rocks of the arid Pilbara region. The genus
Nocticola also occurs in the Philippines, Vietnam,
Ethiopia, South Africa, and Madagascar.
Orthoptera
Many cave crickets (Rhapdidophoridae), which
occur in cave and bush habitats across southern
Australia, are trogloxenes, like some bats. During the
day these moisture-loving insects tend to congregate in
relatively cool, moist, and still air to avoid desiccation.
In the evening, part of the cricket population moves
outside the cave entrance to feed but they return
underground before dawn and so transport organic
matter into the cave. Rhaphidophoridae have a dis¬
junct global distribution in the temperate zones of both
hemispheres. The Macropathinae are considered to be
the basal group and these have a circum-Antarctic dis¬
tribution, suggesting a Gondwanan origin. Generic
diversity is much greater in Australia and New
Zealand than elsewhere. Four genera are restricted to
Australian temperate zones and a further three genera
to Tasmania itself. The remaining three subfamilies
inhabit the Boreal zone, suggesting vicariance owing to
the Mesozoic dissolution of Pangaea.
In contrast, the only truly troglobitic cricket in
Australia is the pigmy cricket Ngamalanguia (Nemobii-
nae: Gryliiidae) (Fig. 3), a monotypic genus endemic to
Cape Range that lacks eyes, ocelli, tegmina, wings, and
auditory tympana; is pale; and has exceptionally long
antennae.
Coleoptera Dystiscidae
Globally, beetles are by far the most intensively
studied cave animals. Chief among them are the
trechine carabid beetles, of which more than 2000 spe¬
cies have been described. Of these, more than 1000
species are troglomorphic, inhabiting caves from peri-
glacial areas of Australia and New Zealand (25 spe¬
cies), eastern Palearctic (ca. 250 species), western
Palearctic ( ca. 600 species), and Nearctic and
Neotropical (ca. 200 species).
Unlike mainland Australia, Tasmanian caves sup¬
port a distinctive cave fauna of carabid beetles from
the tribes Trechini (a strongly hydrophilous group
forming a dominant element of cave fauna of the peri-
glacial areas of Europe, North America, New Zealand,
and Japan) and Zolini (confined to Australasia) each
containing two genera with troglobitic species. In the
periglacial areas of Tasmania, vicariant patterns simi¬
lar to those for opilionids may be deduced for the tre¬
chine and zoline carabid beetles, which form such
a prominent part of the Tasmanian cave fauna.
Harpalinae, a globally widespread and predominantly
phytophagous group, typical of dry country, are con¬
sidered unsuitable for cave colonization, and yet
many genera are represented in caves in Australia.
Two genera of the Calleidini occur in guano caves in
Australia, which suggests, because these beetles are
typically arboreal, the possibility of a reversal from
the arboreal habit typical of this tribe, to an edaphic or
subterranean life.
Although the Cholevidae is well represented in the
more humid parts of Australia, the tribe Leptodirinae
(Bathysciinae), which comprises the predominant com¬
ponent of the rich cholevid beetle fauna of the
Northern Hemisphere, is entirely missing from
Australia and the rest of the Southern Hemisphere. In
the Snowy Mountains area of the mainland, where
periglacial conditions also persisted, is found the only
troglomorphic psydrinid beetle known globally.
Numerous other families of beetles occur in caves
throughout Australia, in both the humid and arid
areas, but most seem to be accidentals. The Australian
troglobitic fauna, especially those that associate with
periglacial areas, differ from those in the Northern
Hemisphere, owing to the composition of the surface
fauna, rather than due to different evolutionary trends.
Recent collections from baited traps lowered into bores
drilled in karst, pisolites, calcretes, and fractured rock
have yielded a sparse but diverse troglomorphic
beetle fauna awaiting study including members of
the families Carabidae, Curculionidae, Staphylinidae
(Pselaphinae), and Tenebrionidae. In addition, Trogidae
are commonly associated with guano deposits derived
from cave roosting bats and swiftlets.
Australia hosts more species of stygobitic diving
beetles (Dytiscidae) than the rest of the world com¬
bined. More than 100 species of Hydroporinae (and
the only two stygobitic species of Copelatinae) are
ENCYCLOPEDIA OF CAVES
218
D
found mainly in shallow calcrete aquifers in the
Yilgarn, western shield, and in Ngalia Basin, Northern
Territory. Several species occur in alluvia of New
South Wales and South Australia.
VERTEBRATES
Caves in the wet-dry (monsoonal) tropics commonly
provide refuge to vertebrates during the dry season
and clearly this temporary habitation has an impact on
the trophic relations of these caves. Among them are
tree frogs (e.g., Litoria caerula), which are also abundant
in uncapped boreholes, and fish, such as the common
eel-tail catfish, Neosilurus hyrtlii, and the spangled
perch or grunter, Leiopotherapon unicolor. In the dry sea¬
son, the fish may survive in caves and underground
water systems and from there they would contribute
to the repopulation of the seasonally inundated
floodplains.
Australia has two described highly troglomorphic
fishes which are sympatric where they occur at
Cape Range. The blind gudgeon, Milyeringa veritas
(Eleotridae) (Fig. 6), is of unknown affinity but inhabits
water ranging from seawater to freshwater in a largely
anchihaline system in Cape Range (a second, unde¬
scribed species is known from Barrow Island). Swamp
eels (Synbranchidae) are represented in Australia by
two species of Ophisternon, of which O. candidum is a
highly troglomorphic species (Fig. 3). Stygobitic speci¬
mens of unknown affinity were recorded from Barrow
Island and the Pilbara during 2010. The genus occurs
widely in the coastal wetlands of the Indo-Malayan
region, with one other troglomorphic species inhabit¬
ing caves in Quintana Roo, Mexico.
Snakes are commonly seen in caves, especially in
the tropical regions where they predate bats (e.g., the
banded cat-snake Boiga fusca ornata). The blind snake,
Ramphotyphlops longissimus, from the Barrow Island
karst has apparent troglomorphies and may represent
the first troglobitic reptile.
Birds are rarely represented in Australian caves
other than as superficial components inhabiting cave
openings. The exceptions are swiftlets ( Collocalia spe¬
cies) that build their nests in the dark zone, on smooth
concave walls high above the cave floor in some tropi¬
cal caves in Far North Queensland and Christmas
Island (Indian Ocean). The nests of some species are
intensively harvested for the gourmet delicacy "birds'
nest soup" in Southeast Asia and India. The Christmas
Island glossy swiftlet ( Collocalia escidenta natalis) is
endemic to Christmas Island where, in the absence of
cave bats, they are the prime source of guano in caves.
A number of other species of Collocalia occur in the
Indian Ocean, Southeast Asia, and Queensland, mostly
nesting in caves. The nests detach from the cave walls
in dry air, a factor that may account for their absence
from the drier tropical areas, such as the Kimberley.
The various subspecies inhabit few of the caves avail¬
able, being known from only five caves on Christmas
Island, whereas the white-rumped swiftlet ( Collocalia
spodiopygus chillagoensis) occurs in less than 10% of
approximately 400 caves at Chillagoe in Queensland.
Bats comprise nearly a third of the Australian mam¬
malian fauna. Seven families of bats, comprising about
30% of the Australian bat fauna, are found in caves.
The 17 species of cave-dwelling bats in Australia are
largely restricted to the tropics and comprise insectivo¬
rous and frugivorous bats and vertebrate predators
(ghost bats. Macroderma gigas). Six species are restricted
to the Cape York peninsula and 11 species occur across
the northern part of the continent, two of them extend¬
ing along the west coast to the arid Pilbara region.
Only four species are restricted largely to the center of
the continent, two being restricted to the western
plains of Queensland and New South Wales.
CONCLUSIONS
In a global context, the most striking features of the
subterranean fauna of Australia are (1) the apparent
age of the lineages present in subterranean environ¬
ments, and (2) the high proportion of geographic
relicts present in the subterranean systems that are
widely separated from their near relatives. Although
much remains to be done to establish consistent pat¬
terns, numerous independent examples suggest similar
processes but at a range of spatial and temporal scales.
In the southeast, there is evidence that Pleistocene gla¬
ciation influenced the cave fauna. But, over most of
mainland Australia, the overwhelming influence seems
to have been relict distributions resulting from increasing
aridity during the Tertiary, particularly in the Miocene.
Numerous terrestrial and aquatic lineages have affinities
with Gondwana, or with Western Gondwana, often at
the generic level. In terrestrial lineages, these are com¬
monly associated with rainforests. Numerous crusta¬
ceans, often lineages entirely comprising stygal species,
and even a fish lineage, have distributions throughout
the area of the former Tethys Ocean. Many lineages
from northwestern Australian anchihaline waters com¬
prise species congeneric with those inhabiting caves on
either side of the North Atlantic.
Acknowledgments
I thank G.D.F. Wilson (non-oniscidean Isopoda), S. Taiti
(Oniscidea), R. Leijs (Amphipoda), T. Karanovic
ENCYCLOPEDIA OF CAVES
DIVERSITY PATTERNS IN EUROPE
219
(Copepoda), I. Karanovic (Ostracoda), M.S. Harvey
(Chelicerata), S. Halse, and S. Eberhard for their com¬
ments and/or unpublished information, and K. Grimes
for the basis of Figure 1 .
See Also the Following Articles
Diversity Patterns in the Dinaric Karst
Diversity Patterns in Europe
Diversity Patterns in the Tropics
Diversity Patterns in the United States
Bibliography
Austin, A. D., Cooper, S. J. B., & Humphreys, W. F. (Eds.), (2008).
Subterranean connections: Biology and evolution in troglobiont
and groundwater ecosystems. Invertebrate Systematics, 22(2),
85-310.
Bradbury, J. H., & Williams, W. D. (1999). Key to and checklist of the
inland aquatic amphipods of Australia. Technical Reports of the
Australian Museum, 14, 1—121.
Eberhard, S. E., & Humphreys, W. F. (2003). The crawling, creeping
and swimming life of caves. In B. Finlayson & E. Hamilton-Smith
(Eds.), Beneath the surface: A natural history of Australian caves
(pp. 127—147). Sydney: University of New South Wales Press.
Finston, T. L., Johnson, M. S., Humphreys, W. F., Eberhard, S., &
Halse, S. (2007). Cryptic speciation in two widespread subterra¬
nean amphipod genera reflects historical drainage patterns in an
ancient landscape. Molecular Ecology, 16(2), 355—365.
Guzik, M. T., Cooper, S. J. B., Humphreys, W. F., & Austin, A. D.
(2009). Fine-scale comparative phylogeography of a sympatric sis¬
ter species triplet of subterranean diving beetles from a single cal¬
crete aquifer in Western Australia. Molecular Ecology, 18(17),
3683-3698.
Hamilton-Smith, E., & Eberhard, S. (2000). Conservation of cave com¬
munities in Australia. In H. Wilkens, D. C. Culver & W. F.
Humphreys (Eds.), Subterranean ecosystems (pp. 647—664).
Amsterdam: Elsevier.
Hoch, H., & Howarth, F. G. (1989). The evolution of cave-adapted
cixiid planthoppers in volcanic and limestone caves in North
Queensland, Australia (Homoptera: Fulgoroidea). Me'moires de
Biospe'ologie, 16, 17—24.
Holmes, A. J., Tujula, N. A., Holley, M., Contos, A., James, J. M.,
Rogers, P., & Gillings, M. R. (2001). Phylogenetic structure of
unusual aquatic microbial formations in Nullarbor caves,
Australia. Environmental Microbiology, 3(4), 256—264.
Howarth, F. G. (1987). The evolution of non-relictual tropical troglo-
bites. International Journal of Speleology, 16, 1—16.
Howarth, F. G., & Stone, F. D. (1990). Elevated carbon dioxide levels
in Bayliss Cave, Australia: Implications for the evolution of obli¬
gate cave species. Pacific Science, 44(3), 207—218.
Humphreys, W. F. (1999). Relict stygofaunas living in sea salt, karst
and calcrete habitats in arid northwestern Australia contain many
ancient lineages. In W. Ponder & D. Lunney (Eds.), The other 99%:
The conservation and biodiversity of invertebrates (pp. 219—227).
Mosman, New South Wales: Transactions of the Royal Zoological
Society of New South Wales.
Humphreys, W. F. (2000). The hypogean fauna of the Cape Range
peninsula and Barrow Island, northwestern Australia. In H.
Wilkens, D. C. Culver & W. F. Humphreys (Eds.), Subterranean
ecosystems (pp. 581—601). Amsterdam: Elsevier.
Humphreys, W. F. (2001). Groundwater calcrete aquifers in the
Australian arid zone: The context to an unfolding plethora of
stygal biodiversity. Records of the Western Australian Museum,
(Suppl. 64), 63-83.
Humphreys, W. F., & Harvey, M. S. (Eds.), (2001). Subterranean biol¬
ogy in Australia 2000. Records of the Western Australian Museum,
(Suppl. 64).
Humphreys, W. F., Watts, C. H. S., Cooper, S. J. B., & Leijs, R. (2009).
Groundwater estuaries of salt lakes: Buried pools of endemic bio¬
diversity on the western plateau, Australia. Hydrobiologia, 626,
79—95. Erratum: Hydrobiologia, 632, 377.
Leys, R., Watts, C. H. S., Cooper, S. J. B., & Humphreys, W. F. (2003).
Evolution of subterranean diving beetles (Coleoptera: Dytiscidae:
Hydroporini, Bidessini) in the arid zone of Australia. Evolution,
57(12), 2819-2834.
Slaney, D. P., & Weinstein, P. (1996). Geographical variation in the
tropical cave cockroach Paratemnopteryx stonei Roth (Blattellidae)
in North Queensland, Australia. International Journal of Speleology,
25(1-2), 1-14.
DIVERSITY PATTERNS IN EUROPE
Louis Deharveng, Janine Qibert , 1 and David C. Culver
Museum National d’Histoire Naturelle, Paris, France,
^Universite Lyon I, France, * American University, U.SA.
INTRODUCTION
To an extent that is unusual in most branches of
zoology systematics, Europe is both a hotspot of
subterranean biodiversity and a hotspot of research into
subterranean biology, both historically and at present.
The scientific study of cave life can be traced back to
Johann von Valvasor's comments in 1689 on the
European cave salamander Proteus anguinus. This species,
the only stygobiotic salamander in Europe, reaches a
length of more than 25 cm, making perhaps the largest
stygobiont known anywhere. It occurs throughout the
Dinaric mountains in northeast Italy, Slovenia, Croatia,
and Bosnia and Herzegovina. During the late eighteenth
century and much of the nineteenth century, living
Proteus were collected and delivered to many scientists
throughout Europe. It was this animal more than any
other cave animal that played a formative role in the
emerging theories of evolution of Lamarck and Darwin.
The first invertebrate was described in 1832, also from
Slovenia, as Leptodims hochenwartii, a bizarre appearing
beetle with an enlarged abdomen and long spindly
appendages.
Besides the caves of the Dinaric region, the cave fauna
of the French and Spanish Pyrenees began to attract atten¬
tion, and the Pyrenean fauna began to be described by the
mid-nineteenth century, hr 1907, the Romanian zoologist
E.G. Racovitza published the enormously influential
"Essai sur les problemes biospeologiques," which set the
agenda for biospeleological research in the coming dec¬
ades. Together with the French entomologist Rene
Encyclopedia of Caves.
© 2012 Elsevier Inc. All rights reserved.
220
D
Jeannel, as well as Pierre-Alfred Chappuis and Louis
Fage, he established in 1907 an association named
Biospeologica. This association had three objectives: (1) to
explore caves and look for subterranean species; (2) to
obtain identifications and descriptions from specialists for
all material sampled; and (3) to publish results in a series
entitled Biospeologica. Ultimately, Biospeologica was
responsible for the inventory of the fauna of more than
1500 caves, mostly in Europe. More than 50 monographic
treatments of the taxonomy and distribution of European
cave fauna were published between 1907 and 1962
(Racovifa, 2005).
Even with more than a century and a half of descrip¬
tion and cataloging of the European subterranean fauna
(Juberthie and Decu, 1994—2000), both species descrip¬
tions and inventories are far from complete. The most
important large-scale inventory project of recent years, the
European project PASCALIS (Protocol for the Assessment
and Conservation of Aquatic Life in the Subsurface) ran
from 2002 to 2004. It developed common protocols
(Gibert, 2001) for the comparison of subterranean aquatic
species diversity at six sites in five countries (France,
Spain, Belgium, Italy, and Slovenia), and brought to light
a large amount of new taxa (Deharveng et al., 2009).
Individual country assessments of subterranean biodiver¬
sity are active as well, and at the beginning of the
PASCALIS project the most advanced of these was in
Italy, where there were more than 6000 records for 899
subterranean species (Stoch, 2001).
DIVERSITY COMPARISONS TO
OTHER CONTINENTS
As of 2000, approximately 5000 obligate subterranean
aquatic (stygobionts) and terrestrial (troglobionts) species
from Europe had been described. By contrast, 1200 had
been described from Asia, 500 from Africa, and 1000 from
North America (Gibert and Culver, 2009). The dominance
of Europe in known subterranean species is the result
of several factors. First, Europe has been better studied
than the other continents (Deharveng et al., 2000). This is
particularly evident in the status of biodiversity assess¬
ment in non-cave subterranean habitats. About half of the
stygobiotic species known from Europe are not primarily
cave-dwellers, but rather live in other subterranean habi¬
tats such as the underflow of streams and epikarst (Pipan,
2005). hr North America, these habitats are little studied
and account for less than a fifth of the known stygobionts.
A peculiar yet widespread terrestrial habitat, the MSS
(milieu souterrain superfidel), which is constituted of small
size voids among scree covered with soil, hosts a number
of troglobionts in Europe; this habitat is unsampled
in North America. In the cases of Asia and Africa,
large geographic areas have not been studied with
respect to subterranean species, both cave and non-cave
(Culver et al., 2001).
Second, subterranean biodiversity in Europe is actu¬
ally higher than on other continents. There is some
empirical evidence to support this. On a worldwide
basis. Culver and Sket (2000) list 20 caves and wells
that are known to have a total of 20 or more stygo¬
bionts and troglobionts. Of these, 13 are in Europe, 3
are in North America, and the rest are scattered else¬
where. This pattern is retrieved on a broader regional
basis. Among 7 large karst areas of Europe and North
America that are among the best documented for their
terrestrial subterranean fauna, the top rankings were
two European sites (Fig. 1). In several regions, espe¬
cially the western Balkans (northeast Italy, Slovenia,
Croatia, Bosnia and Llerzegovina, and Serbia) and the
Pyrenees (France and Spain), even the casual observer
can note the large number of stygobionts and troglo¬
bionts present, relative to caves on other continents. A
possible explanation for the increased diversity of sty¬
gobionts in the western Balkans is the complex biologi¬
cal and geological history of the Dinaric mountains
(Sket, 1999). The amount of available subterranean hab¬
itat in these mountains is large and has had a long his¬
tory of cave development and evolution. In addition,
the complex history of the Mediterranean Sea, includ¬
ing the fact that it almost dried up about 6 million
years ago (the Messinian crisis), may have resulted in a
greater invasion rate of the subterranean realm from
marine waters. Moreover, multiple invasions may have
occurred from surface continental waters. In the case
of terrestrial species, it is possible that invasion of
European caves was enhanced during interglacials of
the Pleistocene relative to North America. This is
because mountain ranges in Europe are largely east-
west oriented, whereas mountain ranges in North
America are largely north—south oriented. Thus, in
FIGURE 1 Species accumulation curves. Incidence functions of
number of troglobionts plotted against the number of 100-km 2 hexa¬
gons in 3 regions of Europe and 4 regions of North America. From
Culver et al., (2006).
ENCYCLOPEDIA OF CAVES
DIVERSITY PATTERNS IN EUROPE
221
North America, species could escape hot summers
either by migrating north or by invading caves. In
Europe, the orientation of the mountains reduced the
migration potential, perhaps resulting in increased
invasion rates of caves and other subterranean habitats.
TAXONOMIC RICHNESS OF THE
EUROPEAN SUBTERRANEAN FAUNA
The subterranean diversity pattern is better known
in Europe than in other continents, yet it is not always
well documented. This is obvious from the extensive
study of stygobiotic fauna conducted during the recent
project PASCALIS on six sites across Western Europe
that resulted in the discovery of at least 109 species
new to science, that is, more than 40% of the collected
species. Patterns are likely to be similar for terrestrial
fauna as reflected in the large number of new taxa
described every year from Europe, though no synthe¬
sis is available.
The distribution of the biota is controlled by a com¬
bination of factors, including geological history, phys¬
icochemical variables, aquifer structure, connectivity
between karst units, and biological interactions resulting
in complex biodiversity patterns, with an obvious hierar¬
chical dimension (Malard et al., 2009). Local (a) and
regional G) diversity is better known than the increase in
diversity over spatial scales ((3), although [3-diversity is
likely to be particularly significant across spatial scales
(aquifer, basin, and region). Put another way, the diver¬
sity in a single cave is always much less than the overall
diversity of caves even within a small area as a result of
reduced opportunities for migration and invasion—as
far as endemism is significant inside the area of interest.
The taxonomic diversity of the European fauna,
with more than 5000 stygobionts and troglobionts, is
too extensive and complex to be reviewed here, but
several highlights are especially noteworthy. The first
of these are the beetles in the terrestrial habitats. Two
families of beetles dominate in temperate zone caves:
the Carabidae and the Leiodidae. Both of these families
reach their zenith of subterranean species richness in
southern Europe, while southern China is emerging
since the 1990s as a second hotspot of biodiversity for
Carabidae. Among Carabidae, troglobiotic species are
distributed in 40 genera of the subfamily Trechinae
(Fig. 2), including seven genera with 18 or more cave
species: Anophthalmus, Aplwenops, Duvalius, Geotrechus,
Hydraphaenops, Neotrechns, and Orotrechus. The genus
Duvalius is especially speciose, with more than 250
species. In contrast, there is only one carabid genus in
North America with more than 20 cave species—
Pseudanophthalmus —although it has more than 200
species.
Among the Leiodidae, a total of 28 genera of the tribe
Leptodirini are known from the Pyrenees. Among them,
Bathysciola, a European genus, has more than 20 species
in this mountain range, often relatively widely distrib¬
uted and living mostly in soil habitats. In contrast, the
27 other genera are almost exclusively constituted of
subterranean species, all oligospecific, often troglo-
morphic and narrowly distributed (Fig. 3F). In the
Dinaric mountains, more than 50 Leiodidae genera are
only known from caves, many exhibiting the highest
levels of troglomorphy known among beetles world¬
wide, with Antroherpon containing 27 species. Once
again, the contrast with temperate North America is
instructive. There are only three genera, a single troglo-
morphic species ( Glacicavicola bathyscioides), and
Ptomaphagus has 19 cave obligate species—all oculated.
Among aquatic species, two groups are especially
noteworthy. One is the amphipod family Niphargidae.
There are nine genera and Niplwrgus alone has more
than 200 species. Taxonomically, the diversity of the
Niphargidae is rivaled by that of the Crangonyctidae
in North America (Holsinger, 1993). Especially spe¬
ciose is the genus Stygobromus, with nearly 200 species.
However, the crangonyctids seem to be less diverse
ecologically than the niphargids, both in terms of mor¬
phological variation and habitat variation. The final
group worthy of special note are the mollusks in the
Dinaric Mountains. In the Slovenian part of the Dinaric
Mountains, for example, there are 37 aquatic obligate
cave snails, one aquatic cave clam (Fig. 4C), and 11
species of terrestrial obligate cave snails. The entire
Dinaric Mountain region has several times that many
species.
TROGLOMORPHY AND RELICTNESS
Aside from its high taxonomic richness, the
European subterranean fauna includes a variety of
taxa outstanding by their degree of troglomorphy,
their phyletic isolation, or their unique adaptive fea¬
tures (Figs. 2, 3, 4).
Subterranean evolution has produced more troglomor¬
phy in Europe than in any other continent. The highest
troglomorphic representatives of the most diversified
groups of terrestrial cave fauna—beetles, springtails, woo-
dlice, aquatic Crustacea, spiders, pseudoscorpions—are
almost all European, Eastern Asia ranking second. As an
example, all Collembolan families present in temperate
caves, except Paronellidae and Arrhopalitidae, have their
most modified troglomorphic species in Europe (/.£’., with
thinner claws or longer antennae, in addition to depig¬
mentation and anophthalmy): Ongulogastrura longisensilla
(Pyrenees) in Hypogastruridae, Ongulonychiurus colpus
(Cantabric range in Spain) in Onychiuridae, Tritomurus
ENCYCLOPEDIA OF CAVES
222
D
FIGURE 2 Troglomorphic Trechinae (Carabidae) of Western Europe. (A) Lessinodytes pivai from Italian Alps; (B) Itaiaphaenops dimaioi from
Italian Alps, the largest cave Trechinae in the world; (C) Sardaphaenops supramontanus grafittii and Sardaphaenops adelphus from Sardinia; (D)
Aphaenops pluto from French Pyrenees; (E) Trichaphaenops obesus from French Alps; (F) Aphaenops cerberus from French Pyrenees. Photos (A) to
(C) courtesy of Enrico Lana and (D) to (F) courtesy of Christian Vanderbergh. Used with permission.
faldfer (Pyrenees) and T. veles (Croatia) in Tomoceridae,
Isotomiella unguiculata (Jura) in Isotomidae, Pseudosinella
cabidochei (Pyrenees) and Verhoeffiella longicornis (Dinaric
karst; Fig. 3E) in Entomobryidae, Oncopodurn lebretoni
(Pyrenees) in Oncopoduridae, Galeriella liciniana (Dinaric
karst) in Sminthuridae, and Megalothomx tubercidatus
(Pyrenees) and Neelus klisurensis (Dinaric karst) in
Neelidae. No explanation has been provided so far for the
prevalence of such high levels of troglomorphy in south¬
ern Europe compared to other regions on Earth.
Caves host more relicts, that is, species without
close phyletic relatives in other habitats of their
ENCYCLOPEDIA OF CAVES
DIVERSITY PATTERNS IN EUROPE
223
FIGURE 3 Some remarkable subterranean species of Europe. (A) Titanobochica magna, a recently described, highly troglomorphic
pseudoscorpion from Portugal; (B) Belisarius xambeui, a relictual troglophilic scorpion of Catalonia, only representative in Europe of the
family Troglotayosicidae; (C) Stenasellus virei hussoni, a frequent stygobiont of central Pyrenees (Isopoda); (D) Thaumatoniscellus spelunccie from
Croatian caves, one of the two species of Thaumatoniscellus in the monotypic subfamily Thaumatoniscellinae (Isopoda); (E) Verhoeffiella longicor-
nis, a highly troglomorphic Entomobryidae Collembola from Croatia; (F) Trocharanis mestrei from Pyrenees, a troglobiotic beetle of
the monospecific genus Trocharanis. Photo (A) courtesy of Sofia Reboleira; photos (B), (C), (F) courtesy of Christian Vanderbergh; and photos (D), (E)
courtesy of Jana Bedek. Used with permission.
distribution area, than any other habitat, and Europe
has more relicts than any other continent. Almost all
are narrow endemics, most are troglobiotic and highly
troglomorphic (Fig. 3), but some among the most
remarkable (such as the millipede Galliobates gracilis of
the monotypic family Galliobatidae, or the scorpion
Belisarius xambeui from Catalonia; Fig. 3B) are only
troglophilic, being found both in caves and soil. Though
most relicts are concentrated in the Dinaric range and
the Pyrenees, a few can be found as far north as
the northern Swiss Alps ( Niphatrogleuma wildbergeri, a
millipede) and as far south as southern Spain (Dalyat
mirabilis, the only European beetle of the South
African-American subfamily Promecognathinae).
At least several suprageneric taxa have their sole
subterranean representatives in Europe. The Dinaric
karst is particularly rich in this respect, with serpulids,
sponges, bivalves (Fig. 4), and the Cnidarian Velkovrhia
enigmatica. Extensive karstic outcrops such as the
Dinarids exposed to large bioclimatic fluctuations dur¬
ing long geological periods offer repeated opportu¬
nities for subterranean colonization and may account
for this richness in unusual cave taxa, often derived
from marine ancestors.
ENCYCLOPEDIA OF CAVES
224
D
FIGURE 4 Some remarkable stygobionts of the Dinaric karst in Croatia. (A) Eunapius subterraneus, the only subterranean freshwater sponge
in the world; (B) Marijugia cavatica, the only freshwater serpulid in the world (Polychaeta); (C) Congeria kusceri, the only stygobiotic bivalve in
the world, and Marifugia cavatica; (D) Troglocaris anophthalmus (Decapoda Atyidae) and Marifugia cavatica on a stalactite. Photos (A), (B), (D)
courtesy of Branko Jalzic and photo (C) courtesy of Helena Bilandzija. Used with permission.
GEOGRAPHIC PATTERNS
WITHIN EUROPE
While the inventory of known subterranean species
is far from complete, existing data provide some
important information about diversity patterns.
The first general pattern is that there is a gradient
in species richness with diversity dropping off to
the north. A comparison of Italy and the United
Kingdom, two European countries thoroughly inven¬
toried, shows this in a striking manner. In the United
Kingdom, a total of 12 troglobionts and 15 stygo¬
bionts, with a further 19 interstitial forms have been
recorded. In Italy, a total of 265 stygobionts, 321 trog¬
lobionts, and a few more from the MSS were known
in 2001. A difference of the same magnitude between
southwestern countries of Europe and Belgium was
highlighted during the PASCALIS project, and was
additionally expected to grow with increasing sam¬
pling effort. Of course, part of this difference is a
result of the covering of the United Kingdom and
Belgium by Pleistocene ice sheets or permafrost, but
undoubtedly other factors are at work to produce this
diversity gradient. As one moves north, mean annual
temperature declines and, all other things being
equal, so does net primary productivity. While little
or no primary production occurs in subterranean
habitats, they are dependent on surface primary
production for food. Thus, food availability decline
with increasing latitude is reflected in the diversity
pattern (Gibert and Deharveng, 2002).
It was recently shown that the north—south increase
in diversity follows the average annual temperature
gradient up to midlatitude only. To the south, diver¬
sity decreases again. Actually, there is a midlatitude
ridge between 42° and 46° north in Europe where bio¬
diversity peaks (Fig. 5). It roughly corresponds to parts
of the mountain ranges of the northern Mediterranean
region. In addition to the Coleoptera, Amphipoda, and
Gastropoda mentioned earlier, many other groups
reach their highest diversity along this ridge, including
isopods, spiders, and springtails. Historical circum¬
stances, especially the extensive shallow embayments
in the Tertiary and the Mediterranean salinity crisis of
ENCYCLOPEDIA OF CAVES
DIVERSITY PATTERNS IN EUROPE
225
FIGURE 5 The midlatitude biodiversity ridge of European troglobionts (redrawn and updated from Culver et al., 2006). Red circles are the
diversity hotspots in Slovenia and Ariege. Red triangles are other probable diversity hotspots. Orange circle is the Ardeche karst; orange trian¬
gles are other moderately rich karsts. Blue triangles are karsts with very few or no troglobionts. The boundary of the Pleistocene ice sheet is
shown as a scored solid line. A pair of dashed lines indicates the hypothesized position of the high diversity ridge.
the Miocene, as well as a large variety of karstic sub¬
terranean habitats, may account for the exceptional
diversification observed today. The presence of exten¬
sive karsts moderately impacted by Pleistocene glacia¬
tions along the ridge area and the long-term high
productivity in surface habitats at these latitudes
may have allowed the persistence of this striking
biodiversity pattern.
The second striking pattern is that of hotspots
nested inside hotspots; along the midlatitude ridge,
subterranean diversity and endemism are not homoge¬
neous, but peak in particular areas. These areas, long
known as exceptionally rich, are all located in moun¬
tainous and humid parts of the ridge. They exhibit an
unusual concentration of cave relictual endemics and
highly troglomorphic species across various taxonomic
groups, both aquatic and terrestrial. The richest area is
the Dinaric range, and the Pyrenees are second (Fig. 6),
while several other hotspots, less rich but also less
documented, exist in various areas (southern Italy,
southern Massif Central in France, southern Catalonia
and Cantabrian Cordillera in Spain, Southern Alps in
France).
The third and the fourth patterns highlighted below
are formally not specific to Europe, but are best docu¬
mented in this continent.
The third pattern that is evident is that the ratio of
troglobionts to stygobionts declines with increasing lat¬
itude from the biodiversity ridge northward (Fig. 7).
The reduction in troglobionts may be largely the result
of Pleistocene effects. No subterranean terrestrial habi¬
tats existed, as far as we know, within the ice sheets.
On the other hand, some subterranean aquatic species
may have survived the Pleistocene in running water
habitats underneath the ice cap. Additionally, it may be
that recolonization has been more rapid among aquatic
species (for example, through fluvial corridors and due
to the hydrological continuum and the connectivity
between aquifers). Stygobionts can be found in almost
all parts of Europe with suitable groundwater habitats,
even though these habitats are often not cave habitats,
but rather other subterranean habitats such as the
underflow of streams. To the south of the biodiversity
ridge, available evidence suggests that the ratio of trog¬
lobionts to stygobionts does not increase any more or
may decrease significantly, but data remain insufficient.
ENCYCLOPEDIA OF CAVES
226
D
FIGURE 6 Hotspots of richness in stygobionts. Each cell across southwestern Europe is 0.2° X 0.2°. From Deharveng et a\., 2009.
The fourth pattern is that of local changes in subter¬
ranean fauna resulting of habitat differences. For
instance, Copepoda, the most diverse group of stygo¬
bionts, were shown to contribute almost equally to the
overall species richness in karst and porous aquifers
and were usually slightly more diverse in nonsatu-
rated compared to saturated habitats. Karst aquifers
studied during the PASCALIS project were often richer
in stygobionts than porous aquifers and unsaturated
zones of the karst often richer than saturated zones.
Overall, however, correlation of habitat types and sty-
gobiotic species richness did not exhibit consistent pat¬
terns across sites (Table 1). Nevertheless, even when
species richness was similar, taxonomic composition
often differed among habitats of the PASCALIS sites, so
that, except in Wallonia and Cantabria, increase in spe¬
cies richness for a given number of samples was clearly
more significant when several habitats and both aquifer
types were included in the analyses. This might reflect
the fact that increased regional richness is correlated on
average to stronger habitat specialization.
At a finer scale, and more than any other factor,
habitat granularity generates a clear-cut pattern of
stygobiont distribution. The large voids in karstic
channels clearly enable a large size range of aquatic
organisms to occur while very tiny spaces in porous
aquifers represent a barrier to larger animals. For
example, in the Vidourle basin in southern France, the
amphipod Niphargus virei and the decapod atyid
Gallocaris inermis, are limited to the karstic aquifer,
while most micro-Crustacea are found in both porous
and karstic aquifers of this site.
The diversity patterns in subterranean terrestrial com¬
munities are less well known regarding habitat types,
but seem to be basically different, because, for unknown
reasons, interstitial animals of the deep soil (euedapho-
bionts, equivalent to the stygobionts of porous aquifers)
only exceptionally colonize caves. Biodiversity of large
subterranean voids is therefore mostly based on rela¬
tively large size invertebrates, while edaphobiont species
are, in fact, more diverse. This claim is based on compi¬
lation of heterogeneous and scattered datasets from the
literature. Actually, the fauna living in "porous" terres¬
trial habitats of caves has been very scarcely studied,
and comparative data between cave and soil fauna at
the same site are lacking.
ASSESSING EUROPEAN BIODIVERSITY
PATTERNS FOR CONSERVATION
The current incomplete state of knowledge of ground-
water biodiversity in Europe remains a major constraint
on successful implementation of its conservation. The
PASCALIS project elaborated a toolbox for the assess¬
ment of this diversity that includes validated methods
ENCYCLOPEDIA OF CAVES
DIVERSITY PATTERNS IN EUROPE
227
FIGURE 7 Ratio of troglobionts (TB) to stygobionts (SB) for western and central European countries where synthetized data are available
(from Gibert and Culver, 2005). The pattern includes non-cave stygobionts.
TABLE 1 Number of Species Collected in Different Habitats of the Six Sites of the PASCALIS Project
Aquifers
Habitats Inside Aquifers
karst
porous
u
s
h
P
Jura (France)
39
41
33
32
32
34
Roussillon (France)
26
43
18
17
35
31
Lessinia (Italy)
58
48
39
30
38
21
Krim (Slovenia)
64
59
34
42
41
48
Cantabria (Spain)
52
32
46
28
26
18
Wallonia (Belgium)
19
16
15
17
10
15
Karst habitats: upper unsaturated zone (u) and lower saturated zone (s); porous analogues: hyporheic zone (h) and phreatic zone (p). In bold italics are highest
values for habitat and aquifer at each site (from Dole-Olivier et al., 2009).
ENCYCLOPEDIA OF CAVES
228
D
for (1) determining the reliability of patterns of regional
biodiversity revealed by mapping of existing data;
(2) obtaining by means of a standardized field sampling
method an unbiased estimate of groundwater biodiver¬
sity in regions for which data were insufficient; and
(3) predicting overall species richness based on biodiver¬
sity indicators in regions with incomplete data sets.
These considerations led to the development of a com¬
mon database that allowed zooming in on subterranean
landscapes in order to consider groundwater biodiversity
at different spatial as well as ecological scales (European,
country, region, local, aquifer types, and habitats). The
project devised a specific action plan for the conservation
of groundwater biodiversity at a European level by
selecting priority regions for conservation using comple¬
mentarity approaches (Michel et al., 2009), by identifying
the spatial scale of relevance for conserving biodiversity
within these regions, and by formulating a series of
appropriate measures for maintaining their biodiversity.
Finally, shifting to an aquifer-based instead of a grid¬
cell—based approach for evaluating diversity patterns
and habitat selection is the next challenge to significantly
increase the relevance of conservation policy in Europe.
Acknowledgments
This work was supported in part by the European project PASCALIS
No. EVK2-CT-2001-002121. We thank Michel Perreau and Arnaud
Faille for information on Leiodidae and Carabidae beetle biodiver¬
sity, and Sofia Reboleira and Mark Judson for information on
Pseudoscorpiones and relictual taxa.
See Also the Following Articles
Diversity Patterns in Australia
Diversity Patterns in the Dinaric Karst
Diversity Patterns in the Tropics
Diversity Patterns in the United States
Bibliography
Culver, D. C., Deharveng, L., Bedos, A., Lewis, J. J., Madden, M.,
Reddell, J. R., et al. (2006). The mid-latitude biodiversity ridge in
terrestrial cave fauna. E cography, 29(1), 120—128.
Culver, D. C., Deharveng, L., Gibert, J., & Sasowsky, I. D. (Eds.),
(2001). Mapping subterranean biodiversity—Cartography de la
biodiversite' souterraine. Charles Town, WV: Karst Waters Institute
(Special Publication 6).
Culver, D. C., & Sket, B. (2000). Hotspots of subterranean biodiver¬
sity in caves and wells. Journal of Cave and Karst Studies, 62(1),
11-17.
Deharveng, L., Dalens, H., Drugmand, D., Simon-Benito, J. C., da
Gama, M. M., Sousa, P., et al. (2000). Endemism mapping and
biodiversity conservation in Western Europe: An arthropod per¬
spective. Belgian Journal of Entomology, 2(1), 59—75.
Deharveng, L., Stoch, F., Gibert, J., Bedos, A., Galassi, D. M. P.,
Zagmajster, M., et al. (2009). Groundwater biodiversity in Europe.
Freshwater Biology, 54(4), 709— 726.
Dole-Olivier, M.-J., Castellarini, F., Coineau, N., Galassi, D. M. P.,
Martin, P., Mori, N., et al. (2009). Towards an optimal sampling
strategy to assess groundwater biodiversity: Comparison across
six European regions. Freshwater Biology, 54(4), 777—796.
Gibert, J. (2001). Protocols for the assessment and conservation of
aquatic life in the subsurface (PASCALIS): A European project.
In D. C. Culver, L. Deharveng, J. Gibert & I. D. Sasowsky (Eds.),
Mapping subterranean biodiversity—Cartography de la biodiversite'
souterraine (pp. 19—21). Charles Town, WV: Karst Waters Institute
(Special Publication 6).
Gibert, J., & Culver, D. C. (2009). Assessing and conserving ground-
water biodiversity: An introduction. Freshwater Biology, 54(4),
639-648.
Gibert, J., & Culver, D. C. (2005). Diversity patterns in Europe.
In D. C. Culver & W. B. White (Eds.), Encyclopedia of caves
(pp. 196—201). Amersterdam, The Netherlands: Elsevier Press.
Gibert, J., & Deharveng, L. (2002). Subterranean ecosystems: A trun¬
cated functional biodiversity. Bioscience, 52(6), 473—481.
Holsinger, J. R. (1993). Biodiversity of subterranean amphipod crus¬
taceans: Global patterns and zoogeographical implications.
Journal of Natural Elistory, 27, 821—835.
Juberthie, C., & Decu, V. (Eds.), (1994—2001). Encyclopaedia biospeologica
(Vols. 1—3). Moulis, France: Societe Internationale de Biospeologie
(in French).
Juberthie, C., & Juberthie-Jupeau, L. (1975). La reserve biologique du
laboratoire souterrain du C.N.R.S. a Sauve (Card). Annales de
Spele'ologie, 30, 539—551 (in French).
Malard, F., Boutin, C., Camacho, A. I., Ferreira, D., Michel, G., Sket, B.,
et al. (2009). Diversity patterns of stygobiotic crustaceans across
multiple spatial scales in Europe. Freshwater Biology, 54(4), 756—776.
Michel, G., Malard, F., Deharveng, L., Di Lorenzo, T., Sket, B., & De
Broyer, C. (2009). Reserve selection for conserving groundwater
biodiversity. Freshwater Biology, 54(4), 861—876.
Pipan, T. (2005). Epikarst—A promising habitat. Ljubljana, Slovenia:
ZRC Publishing.
Racovifa, G. (2005). L'entreprise Biospeologica. Sa creation, son acti-
vite et ses realizations. Endins, 28, 25—34 (in French).
Sket, B. (1999). High biodiversity in hypogean waters and its endan-
germent—The situation in Slovenia, Dinaric Karst, and Europe.
Crustaceana, 15, 125—139.
Stoch, F. (2001). Mapping subterranean biodiversity: Structure of the
database and mapping software SKMAP and report of status for
Italy. In D. C. Culver, L. Deharveng, J. Gibert & I. D. Sasowsky (Eds.),
Mapping subterranean biodiversity—Cartography de la biodiversite' souter¬
raine (pp. 29—35). Charles Town, WV: Karst Waters Institute (Special
Publication 6).
DIVERSITY PATTERNS IN THE
DINARIC KARST
Boris Sket
Univerza v Ljubljani, Slovenia
WHAT IS THE DINARIC KARST?
The Dinarides or Dinaric Alps are a part of the Alpine
system along the east coast of the Adriatic, the western
part of the Balkan Peninsula in southern Europe (Fig. 1).
In the extreme northeast of Italy and the west of Slovenia
they are detached from the Southern Calcareous Alps by
a narrow belt of the so-called isolated karst. The Dinaric
Encyclopedia of Caves.
© 2012 Elsevier Inc. All rights reserved.
DIVERSITY PATTERNS IN THE DINARIC KARST
229
FIGURE 1 Position of the Dinaric karst. Mtg = Montenegro. After
Sket (1997).
karst with the surrounding belt of the isolated karst
patches covers approximately the southeastern halves of
Slovenia, Croatia, and Bosnia and Herzegovina, and the
whole of Montenegro, extending into southwestern
Serbia and Kosovo. In northwestern Albania it extends
up to Hellenides.
The Dinaric land appeared atop the Adriatic micro¬
plate out of the sea approximately 30 million years
ago as an island that separated the sea of Paratethys
from the then still existing narrow part of the Tethys
Ocean. Its flora and fauna originated and developed
in comparative isolation from the rest of Europe.
After orogenic faulting and erosion of the flysch layer,
karstification began—most probably at the end of the
Pliocene. It caused the origin of cave habitats, hydro-
graphic changes, and splintering of many biotic distri¬
bution areas.
Tropical or subtropical climates during the Tertiary
were followed by intermittent glaciations in the
Pleistocene. However, the closest extensive glacier in the
region was in the Julian Alps in northwestern Slovenia,
whereas in the Dinarides proper, only some of the high¬
est peaks were glaciated. The dry land extended during
the last glacials far south into today's Adriatic Sea,
enabling the continental biota to move away from their
former proximity of the coldest points. It has been
supposed that after glaciations, 10,000 years ago, the
whole Dinaric karst, except for its highest elevations and
a narrow coastal strip, before man's impact, had been
covered by extensive forests.
Humans began to colonize the territory (Sket, 1997) at
least by the end of the Riss glacial or 65,000 years BP.
With the onset of agriculture and animal husbandry
(Mesolithic, 10,000—4300 BC), more forest clearing
occurred. Illyric tribes depended primarily on domestic
animals, migratory grazing (transhumance), and storage
of hay. Wintering of large herds caused higher pressure
on forests' existence. The Copper to Iron Ages
(3000—200 BC) meant additional forest devastation for
mining and the primitive metallurgy. In the Halstatt
period, southern Slovenia was already quite densely
inhabited. Grave constructions testify to the existence of
bare rocky grounds in Lika (Croatia) in 500 BC and at
the coast in Bronze and Roman times. Up until 60 years
ago, before the reforestation started to be successful,
very large parts of the Dinarides were barren rock or
nearly so. The measures that are responsible for
comparatively green karst sites nowadays were the
introduction of the European black pine (Pinus nigra) for
reforestation, a ban on goats, and a remarkable reduction
in the human population in karst areas; the population
moved to the towns and to developing industrial centers
as well as to the fertile fields of the Pannonian Plain.
Today's Dinaric karst extends to more than 56,000 km 2 ;
it is an approximately 100-km-wide and 600-km-long
belt between the Adriatic Sea and the northeast hills and
plains. It includes most Adriatic islands and drowned
caves below sea level, and reaches some peaks having
2000-m elevations. Mountain ridges generally run paral¬
lel to the Adriatic coast (northwest—southeast). One of
the specific characteristics of this karst are poljes, which
are large, closed, periodically flooded depressions, typi¬
cally without a surface in- and outflow.
Thick carbonate layers, mainly Triassic limestones
(but also dolomites and limestones of other ages), are
highly porous; in the approximately 7000-km 2 area of
the Dinaric part of Slovenia only, close to 7000 caves
have been registered and investigated.
In the outer Dinarides (toward the Adriatic) the cli¬
mates are Mediterranean or sub-Mediterranean. The
mean yearly temperatures are between 12° and 17°C.
Precipitation is mainly about 1000—2000 mm yr _1 , but
is unevenly distributed and more than 50% in the cold
half-year; this causes regular droughts in summer and
washing away of soil in winter. The climate of the
inner Dinarides is more continental and cooler and
there is less precipitation. The nonfluctuating tempera¬
tures of underground habitats approximately equal
mean yearly temperatures of the region, which are
8—16°C higher than winter temperatures even in sub¬
tropical regions. The floods in rainy periods cause
some temperature fluctuations and enrich the under¬
ground with fresh food resources.
The comparatively isolated position of the Dinaric
territory along with high precipitation, its tropical past,
and very dynamic geomorphologic changes allowed
development of a very diverse flora and fauna. This is
ENCYCLOPEDIA OF CAVES
230
D
attested to by the fossil molluscan faunas of Pliocene-
Pleistocene lakes as well as by the diversity of survivors.
The territory's position on the edge of formerly glaci¬
ated regions allowed it to be a refuge for thermophile
elements during the Pleistocene. Particularly useful
refuges are the thermally conditioned hypogean habi¬
tats, where animals are not exposed to unfavorable win¬
ter temperatures.
COMPOSITION AND DIVERSITY OF
SUBTERRANEAN FAUNA
History
Disregarding the remote and isolated first mention
of a cavefish in China in 1540, a series of zoological
discoveries in caves started in the extreme northwest
part of the Dinarides. The amphibian Proteus anguinus
was first mentioned by Valvasor in 1689 and scientifi¬
cally described by Laurenti in 1768; however, it was
first seen in a cave in 1797. The next was the most
unusual cave beetle, found by the local cave worker
Luka Cec in the famous cave Postojnska Jama,
described by Schmidt in 1832 as Leptodirus hochenwartii.
These and a great number of other discoveries in the
nineteenth century were in the Kras (= Carso = Karst)
of Slovenia, which gave its name to the geomorpholo-
gic phenomenon of the karst. Discoveries of cave ani¬
mals in other parts of the world (first in the Caucasus,
Appalachians, Pyrenees, and New Zealand) occurred
in the middle of the century, and in other parts of the
Dinarides even toward its end. Thanks to such a his¬
tory, the "classical Karst" won the appellation of the
"cradle of speleobiology (biospeleology)." Hamann's
(1896) book Europaische Hohlenfauna, which even at that
time already cited 400 references, was still devoted
mainly to the cave species from Slovenia.
The most recent census of the Dinaric obligate sub¬
terranean fauna (Sket et al., 2004) revealed more than
900 species, 600 terrestrial and 330 aquatic troglo-
bionts. If we compare this to 930 and 420 troglobionts
in the whole of North America, the importance of
such numbers is evident. Among regions of approxi¬
mately the same size class in the world, the Dinaric
karst is by far the richest in the obligate subterranean
fauna. This does not seem to be a consequence of the
state of investigations. To the troglobionts, one should
add some regularly occurring eutroglophiles as well
as a high number of normal surface animals (troglox-
enes or accidentals) that occasionally, but with a high¬
er or lower regularity, happen to occur in caves. Some
representatives of the Dinaric fauna are shown in
Figure 2.
Terrestrial Cave Fauna
By far the most diverse group of troglobionts are bee¬
tles (Coleoptera), represented by more than 200 species.
This should not cause wonder, because Coleoptera are
by far the richest animal group in the world; neverthe¬
less, its representation underground is very biased; it is
represented mainly by members of only two families.
In addition, the other insect groups—which are still ex¬
tremely diverse on the surface—are hardly represented
underground at all. The most important in Dinaric
caves is the family Cholevidae with its subfamily
Leptodirinae (known also as Catopidae: Bathysciinae).
Approximately 175 species are distributed in as many
as 50 genera. Although this is originally an edaphic (z.e.,
soil-inhabiting) group, only a few cave-inhabiting spe¬
cies are not specialized to the cave habitat. And only
few Dinaric genera are not endemic to this region: the
troglophilous Phaneropella lesinae Jeannel exhibits a
transadriatic distribution and its congenerics occur also
far in the Asia Minor. The edaphic and troglophilous
Bathyscia montana Schioedte may serve as an archetype
of edaphic leptodirines: it is less than 2 mm long, egg
shaped, with short legs and antennae. Some troglobionts
are similarly built, but most of the others attained differ¬
ent degrees of troglomorphism. They are regularly big¬
ger and at least slightly elongated.
The higher troglomorphic species may be pholeuo-
noid, which means that they have a spindle-shaped
abdomen and a neck-shaped prothorax. Such species
include Parapropus spp. in the northwestern part of the
Dinarides; similarly shaped cave beetles may occur
also in other karst regions (such as Pholeuon in
Romania). Only in the Dinaric karst have the beetles
attained the highest, that is, the leptodiroid, degree of
troglomorphism. In the name-giving Leptodirus hochen¬
wartii Schmidt from Slovenia and Croatia, the legs and
antennae are extremely elongated, the "neck" (protho¬
rax) is nearly cylindrical, and high-vaulted elytrae
make the abdomen egg shaped or nearly globular
(Fig. 2F). Of similar shape are more than 25 member
species of the genus Antroherpon in the southeastern
part of the Dinaric karst. The high number of species
probably indicates the effectiveness of such a body
transformation. Nowhere else in the world has this
group of beetles (Cholevidae) achieved such diversity
as in the Dinaric caves.
The other beetle group in Dinaric caves is the family
Carabidae with 12 genera and 80 troglobiotic species of
its subfamily Trechinae. The carabid beetles in surface
soil are predators; they are twice as big and longer
legged than their scavenger cholevid neighbors. The
northwestern Anophthalmus spp. or the southeastern
Neotrechus spp. are, however, larger, more slender,
and with longer appendages than the epigean or
ENCYCLOPEDIA OF CAVES
DIVERSITY PATTERNS IN THE DINARIC KARST
231
FIGURE 2 Some representatives of the Dinaric subterranean fauna. (A) Typhlogcimmarus mrazeki (Amphipoda), (B) Mnclwerites spelaeus
(Coleoptera: Pselaphidae), (C) Monolistra monstruosa (Isopoda: Sphaeromatidae), (D) Niphargus orcinus (Amphipoda: Niphargidae), (E) Zospeum
kusceri (Gastropoda: Carychiidae), (F) Leptodirus hochenwartii (Coleoptera: Cholevidae), (G) Synurella ambulans (Amphipoda: Crangonyctidae),
(H) Delaya bureschi (Oligochaeta: Haplotaxidae), (I) Stalita taenaria (Araneae: Dysderidae), (J) Lanzaia vjetrenicae (Gastropoda: Hydrobioidea).
ENCYCLOPEDIA OF CAVES
232
D
troglophilous Trechus spp. Some other troglobionts such
as the Aphaenopsis spp. are even more troglomorphic,
but this transformation never reached the degree it has
in cholevids. The Dinaric Trechinae are also less excep¬
tional; their diversity is approached by other cave fau¬
nas, such as those in North America. To the same
family, but its subfamily Pterostichinae, belong also two
comparatively large, approximately 10-mm-long species
of Laemostenes ; although also living outside caves, they
are worth mentioning because they are present in nearly
all caves in the whole Dinaric karst and far outside it.
The family Pselaphidae (Fig. 2B) is represented by
only 20 troglobiotic species. These are tiny beetles
with very slender thoraxes and wide abdomens with
shortened elytrae. Their surface relatives are also soil-
dwellers. The related extensive family Staphylinidae,
with a nearly tap-shaped body with very short elytrae,
resulted in the Dinarides in no troglobiotic species,
but a number of species occur as troglophiles or occa¬
sional guests in the entrances of the caves. Examples
include some tiny species of Atheta and some larger
Glyptomerus spp. Some tiny weevils (Curculionidae)
are also known from Dinaric caves. Some species of
the genus Troglorrhynchus are often regarded as tro¬
globionts. However, they depend on tree roots that
are not as common and extensively developed in karst
caves as they are in lava tubes, instead they are shal¬
low underground. Troglorrhynchus spp. are in fact, as
soil inhabitants, most probably merely guests in caves.
Most troglobiotic beetles in the Dinarides are eyeless
and have reduced hind wings but none is really pig¬
mentless; although translucent, most are dark brown,
whereas some are yellowish brown in color. Although
carabid beetles are more or less evenly distributed along
the Dinaric area, slightly more diversified in the north¬
west, the specific cholevids become more diverse toward
the southeast; the southern Herzegovina, with neighbor¬
ing parts of Croatian Dalmacija and Montenegro, by far
being the richest.
The next richest groups of troglobionts are the spi¬
ders and false scorpions. The genus Troglohyphantes is
very characteristic among spiders (Araneae). Its small
representatives of mostly less than 4 mm in body
length span their small sheet webs either in the forest
litter, in micromammalian burrows (they are microca-
vernicole), in entrance parts of the caves, or deeper in
them. Only the latter (deep in caves) are mostly pig¬
mentless and eyeless; there are approximately 25
troglobiotic species. In contrast, the always more than
5-mm-long spiders of the family Dysderidae hunt
without webs. Among 25 species of 7 mainly fully tro¬
globiotic genera belongs also the first scientifically
described cave spider Stalita taenaria Schioedte.
These are shiny brown to orange in the cephalothorax
and legs and dirty white on the abdomen. Not less
than 7 spider families are represented by troglobiotic
species in the Dinarides. The strangest might be
the family Anapidae; its species Pseudnnapis relicta
Kratochvil from Montenegro measures only 1.4 mm in
length and its relatives are geographically very far
away.
Similarly numerous are false scorpions (Pseudoscor-
piones), which, however, are less diverse, representing
only three families, which have similar body shapes.
The most specialized troglobiotic species are those
with extremely slender and elongated pincers on the
second pair of appendages and an up to 8-mm body
length. They are very different from their relatives that
inhabit forest soil. Neobisium spelaeum Schioedte was
the first species described from caves.
The other groups represented by troglobiotic spe¬
cies are (in descending order) millipedes (Diplopoda),
woodlice (Isopoda: Oniscidea), snails (Gastropoda),
centipedes (Chilopoda), harvestmen (Opiliones), and
planarians (Turbellaria: Tricladida). Remarkable milli¬
pedes are species of the genus Apfelbeckia that are most
probably subtroglophile or even trogloxene. At
approximately 10 cm they are the largest European
millipedes. During the hot and dry Mediterranean
summers, they are found only at the entrances of
caves, giving those rooms a sweetish scent because of
their defense glands. Among snails, widely distributed
are tiny species of Zospewn (Fig. 2E), a genus present
also in the Alps and Pyrenees; their shells are about
2 mm high, ovoid, may be seen on wet cave walls or
may be washed out of the earth from the cave bottom.
In the southeastern parts of the Dinarides, bigger snail-
shells may be seen, such as those of Aegopis spelaeus
A.J. Wagner, which can be up to 20 mm in diameter.
Terrestrial is also the small prosobranch Pholeoteras
euthrix Sturany with a hairy shell. The pigmentless
species of terrestrial planarians from Slovenia has not
yet been described.
Like everywhere, bats are an important group of
subtroglophile cave inhabitants. However, very large
colonies hardly exist anymore—nowadays, a colony of
1000 specimens is a very high number for the area.
One such year-round cave species is the long-fingered
bat ( Myotis capaccinii Bonaparte). Particularly bound to
caves is the Schreibers' bat ( Miniopterus schreibersii
Kuhl). The most often seen species, but more solitary,
are different horseshoe bats ( Rhinolophus spp.). Bats are
important vectors of food resources between the sur¬
face and the underground. Individuals of other mam¬
mal species may also be present deep in caves,
including the rodent the edible dormouse ( Glis glis
Linne), which was first scientifically described just
from the beech forests of the Dinarides, and some mar¬
tens ( Martes foina Erxleben). No birds enter the totally
dark parts of Dinaric caves. However, the twilight
ENCYCLOPEDIA OF CAVES
DIVERSITY PATTERNS IN THE DINARIC KARST
233
areas near the entrances used to be common nesting
places for large colonies of rock doves ( Columba livia
Gmelin), giving a number of caves the name Golobina,
Golubinka, and so on ( golob or golub is the Slovene or
Croatian name for Columba spp.). Unfortunately, their
populations have been heavily decimated or dispersed
by hunters or by other human actions. On the other
hand, groups of feral pigeons joined their wild rela¬
tives in some caves.
Aquatic Subterranean Fauna
As is generally true throughout the world, by far
the prevailing group in Dinaric hypogean waters is the
crustaceans (Sket, 1999). Peculiar, however, to the
Dinarides is their richness in snails (Gastropoda);
therefore, we discuss them first here.
With approximately 130 species, the Dinaric aquatic
gastropods represent close to half of the world's
known aquatic troglobiotic snail fauna. They are also a
very common appearance in Dinaric cave waters and
their empty shells may accumulate in more or less
pure piles (thanatocenoses), even some meters long.
Nearly all of them belong to the group Hydrobioidea
whose formal subdivision is still actively changing.
These are snails with a discoid, conical to nearly rod¬
shaped shell, with a simple or highly elaborated, ear¬
like mouth and a smooth to strongly ribbed whirl. The
mostly 2- to 3-mm-high shells are approximately twice
as big as those of their North American relatives,
which is a difficult-to-explain curiosity. The shells of
Hadziella species (from the northwest Dinarides) are
flat; Hauffenia are mostly widely conical, Iglica are very
slender ( iglica in Slavonic languages means "small nee¬
dle"); the Lanzaia (Fig. 2J) and Plagigeyeria spp. from
the southeast exhibit very diversely ornamented,
ribbed shells with elaborate mouths. A common mem¬
ber of pulmonate snails (Pulmonata: Ancylidae) is the
generalist, genetically/cytologically diverse and there¬
fore widely distributed species Ancylus fluviatilis O.F.
Mueller with a cap-shaped shell without a spiral. It
may also be present in caves and such a cave popula¬
tion might be clinally (gradually) troglomorphized:
skin pigmentation gradually disappears and eyes
become more and more reduced along the sinking
stream in the Postojna-Planina Cave System. Acroloxus
tetensi (Kuscer), with a similarly cap-shaped shell, is a
troglobiotic species of Acroloxidae.
Among crustaceans, the most numerous are copepods
and amphipods, with approximately 60 species each. Of
amphipods (Amphipoda), not less than 45 species are of
the genus Nipltargus (Fig. 2D). This is the most diverse
genus of the group, spread over almost all of Europe
(and in Iran and the Arabian Peninsula in Asia). It
inhabits all freshwater habitats, including forest ditches,
brooks, caves, and interstitial waters. Specimens of dif¬
ferent species may be 2 to 30 mm long and extremely
diverse in their body shapes. However, all of these spe¬
cies are absolutely without eyes. The same variability
span is achieved in the Niphargus species of the
Dinarides. One of the smallest species, N. transitivus
Sket, inhabits interstitial and karst waters in the extreme
northwest. The slender and up to 20-mm-long N. stygius
Schioedte, the type species of its genus, inhabits mainly
percolation waters in Western Slovenia; the very
stout and large N. orcinus Joseph and its relatives are
spread throughout larger cave water bodies along the
Dinaric karst. Particularly interesting is the long-legged
N. balcanicus Absolon with a densely spiny back,
from Herzcegovina; similar is N. dolichopus Fiser,
Trontelj, and Sket with even longer pereopods, from
Bosnia. The related Niphargobates orophobata Sket from
the fissure system in Slovenia has the only close relative
in the Greek island of Kriti (Crete). Another relative of
Niphargus, Carinurella paradoxa Sket, which inhabits inter¬
stitial waters within the northwest Dinaric karst, is
able to roll into a ball. The up to 25-mm-long and stout
Typhlogammarus mrazeki Schaeferna (Fig. 2A) is an
omnipotent raptor that can catch a shrimp swimming or
climb up a vertical wall with a thin layer of water.
The extremely fragile Hadzia fragilis S. Karaman has par¬
ticularly numerous relatives in the Mediterranean and
in the Caribbean. It is a mainly anchihaline species but a
few populations are present in continental freshwaters.
Copepods (Copepoda) are represented by some gen¬
eralist and a number of troglobiotic species, many com¬
ing from the genera Diacydops and Acanthocydops.
Cyclopoids are particularly richly represented in larger
bodies of water, where they may even be accompanied
by the explicitly planctonic diaptomids (Calanoida:
Diaptomidae); four Dinaric cave species is an enormous
number for freshwater Calanoida. Troglodiaptomus sketi
Petkovski exhibits a holodinaric distribution; the others
are more endemic. Harpacticoids are particularly well
represented in the percolation waters in fissure systems.
Their genera Elaphoidella with 20 and Parastenocaris with
15 species are the leading groups, and numbers are still
growing.
Isopoda are not numerous, but are very diversely
represented. The asellids (family Asellidae) are repre¬
sented by the troglomorphic races of Asellus aquaticus
Linne and by some troglomorphic Proasellus species,
such as P. slovenicus Sket in the northwest or the diversi¬
fied species P. anophthalmus S. Karaman in the southeast.
Very characteristic is the rich group of aquatic pill-bugs
Monolistra (family Sphaeromatidae, Fig. 2C), some of them,
for example, M. spinosissima Racovitza, with long spines
on their backs. The largest isopods are cirolanids
(Cirolanidae) Sphaeromides, distributed close to the
ENCYCLOPEDIA OF CAVES
234
D
Adriatic coast but without any known ecological or his¬
torical connection to the sea. The water fleas (Cladocera),
although with only three tiny (less than 0.5-mm-long)
troglobiotic Alona species, are relatively richly repre¬
sented. The generally interstitial Microparasellidae,
such as Microcharon spp. or Microparasellus sp., are not
numerous, but they also occur in caves or in gravel beds
outside them.
Thermosbaenacea are represented by an anchihaline
coastal species, Monodella lwlophila S. Karaman, and a
mainly interstitial freshwater species Limnosbaenci
finki Mestrov and Lattinger. All troglobiotic decapod
shrimps belong to the family Atyidae (Sket & Zaksek,
2009); they are very common in Dinaric caves and up
to three species may occur in the same cave (in
Vjetrenica). The most widely distributed species aggre¬
gate Troglocaris (Troglocaris) anophthalmus Kollar is
holodinaric and split into a number of independent
species. Members of subgenera Troglocaridella and
Spelaeocaris are southeastern merodinaric; five species
have been recognized in the latter.
Also worth mentioning are 13 epizoic to parasitic
species of Turbellaria Temnocephalida, belonging to
the family Scutariellidae and living mainly on atyid
cave shrimp. Besides these troglobiotic species, only
one epigean species is from Europe, again from the
same, Dinaric, region.
Three species of true filter feeders should be
mentioned, because these types of animals are very rare
in caves. Besides some trogloxene or eutroglophile spe¬
cies, the only troglobiotic freshwater sponge (Porifera:
Spongillidae), Eunapius subtermneus Sket et Velikonja,
inhabits Dinaric caves near Ogulin (Croatia); it is of a
softer consistency than the surface species and forms
only a few gemmulae. The only troglobiotic clam
(Bivalvia: Dreissenidae), Congeria kusceri Bole, is similar
to and related to the well-known zebra clam; it fastens
itself to rocks by means of byssal threads. The only trog¬
lobiotic tube worm (Polychaeta: Serpulidae), Marifugia
cavatica Absolon et Hrabe, fastens itself to rocks with its
less than 1-mm-wide calcareous tubes. It is particularly
interesting that generations of tubes may accumulate to
build meter-thick tufa-like layers. The only troglobiotic
cnidarian (Hydrozoa: Bougainvilliidae), Velkovrhia enig-
matica Matjasic et Sket, is ecologically similar. A colonial
species, it attaches to the substratum by means of sto¬
lons, with sessile medusoids still distinctly developed.
All these putative species exhibit a holodinaric
distribution.
The only troglobiotic vertebrate in Europe is the usu¬
ally approximately 20-cm-long cave salamander Proteus
anguinus Laurenti. Its distribution area is holodinaric
with some localities in the small Italian part of the
Dinaric karst and with the extreme southeastern reliable
localities in southeast Herzegovina. At present, two
subspecies are formally described, although this does
not match its whole diversity. All proteus populations
exhibit outer gills, only three toes on the first legs, and
two on the hind legs. The holodinarically distributed
troglomorphic race P. a. anguinus has a more or less col¬
orless skin and an elongated head with reduced eyes,
hidden below the skin. The nontroglomorphic P. a. par-
kelj Sket & Arntzen is limited to a tiny area in the
extreme southeastern part of Slovenia. It is very darkly
(sometimes black) pigmented, has normally developed
eyes, and different body proportions. However, both
races are obligate cave-dwellers and both come out of
springs on some nights.
In the central and southern parts of the Dinaric
karst, a number of endemic cyprinid fishes (Pisces:
Cyprinidae) with small distribution areas are present.
Most of them are known as regular periodical coloni¬
zers of caves (subtroglophiles), leaving the open waters
of poljes before the retreat of the water and appearing
again soon after a flood. They were observed to do both
directions of migrations actively, not because they were
directly forced to do so by hydrological events. In
ancient times some species were economically impor¬
tant; special constructions were placed in caves to catch
them in masses since they were a highly prized export
article of the otherwise poor karst regions. A number
of species used to be attributed to the genus Phoxinellus,
earlier called Pamphoxinus. They exhibit different degrees
of scale reduction; P. alepidotus Heckel possesses only a
few scales along the lateral line. These fishes have been
recently mainly displaced to Telestes and Delminichthys ;
taxonomically the most particular species is Aidopyge
huegelii Heckel.
BIOGEOGRAPHICAL PATTERNS IN THE
DINARIC UNDERGROUND
Widely Spread Taxa
Species or species groups may inhabit very different,
wide areas within the Dinaric karst or—in a few cases—
beyond it. The latter may be particularly instructive,
possibly explaining to us the noncontiguous distribution
areas of even highly specialized cave species. Such is
the case with the isopod Asellus aquaticus, amphipod
Synurella ambulans F. Mueller (Fig. 2G), and gastropod
Ancylus fluviatilis. They are highly generalist species of
approximately European distribution that penetrate
underground particularly along sinking rivers. In the
Dinaric karst, some cave populations of these species
became troglomorphic and troglobiotic. Similar but less
well known is the situation with some generalist cyclo-
poid copepods.
ENCYCLOPEDIA OF CAVES
DIVERSITY PATTERNS IN THE DINARIC KARST
235
Transdinaric Distribution
Some groups of species inhabit most of the Dinaric
karst, but occur also in areas to the east or to the west.
Such is the shrimp genus Troglocaris ; relatives of the
handful of Dinaric species are present in Georgia
(Sakartvelo) in the east. Dinaric species of Sphaeromides
have close relatives in Bulgaria and eastern Serbia.
For both groups, molecular analysis excluded the
French species from being congeneric. The rich genus
Monolistra covers the whole Dinarides and reaches far
into the southern calcareous Alps.
Among terrestrial animals, transdinaric are the gas¬
tropods Zospeum, a genus reaching from the southeast¬
ern Dinarides, through the southern Alps, far into the
Pyrenees.
Holodinaric Distribution
Also the scattered holodinaric distribution of some
taxa may be a consequence of multiple (i.e., polytopic
and/or polychronous) immigration with a possible
subsequent extinction of surface populations. In some
cases, the immigration was followed by speciation; in
fact, in all cases studied by DNA analysis, a split into
distinct (biological) species could have been shown.
In some other cases, the holodinaric area might be
inhabited by a homogeneous species. The "holodina¬
ric" area, starting with the northwestern border of the
Dinaric karst, finishes with the southeasternmost
Flerzegovina. In all holodinaric species (or genera),
the taxon is never contiguously spread over most of
the area; it is instead split into groups of populations
in hydrographically isolated karst areas. Such a distri¬
bution exhibits the most unique cavernicolous repre¬
sentatives of their higher groups: the amphibian
Proteus anguinus, the clam Congeria kusceri, the tube
worm Marifugia cavatica, and the cnidarian Velkovrhia
enigmatica. Other cases are particular representatives
of bigger groups, for example, one gastropod species,
Zospeum amoenum Frauenfeld, the spider Parastalita
stygia Joseph, the amphipod Niphargus steueri Schellenberg
and some others.
Merodinaric Distribution Patterns
The northwestern merodinaric elements are, for
example, the small beetle genus Leptodirus, the flat-
shelled hydrobioid snails Hadziella, and the monolistrine
(Isopoda) subgenus Microlistra. Their approximate counter¬
parts in the southeastern merodinaric area are the beetle
genus Antroherpon with numerous leptodiroid species,
and hydrobioid snail genera Lanzaia and Plagigeyeria.
No explanation for this bipolarity of the continental
Dinaric cave fauna has been presented until now. One
has to emphasize that both continental merodinaric
areas, when combined, do not correspond to the holodi¬
naric area. A paralittoral merodinaric distribution along
the Adriatic coast is exhibited by the anchihaline amphi¬
pod Hadzia fragilis S. Karaman, the thermosbaenacean
Monodella halophila S. Karaman, and in a big part of
the belt also by the amphipod Niphargus hebereri
Schellenberg. These predominantly anchihaline species
are not present in the brackish subterranean waters
of the Kvarner (Quarnero) Gulf, which was inundated
by the sea only late in the Pleistocene; this shows us
clearly that also the paralittoral distribution pattern has
a historic—not purely ecological—background.
Smaller Distribution Areas
Most species and subspecies, but also some genera,
exhibit smaller distribution areas within one of the mero¬
dinaric areas. The most interesting fact is that most of
these areas do not follow the recent hydrological divides.
One species (or subspecies) may either cross the borders
of a divide or may be limited just to a part of the actual
drainage area. This has been more properly studied for
species of the isopod genus Monolistra and of some
hydrobioid snails. It has been supposed that such distri¬
bution areas had been achieved in geologically past
drainage areas (sometimes still prekarstic, on the surface)
and maintained by competition between related species
today. Most other small distribution areas are not specifi¬
cally explainable; they may be either relics of a formerly
wider area or results of locally limited immigrations
underground. A number of animal species is known
from only one locality, either a cave or a spring; the most
striking is the amphipod Niphargobates orophobata Sket
from just one jet of percolating water in Planinska Jama
(Slovenia).
Endemism
Only a negligible number of obligate subterranean
species cross the borders of the Dinaric region. Such
is the case of some amphipods, snails, and beetles
crossing the slightly indistinct border between the
Dinaric and South Alpine regions, and such is the
case of some copepod species, which may even
exhibit wider distribution areas. On the other hand,
also comparatively high is the endemism between
regions within the Dinarides. At the moment, the
only distribution data for administrative regions
(states) that are known definitely do not reflect the
natural biogeographical units. Nevertheless, the num¬
ber of endemic species (for countries) is mostly
40—60%, sometimes even much higher. Consideration
of nominal subspecies would increase the endemicity
remarkably as has been studied in some large genera.
ENCYCLOPEDIA OF CAVES
236
D
while even a number of endemic genera exist, scat¬
tered to different taxonomic groupings; they are par¬
ticularly numerous in the cholevid beetles. Only 5 of
the 44 present cholevid genera are not endemic for
the Dinaric karst and at least 23 genera seem to be
endemic for one of the states (Slovenia, Croatia, or
Bosnia and Herzegovina). Because political borders
are crossing regions with homogeneous faunistic
assemblages, the endemism degree would certainly
be much higher in natural biogeographic provinces.
HOTSPOTS WITHIN THE HOTSPOT
In this highly biotically diverse area, some cave sys¬
tems have particularly rich cave faunas. When search¬
ing for caves with more than 20 obligate cave
inhabitants, 20 such caves or cave systems were traced
and as many as 6 of them were in the Dinaric karst
(Culver and Sket, 2000). The Postojna-Planina Cave
System (PPCS) is, with 99 such species, the richest
among them. The system consists of 17 km and 6 km
of passages connected by 3 km of flooded corridors
not yet mastered by cave divers. The dry parts of the
Postojnska Jama is one of the oldest (since 1818) and
most famous tourist caves in the world, with a small
railway in it; some parts of the system are used exten¬
sively by tourists, but there are still many "wild" parts
that are rich in fauna. This is the type locality for a
number of "first cave" animals, including the first
described troglobiont, the beetle Leptodirus hochenwartii,
the first cave spider Stalita taenaria, and many others.
Altogether nearly 60 animal species or subspecies were
first found and described from this system, and the
European cave salamander, Proteus anguinus, was first
seen in its natural habitat in Crna Jama in 1797. It is
also a site of long-term ecological studies. Dry parts of
the system are inhabited by 37 species, 9 of which are
beetles. There are 62 aquatic troglobiotic species; par¬
ticularly numerous are crustaceans, snails, and oligo-
chaetes. The main artery of the system is the sinking
river Pivka, which sinks at the Postojnska Jama and
resurges from the Planinska Jama; there is approxi¬
mately 10 km of underground bed in-between. A
mixed fauna of troglobiotic, eutroglophile, and troglox-
ene species inhabits it. The last ones are mainly insects;
aquatic larvae of some of them are present all along
the course. Populations of some eutroglophile species
exhibit a clinaly (gradually) increased troglomorphy
along the stream. These caves are in Slovenia, which is
the richest in the world for aquatic subterranean fauna.
The next four richest caves are also in Slovenia: Krizna
Jama, Logarcek, Sica-Krka System, and Grad.
In Herzegovina, which seems to be the richest area
in the world for terrestrial cave fauna, evidence is
growing that a number of additional caves will be cat¬
apulted to the status of the richest ones. Until now,
only Vjetrenica in Popovo Polje seems to reach the
PPCS in number of troglobiotic species. This complex
cave system has 6 km of passages, which include a
number of small streams, pools, and trickles of water.
The surface animals (trogloxenes) are very few here.
There is no sinking stream, but water jets after rains
import organic debris through crevices and shafts by
which the rich fauna are being fed. Among troglo-
bionts there are 10 species of amphipods, some of
which are very large, and three species of decapod
shrimps. There are approximately 10 species of beetles.
Particularly noteworthy are the amphibious catopid
beetle Hadesia vasiceki J. Mueller and the amphipod
Typhlogammarus mrazeki Schaeferna, which occupy the
cave hygropetric (described below).
SPECIAL ASSEMBLAGES^
In an area with such a high number of faunal ele¬
ments and, therefore, comparatively high pressure for
interspecific competition, one can expect further eco¬
logical specialization of species. Therefore, besides a
number of more or less "trivial" communities and
synusiae, composed of species that sometimes merely
by chance are locally combined, a small number of
characteristic groupings with particular ecological
demands or abilities exist.
Ice caves are scattered all over the area. They are
usually sack-shaped and surrounded by forest in mod¬
erately higher elevations and bear ice deposits during
the whole year. Particularly wet soil with close to zero
temperature is the ecological characteristic of the prox¬
imity of such a subterranean glacier in summer. This
environment is usually particularly rich in troglobiotic
and troglophilic (e.g., Nebria spp.) Coleoptera. However,
the highly troglomorphic leptodirine beetles of the
genus Astagobius seem to have specialized to such an
environment.
The cave hygropetric is a rocky wall, usually sinter,
overflown by a thin layer of slipping water. A number
of leptodirine beetles seem to be specialized for climb¬
ing in such a semiaquatic environment and collecting
organic particles brought by water. All of them—
although not closely related—have a pholeuonoid
body shape, with very strong claws, and with particu¬
larly hairy mouth apparatus. The best-known repre¬
sentative of such beetles is Hadesia vasiceki. These
obligate inhabitants may join some nonspecialized
semiterrestrial or semiaquatic guests using such envir¬
onments as hunting grounds or simply as distribution
paths.
ENCYCLOPEDIA OF CAVES
DIVERSITY PATTERNS IN THE DINARIC KARST
237
A number of aquatic and probably some terrestrial
species are nearly limited to the percolation-water-
filled crevice systems, which include the epikarst ;
however, such species are primarily found in percola¬
tion-water-fed rimstone pools or puddles on cave
walls or bottoms. Morphologically adapted to such
narrow places seem to be the harpacticoid copepods of
the genera Parastenocaris and Elaphoidella in particular,
while the tiny amphipod Niphargobates exhibits strong
and curved legs and claws appropriate for climbing in
less narrow crevices.
Thermal waters are a trophically particularly inhos¬
pitable environment. They force higher metabolism
intensities and, because they are purified as they make
their way through greater earth depths, they offer par¬
ticularly little food and little oxygen. In hypothermal
waters of 15—28°C on the northwestern edges of the
Dinaric karst, some stenasellid isopods, such as the
thermophile race of Protelsonia hungarica Mehely and
tiny gastropods Hadziella thermalis Bole, are present. In
Slovenia, close to the Pleistocene Alpine glacier, stena-
sellids were found only in such an environment. It
seems, in fact, that biota are limited to trophically less
inhospitable springs, because the filtration of huge
quantities of water from depths gave no results.
Sinking streams are streams that, after some course
on the surface, flow underground. The subterranean
bed of a sinking stream is, in fact, an ecotone environ¬
ment. Ecological conditions in them are more fluctuat¬
ing, are richer in food resources, and more diverse
than in "autogenous" waters. Some surface species
may penetrate along such streams for shorter or longer
stretches underground. Among them are adaptation-
prone eutroglophile species such as Asellus aquaticus.
Along such a subterranean bed, a rich assemblage of
selected surface and subterranean animals is formed;
the fauna changes its quantitative and qualitative com¬
position with gradual ecological changes leaving the
sinking point and the surface influences. Although
some troglobionts, for example, Troglocaris shrimp
and even Proteus and some Niphargus spp., find such
a rich environment particularly inviting, others strictly
avoid it. On the other hand, this is the only cave
environment where a number of insect larvae (of
Ephemeroptera, Plecoptera, and Trichoptera) may be
present.
POLLUTION AND PROTECTION
A number of threats face the Dinaric cave fauna. In
the nineteenth and early twentieth centuries, some ani¬
mals had been caught in large numbers for trading pur¬
poses. The amphibian Proteus angidnus was reportedly a
popular pet in parlor aquaria, while the rare cave beetles
won high prices among amateur entomologist-collectors.
Some dealers organized genuine chains of collectors
all along the Dinaric karst. Because they were using
baited traps they were potentially able to affect some
populations.
In the decades of economic development of the
traditionally poor and economically passive Dinaric
karst after the Second World War, a number of hydro-
technical projects have been accomplished. Surface
and underground dams and artificial tunnels changed
the hydrology of some parts of the Dinaric karst
remarkably. We may expect that some previously iso¬
lated population-races might come in contact and fuse
again or at least that some populations were "pol¬
luted" by repeated input of foreign genes. New com¬
petition assemblages may have formed, threatening
the existence of some species. The hydrological regime
of the underground in wide surroundings of those
constructions definitely changed, also changing the liv¬
ing conditions (and sometimes threatening the exis¬
tence) of cave populations and species. For example,
the live deposits of the Marifugia-tuta with its builder
and many interstitial guests is now dead, due to the
lack of regular long-lasting floods in Popovo Polje and
in the sink-cave Crnulja. This unique phenomenon has
been destroyed. Such is also the destiny of the massive
colony of the clam Congeria kusceri in Zira Jama.
Of course, the most serious threat to the cave biota
is the omnipresent and diverse pollution. The net-
shaped underground hydrological connections make
the direction of pollutant outflows unpredictable and
their effect very wide-reaching. In the case of the sink¬
ing river Pivka, a moderate amount of organic pollu¬
tion may enable surface animals to increase their
success in competitive situations to the detriment of
troglobionts. In this way, pollution may indirectly
extinguish cave faunas. However, spills of generally
poisonous materials are not a rare event in the karst
and only the fact that such spills are also detrimental
to the precious resource of potable water allows us to
efficiently protect the subterranean fauna.
However, as early as 1920, the Section for Nature
and Natural Monuments Conservation of the Slovene
Museum Society wrote and published a memorandum
that outlined a very complete and detailed nature con¬
servation plan. Beside the Alpine sites and biota, one of
its main subjects was specifically karst caves as a
whole as well as the particularly interesting cave fauna.
A paragraph about a total ban on the commercial
exploitation of cave fauna was included, saying that it
should only be exploited for scientific and educational
purposes. This was—very generally—realized in 1922.
This early attempt at nature protection, initiated in a
high degree by cave fauna, was followed by a number
of species protection acts and other environmental
ENCYCLOPEDIA OF CAVES
238
D
legislation in all Dinaric countries. So, the legal back¬
ground for the many-sided protection of this hotspot
exists—but so do many practical hindrances for its
practical implementation.
See Also the Following Articles
Dinaric Karst: Geography and Geology
Diversity Patterns in Australia
Diversity Patterns in Europe
Diversity Patterns in the Tropics
Diversity Patterns in the United States
Bibliography
Culver, D. C., & Sket, B. (2000). Hotspots of subterranean biodiver¬
sity in caves and wells. Journal of Cave & Karst Studies, 62, 11—17.
Hamann, O. (1896). Europdische Hohlenfauna. Jena: Hermann
Costenoble.
Sket, B. (1997). Biotic diversity of the Dinaric karst, particularly in
Slovenia: History of its richness, destruction, and protection.
In I. D. Sasowsky, D. W. Fong & E. L. White (Eds.), Conservation
& Protection of the Biota of Karst (pp. 84—98). Charles Town, WV:
Karst Waters Institute (Special Publication 3).
Sket, B. (1999). High biodiversity in hypogean waters and its endan-
germent—The situation in Slovenia, Dinaric karst, and Europe.
Crustaceana, 72(8), 767—779.
Sket, B., Paragamian, K., & Trontelj, P. (2004). A census of the obli¬
gate subterranean fauna in the Balkan Peninsula. In H. I.
Griffiths & B. Krystufek (Eds.), Balkan Biodiversity. Pattern and
Process in Europe's Biodiversity Hotspot (pp. 309—322). Dordrecht,
The Netherlands: Kluwer Academic Publishers.
Sket, B., & Zaksek, V. (2009). European cave shrimp species
(Decapoda: Caridea: Atyidae), redefined after a phylogenetic
study; redefinition of some taxa, a new genus and four new
Troglocaris species. Zoological Journal Linnean Society, 155(4),
786-818.
Valvasor, J. W. (1689). Die Ehre des Herzogthums Crain. Nuremberg,
Germany: W. M. Endtner.
DIVERSITY PATTERNS IN
THE TROPICS
Louis Deharveng and Anne Bedos
Museum National d’Histoire Naturelle de Paris
The intertropical belt has an ecological definition that
roughly matches its geodesic limits (i.e., between
23°27'N, the tropic of Cancer, and 23°27'S, the tropic
of Capricorn; Fig. 1). Tropical climate, ranging from
lowlands to the nival belt, can be characterized by a low
thermal amplitude across seasons and usually high
humidity, though deserts are also included in the belt
(parts of Saudi Arabia, southern Africa, Sahara,
Australia, and India). These climatic characteristics of
humid tropics are also those of cave microclimate, aside
from darkness. Caves are actually largely azonal habi¬
tats, with a very similar environment across all latitudes
and all macroclimates. Beyond local variations, tempera¬
ture is the most basic physical factor distinguishing abio¬
tic environments in tropical versus nontropical caves.
HISTORICAL CONTEXT
As early as 1914, Jeannel and Racovitza claimed
that troglobionts, though rare, were well-diversified in
the tropics. This statement was overlooked by other
authors for almost 60 years. Until the 1970s, authors
who worked on the cave fauna repeatedly stressed
that true subterranean animals were rare or absent in
the tropics (Annandale et al. 1913; Leleup, 1956). Why,
they argued, would animals develop adaptations to an
environment that is similar to that outside the caves in
its weak thermal amplitude and permanent high
humidity? In the same vein, it was speculated that
higher energy input in tropical caves would have less¬
ened the "selection pressures for energy-economizing
troglobite adaptations" (Mitchell, 1969), accounting for
the scarcity of some troglomorphic traits among cave
invertebrates of the tropics.
The idea that tropical troglobionts are scarce and
mostly limited to aquatic species was based on surveys
of a few tropical caves, mostly from Malaysia (Batu
Caves), Burma (Farm Caves), and Congo. The rediscov¬
ery that a rich troglobiotic community including highly
troglomorphic species might exist in the tropics was pro¬
vided by Flowarth (1973) based on a study of the fauna
of Hawaiian lava tubes. Since then, the presence of
troglobionts has been confirmed in other parts of the tro¬
pics, in volcanic as well as limestone caves (Chapman,
1980; Humphreys, 2000; Gnaspini and Trajano, 1994;
Deharveng and Bedos, 2000; Deharveng et al. 2009).
CURRENT STATE OF KNOWLEDGE
Cave-restricted species (troglobionts and guano-
bionts) continue to be discovered in virtually every
newly surveyed tropical karst, along with a profusion
of troglophilic and guanophilic forms. This rapid prog¬
ress should not hide the fact that most tropical karsts
have not yet been biologically investigated. Enormous
geographic (Figs. 1,2) and taxonomic gaps still hamper
our global understanding of tropical cave biodiversity.
Unevenness in taxonomic coverage is illustrated, for
instance, by the Mulu cave fauna in Sarawak, which
probably includes the largest number of named troglo¬
bionts in the tropics, but which has barely been sur¬
veyed for aquatic and terrestrial microarthropods.
Another striking example is the giant rhaphidophorid
crickets of Southeast Asia which, in spite of being the
Encyclopedia of Caves.
© 2012 Elsevier Inc. All rights reserved.
DIVERSITY PATTERNS IN THE TROPICS
239
FIGURE 1 Main documented spots of tropical subterranean biodiversity, anchihaline and interstitial habitats excluded. Cave systems cited
in Tables 1 and 2 shown as circles, others as triangles; in orange, more than 10 troglobiotic species; in blue, less than 10 troglobiotic species.
After Jubertliie and Decu (1994—2001), updated.
FIGURE 2 Main records of obligate subterranean species in Southeast Asia ( after Deharveng and Bedos, 2000, updated). Red spots: best stud¬
ied caves or karsts, with more than four obligate subterranean species. 1: Masan-dong (Mulun karst); 2: Halong Bay Caves; 3: Pu Luong karst;
4: Ke Bang karst; 5: Tham Thon (Khammouane karst); 6: Tan Phu lava tubes; 7: Hon Chong—Kompong Trach karst; 8: Vang Vieng; 9: Tham
Chiang Dao; 10: Farm Caves; 11: Tham Phulu; 12: Phang Nga Caves; 13: Batu Caves; 14: Niah Cave; 15: Air Jernih system (Mulu karst); 16:
Baai system (Sangkulirang karsts); 17: Sangki system (Gunung Seribu); 18: Gunung Sewu karst; 19: Gua Salukkan Kallang (Maros karst); 20:
Muna karst; 21: Batu Lubang (Halmahera); 22: Fakfak karst; 23: Montalban.
most important invertebrates in food webs of tropical
caves (Deharveng and Bedos 2000), remain unde¬
scribed in most cases. More generally, and in contrast
with temperate cave fauna, the majority of tropical spe¬
cies listed in the literature are undescribed and their
ecological status is often uncertain, due to the scarcity
of faunistic surveys in surface habitats. Moreover,
although the sampling of cave faunas is enough to
derive rough biogeographic patterns, it is usually
insufficient for the detection of evolutionary radiations
among tropical cave groups. Each time fine-grain stud¬
ies have been carried out, however, such radiations
associated with narrow endemism among species have
been recognized. If confirmed to be general, such a
pattern would lead to a considerable increase in esti¬
mates of species richness in tropical caves. In all these
ENCYCLOPEDIA OF CAVES
240
D
aspects, we are obviously at the dawn of the biological
exploration of tropical cave biodiversity.
SUBTERRANEAN HABITATS
A subterranean fauna is encountered in any kind
of consolidated or unconsolidated rocks, as long as
obscure, climatically stable, interconnected, and humid
voids are available, together with minimal food input
from the surface. The combination of void size available
for fauna—small versus large—and their contents—air,
freshwater, or salty water—allows the recognition of the
basic types of subterranean habitats. In addition, terres¬
trial cave habitats can be further split into low-energy
and high-energy, based on the nature and amount of
food resources.
Traditionally, terrestrial interstitial habitats, that is,
the deep soil—the richest compartment of the hypogean
domain—is not dealt with by subterranean biologists.
All other habitats (terrestrial low-energy and high-
energy cave, aquatic interstitial, aquatic cave, marine/
anchihaline interstitial, marine/anchihaline cave) are
relevant to subterranean studies. They host their own
fauna, which has diversified according to patterns that
differ in temperate and tropical regions.
Aquatic Interstitial
Aquatic interstitial communities may be very diverse
and are richer in species than the noninterstitial aquatic
fauna of caves. Of particular interest is the epikarstic
component of this fauna, which is brought into caves
by dripping water from stalactites and is thus relatively
easy to study in terms of composition and dynamics.
The interstitial fauna is mostly composed of minute
microcrustaceans, with copepods being dominant. In
the tropics, however, most of our knowledge on this
fauna comes from nonkarstic, unconsolidated lowland
terrains, and the caves most intensively surveyed for
subterranean invertebrates have been at most barely
surveyed for their interstitial fauna.
Anchihaline and Marine Caves
Littoral karsts of the tropics and subtropics often
host a characteristic and rich fauna of anchihaline (or
anchialine) and marine cave species, which is much
more diversified than in temperate regions (Iliffe 2000).
When present, anchihaline/marine species may locally
account for the largest part of subterranean biodiversity,
especially for oceanic islands. At the extreme, the
Bahamas probably do not have a single terrestrial tro-
globiont, whereas more than 45 anchihaline stygobionts
are listed from its submarine passages, including the
majority of the 24 known species of Remipedia.
Globally, however, the anchihaline fauna does not rep¬
resent the most important component of tropical subter¬
ranean biodiversity, because of its narrow geographic
range. Many favorable coastal karsts of Southeast Asia
(Phang Nga Bay in Thailand, Halong Bay in Vietnam),
have no or very few anchihaline/marine species
described, whereas their freshwater cave fauna is rich.
So far, the anchihaline fauna has been mostly sampled
in neotropical islands and, more recently, in Cape
Range (Australia), but it remains poorly investigated in
Southeast Asia, the Pacific, and Africa. Several taxa
linked to anchihaline/marine caves of the tropics are
remarkable by their phyletic isolation (Remipedia,
Mictacea), or affinities with deep-sea species (e.g.,
Thetispelecaris reinex, from the Bahamas, in the order
Bochusacea). Other anchihaline stygobionts of the tro¬
pics include fish and a number of Crustacea, of which
several have representatives both in anchihaline and
freshwater habitats.
Cave Freshwater
Cave freshwater habitats typically host large-sized
stygobionts and stygophiles. Microcrustaceans, which
are typical of interstitial habitats, are also represented,
either as permanent populations or brought in from
the epikarst by percolating water. The largest fresh¬
water stygobionts have been well studied in several
tropical regions, such as Cuba, Thailand, and South
Sulawesi, but data are lacking for most karsts of the
tropics, including Central America and several large
areas of Southeast Asia. In the tropics, this fauna is
often rather elusive, but may be found locally as high-
diversity assemblages. As a rule, phreatic water in cave
passages hosts the richest communities of stygobionts,
in contrast to cave puddles and streamlets (even endog¬
enous) that are often devoid of large stygobionts. The
richest stygobiotic fauna of continental Southeast Asia is
found in the food-poor phreatic waters of Tham Phulu
(Thailand), which host no less than six highly trog-
lomorphic species that can co-occur in a single puddle:
Dugesia deharvengi, Heteroclwetella glandularis, Theosbaena
cambodjana, Aecjidgidiella aquilifera, Stenasellus rigali, and
Siamoporus deharvengi).
Aquatic habitats rarely show the striking contrast
between high- and low-energy habitats observed in
tropical terrestrial habitats, because guano that falls
in running water is readily dispersed. When present,
however, guano-rich puddles may host a few stygo¬
bionts, but there is no evidence that any are restricted
to this habitat.
A characteristic feature of aquatic subterranean
habitats in temperate regions is the overwhelming
ENCYCLOPEDIA OF CAVES
DIVERSITY PATTERNS IN THE TROPICS
241
dominance of Crustacea and low representation of
insects: the same is seen in the freshwater of tropical
caves. However, the dominant taxonomic groups are
different. In particular, stygobiotic crabs are totally
absent in temperate caves, whereas they are frequent
and diversified in the tropics (Guinot 1994).
Tropical freshwater stygobionts are often related to
anchihaline or marine taxa. The recently described,
highly troglomorphic freshwater hymenosomatid crabs
of Kalimantan and Sulawesi (Fig. 3) belong to an
essentially marine family, and the same pattern is seen
in many tropical caves Caridea. In other cases, such as
Stenasellus isopods, the marine origin is more remote
and all members of the family are freshwater stygo¬
bionts, ranging from Southeast Asia to Africa and
Europe. At least a few species of oniscid isopods
('Thailandoniscus annae from Thailand and several unde¬
scribed Sinoniscus from southern China) and several
freshwater beetles are secondary stygobionts, having
colonized freshwater from terrestrial ancestors.
Oligotrophic Terrestrial Habitats and Troglobionts
Food-poor habitats are extensive in tropical caves.
These include all deep cave passages that are free of
bat and bird colonies, and out of reach of flooding by
exogenous streams. As in temperate regions, these
habitats are typically home to troglobiotic inverte¬
brates, with a low representation of troglophilic species
and an absence of guanobionts. Food resources consist
of scattered feces of bats, birds, and crickets; roots;
organic matter brought down by percolating water;
and old debris from exceptional flooding episodes.
Cave-restricted troglobionts, though usually rare,
are better known than other categories of tropical cave
fauna because their unusual morphology, odd biology,
frequent relictual status, and rarity attract the interest
of biologists.
FIGURE 3 Sulaplax ensifer Naruse, Ng, and Guinot, 2008 from a
cave of Muna island (Sulawesi). Photo courtesy of Tohru Naruse. Used
with permission.
The tropical faunas of troglobionts exhibit striking
local variations in their taxonomic composition, but
they also share common characteristics across the
whole intertropical belt. Local variations concern all
groups, in particular the dominant ones: millipedes,
woodlice, springtails, and crickets. Haplodesmidae
millipedes (Fig. 4) are, for instance, highly diversified
and often abundant in Southeast Asia caves, but absent
from tropical America and Africa caves, where no
millipede radiation has been recorded so far. Crickets,
which often constitute the basis of invertebrate food webs
in the tropics, exhibit a clear-cut distribution pattern:
Rhaphidophoridae (Gryllacridoidea) in Southeast Asia
and Phalangopsidae (Grylloidea) in Africa and tropical
America. Patterns in troglobiotic springtails are also
sharply contrasted, but at a finer level; in Southeast Asia
the dominant troglobiotic groups are distributed as a com¬
plex mosaic: Sinella (Entomobryidae) in southern China
and northeastern Thailand, Lepidonella (Paronellidae) in
Vietnam, Troglopedetes (Paronellidae) in western Thailand,
Cyphoderopsis (Paronellidae) from southern Thailand to
Java, and Pseudosinella (Entomobryidae) in Sulawesi
and Papua. Needless to say, the cause of these complex
biogeographical patterns is unknown.
On the other hand, several groups have troglobiotic
representatives across large parts of the tropics, whereas
they are absent or very rare in temperate caves. Examples
include scorpions and various poorly studied spider fam¬
ilies among the predators, nocticolid cockroaches (Fig. 5)
and paronellid Collembola among the decomposers, or
fulguromorph Hemiptera among the root-feeders. The
absence or rarity of groups that are highly diversified in
northern temperate regions is a further characteristic of
FIGURE 4 Radiation of the millipede genus Eutrichodesmus,
illustrated by four recently described species from Vietnamese caves
(Golovatch et ah, 2009): (A) Eutrichodesmus regularis; (B) E. armatocaudatus;
(C) E. asteroides; (D) E. aster. Photos bi/ Louis Deharveng.
ENCYCLOPEDIA OF CAVES
242
D
FIGURE 5 Nocticolidae cockroach from Saripa Cave in Maros
(Sulawesi). Photo courtesy of Jean-Yves Rasplus. Used with permission.
the tropical troglobiotic fauna, which is shared with
southern temperate faunas. Examples of such groups are
Onychiuridae and Hypogastruridae among springtails,
and several groups of beetles.
At a finer scale, the level of geographic diversification
of troglobiotic taxa across fragmented karst landscapes
is of prime importance for biodiversity evaluation
and conservation. No study has directly addressed this
issue for the tropics, but indirect information is available
from various taxonomic and faunistic studies concern¬
ing the Hawaiian cave fauna as well as radiations of
Trechinae beetles, Telemidae spiders, Haplodesmidae
and Glomeridae millipedes in southern China, and
those of Troglopedetes springtails in western Thailand.
All these data indicate that levels of endemism are prob¬
ably higher in tropical than in temperate caves. Besides
high levels of endemicity, two characteristics of troglo¬
biotic taxa interest biologists: their often highly modified
morphology (troglomorphy) and their frequent phyletic
isolation (relictness). Both features are discussed sepa¬
rately below.
Guano and Guanobionts
A considerable proportion of cave terrestrial
communities in the tropics are dependent on guano,
produced by bats and locally by swiftlets (Fig. 6)
or other birds. Troglobionts exploit scattered bat or
swiftlet droppings, or even small guano patches, but
they avoid guano accumulations, which are inhabited
by guanophiles or guanobionts, often referred to as
troglophiles on a morphological basis. Guanobionts
complete their whole life cycle in or around piles of
guano, and should be rather considered as a category
of troglobionts as suggested by Chapman (1986) and
Gnaspini (this book), even if they usually lack troglo-
morphic characters.
The arthropods foraging on and around guano
piles are diverse in terms of their size and abundance.
Giant arthropods (tailless whip-scorpions, Sparassidae
spiders (Fig. 7), scutigerid centipedes, crickets, cock¬
roaches) are the most conspicuous and, though not
strictly linked to guano, they are frequent on the walls
of most high-energy caves in the tropics. The guano
microarthropods, comprised of a multitude of medium
to small species (mites, springtails, millipedes, beetles,
flies, and moth larvae), are present as huge popula¬
tions foraging on and in guano piles, where guano¬
bionts and more opportunistic species are mixed. The
guano fauna includes both widespread species (often
pantropical, e.g., the springtail Xenylla yucatana, or
even cosmopolitan, e.g., the cockroach Periplaneta amer-
icana) and species with more restricted ranges. The lat¬
ter category might be more important than previously
thought, as shown by the recent discovery of a huge
radiation of typically guanobiotic Cambalopsidae milli¬
pedes in Southeast Asia, each karst having one or two
endemic species (Fig. 8).
Though several of the dominant groups in guano-
related habitats are known across the whole tropics
(Gnaspini, this book), regional differences exist for
others. In tropical Asia for instance, the dominant
medium-size arthropods on guano piles are
Cambalopsidae millipedes, and giant Sparassidae
spiders (Fig. 7) are frequent on cave walls of food-
rich caves, whereas both taxa are absent from the
guano caves of Africa and tropical America.
Troglophilic and Stygophilic Species
The most abundant and most conspicuous animals
of tropical caves are species that are not strictly
linked to subterranean environments but spend part
of their life in caves, either seasonally or daily (sub-
troglophiles), or that have both cave and non-cave
populations (eutroglophiles). These troglophiles are
present in all the habitats listed above. Many taxa
cited in the Encyclopaedia Biospeologica (Juberthie and
Decu, 1994—2001) for tropical countries are troglo¬
philes, accounting for the largest part of the fauna
encountered in caves, far ahead of troglobionts.
First among troglophiles are bats (often classified
as trogloxenes), with many species restricted to the
tropics. Tropical caves commonly host very large bat
colonies; two million individuals of Tadarida plicata
live, for instance, in Khao Chong Pran Cave in
Thailand. However, most bat species live as scattered
individuals or small groups in caves. Taxonomic and
functional diversity of cave bats are much higher in
the tropics than in temperate regions. In addition to
the classical insectivorous bats, several large fruit-bats
ENCYCLOPEDIA OF CAVES
DIVERSITY PATTERNS IN THE TROPICS
243
FIGURE 6 Swiftlets. Left: young, Ngalau Indah near Lintau Buo (Sumatra) ( Photograph by Louis Dehcirveng). Right: female in its nest, Santo
(Vanuatu). Photo courtesy of Bernard Lips. Used with permission.
FIGURE 7 Heteropoda sp. eating a giant cave cockroach Miroblatta
baai Grandcolas (Baai Cave, Kalimantan). Photo by the authors.
often roost in tropical caves; blood-sucking and fish¬
eating species are even known in caves of tropical
America. Morphological diversity is also striking.
Among others, tropical bats include one of the smallest
mammals in the world ( Craseonycteris thonglongyai,
from Thailand) and the strange naked bat Cheiromeles
torquatus, from Sunda islands.
Bats are not the only troglophilic flying vertebrates.
Several birds are linked to caves in the tropics. The
most regular and conspicuous ones are different species
of swiftlets in Southeast Asia, and the oilbird ( Steatornis
caripensis) in tropical America. Like bats, swiftlets
(Fig. 6) may live in large colonies, which are often
managed and exploited for their nests by local villagers
in a remarkably sustainable way. Non-flying troglo-
philes are much rarer among vertebrates, the most
conspicuous and frequent in Southeast Asia being the
bat-eating snake Orthriophis taeniurus ridleyi.
Bats and birds can provide the basis of cave food
webs, as they import, through their feces, most of the
food resources exploited by terrestrial and, to a lesser
extent, aquatic cave invertebrates. Like other caverni-
coles, troglophilic species exploit this widespread
resource, as well as debris brought into caves by flooding.
Several of the largest species of crickets, tailless whip-
scorpions, and giant Sparassidae spiders are known (or
assumed) to leave caves every night for feeding, but this
is certainly not the case for populations living in deep
recesses of the caves. Actually, in spite of the widespread
occurrence and conspicuousness of these troglophiles,
very little is known about their biology. Regarding sea¬
sonal subtroglophiles, information is practically nonexis¬
tent. Small-sized species encountered in caves rather
qualify as eutroglophiles; they are often abundant on
cave guano or organic debris, as well as in surface soils.
They include numerous springtails and mites, various
beetles and bugs, but also larger species such as milli¬
pedes, spiders, woodlice, cockroaches, and ants.
Stygophilic species are also numerous in tropical caves,
including fish and various invertebrates. Large-sized and
big-eyed Macrobrachium (Palaemonidae) are the most con¬
spicuous aquatic animals in many caves of Southeast
Asia; they are common even far inside caves, usually
excluding stygobionts from the habitats they occupy.
BROAD-SCALE PATTERNS OF
TROGLOBIONT SPECIES RICHNESS
Geographical Patterns of Species Richness
The uneven sampling of tropical caves and biased
taxonomic coverage of their fauna result in a heteroge¬
neous data set, from which some general patterns of
species richness can however already be drawn
ENCYCLOPEDIA OF CAVES
244
D
FIGURE 8 The distribution of Cambalopsidae millipedes in Southeast Asia (synthetized from the 2007—2009 works of Golovatch and colla¬
borators). Each dot represents a locality (usually a guano cave, rarely a surface habitat locality) with one or two endemic species. Distribution
data for the tramp species Glyphiulus granulatus are not displayed on the map. Empty triangles correspond to unidentified species.
1. Richness in troglobionts or guanobionts is rarely
very low in the tropics, and never null as in
temperate regions affected by glaciations. However,
it varies considerably from place to place, and
allows the recognition of both hot- and cold-spots
of species richness (Table 1).
2. The terrestrial fauna appears to be more diverse
in the Oriental and Australian regions than in the
Neotropics or Africa. This is obvious from the list
of cave species given by country in the Encyclopaedia
Biospeologica (Juberthie and Decu, 1994—2001), and
from updated data for the richest tropical caves,
summarized in Table 1. Paucity in troglobionts is
observed in many countries of tropical America,
with the possible exceptions of highlands
(Guatemala, Venezuela) and some subtropical
regions of Mexico (Reddell 1981). The pattern is
different for stygobionts, with comparatively higher
richness in the caves of tropical America, Africa,
and Australia than in Southeast Asia. Uneven
sampling effort might, however, explain much of
the difference, as illustrated by recent findings such
as remarkable stygobiotic Hymenosomatidae crabs
of Kalimantan and Sulawesi (Fig. 3).
3 . Locally, large karst blocks tend to be richer in
troglobionts than smaller ones, having a statistically
wider diversity of habitats. In Sarawak, the Air
Jernih system in the Mulu karst has significantly
more cave-restricted species than the Niah Cave
in the small isolated hill of Gunung Subis (Table 1).
However, this does not necessarily apply to
guanobionts, as illustrated by their high diversity
in caves developed in small limestone outcrops
such as the Batu Caves or Niah Cave in Malaysia.
4. "Most diversity in caves is expressed regionally
rather than locally" (Culver and Sket, 2000). This low
alpha- versus high beta-diversity reflects the high
degree of micro-endemism among subterranean
groups. The available evidence suggests that the
impact of fragmentation on speciation might be
stronger in tropical than in temperate karsts. A good
illustration is the radiation of the guanobiotic
Cambalopsidae millipedes that has been recently
found in Southeast Asia caves (Fig. 8) with more
than 50 species. The absence of MSS ("Milieu
Souterrain Superficiel") in the tropics, except in
some volcanic terrains, certainly contributes to
accelerating diversification by limiting gene flow
between populations of different karst units.
5 . The richest hotspot of the tropical subterranean
fauna today is the Maros karst in South Sulawesi,
with 28 troglobionts and stygobionts in the
ENCYCLOPEDIA OF CAVES
ENCYCLOPEDIA OF CAVES
TABLE 1 Biodiversity and Faunistic Composition of Some Well-Studied Tropical Caves
SKT
System
(SUL)
Air
Jemih
System
(MAL)
Sangki
System
(SUM)
Ma
San
Dong
(CHI)
Tham
Chiang
Dao
(THA)
Batu
Lubang
(HAL)
Areias
System
(BRA)
Tham
Non
(LAO)
Tham
Thon
(LAO)
Hang
Mo So
(VIET)
Hang
Nuoc
(VIET)
Kulumuzi
Cave
(TAN)
Siju
Caves
(ASS)
Batu
Caves
(MAL)
Olhos
d'Agua
Cave
(BRA)
Farm
Caves
(MYA)
Niah
Great
Cave
(SAR)
Mollusca
1
Hirudinea
1
Arachnida
Acari
1
1
1
1
1
1
1
Araneae
3
5
2
1
2
4
1
2
3
1
Opilionida
1
1
1
1
1
1
1
2
1
1
1
1
1
1
Palpigradida
1
1
Pseudoscorpionida
1
1
1
1
1
1
1
Schizomida
1
1
1
1
1
1
1
Scorpionida
1
Myriapoda
Chilopoda
1
Diplopoda
Glomeridesmida
1
Diplopoda
Glomerida
1
1
1
1
Diplopoda Julida
1
1
1
1
1
Diplopoda
Spirostreptida
1
1
1
Diplopoda
Polydesmida
2
2
2
2
1
2
3
2
2
1
1
1
Diplopoda
Stemmiulida
1
1
Diplopoda
Callipodida
1
Crustacea
Isopoda Oniscidea
3
2
3
2
2
3
2
2
1
1
3
1
2
2
Collembola
Entomobryomorpha
2
1
2
2
2
2
3
1
2
2
3
1
1
1
Poduromorpha
1
2
Symphypleona
1
1
1
1
1
2
Diplura
Campodeidae
1
1
1
1
Zygentoma
Nicoletiidae
1
Insecta
Blattodea
1
1
1
1
1
1
1
( Continued )
ENCYCLOPEDIA OF CAVES
TABLE 1 ( Continued )
SKT Air Sangki Ma Tham Batu Areias Tham Tham Hang Hang Kulumuzi Siju Batu Olhos Farm Niah
System Jernih System San Chiang Lubang System Non Thon Mo So Nuoc Cave Caves Caves d'Agua Caves Great
(SUL) System (SUM) Dong Dao (HAL) (BRA) (LAO) (LAO) (VIET) (VIET) (TAN) (ASS) (MAL) Cave (MYA) Cave
(MAL) (CHI) (THA) (BRA) (SAR)
Coleoptera 1
Carabidae
Coleoptera
Pselaphidae
Coleoptera
Leiodidae
Dermaptera
Hemiptera
Hymenoptera
Formicidae
1 12 11
1 1 1
1
1 1
1
1
Orthoptera
1
1
1
1
Troglobionts
21
15
17
15
15
15
13
16
15
14
12
10
7
4
4
4
1
Turbellaria
1
1
1
Mollusca
1
1
Crustacea Amphipoda
1
1
1
1
1
1
(Bogidiellidae)
Copepoda 1
Decapoda Anomura 1
(Aeglidae)
Decapoda 11 1 1
Brachyura
Decapoda Natantia 3 1 11
Isopoda Anthuridae 1
Isopoda Oniscidea 1 2
Isopoda 1 1
Stenasellidae
Ostracoda
1
Syncarida 1 1
Pisces
1
1
1
Stygobionts
7
6
3
2
1
1
3
0
1
0
1
0
1
3
3
0
2
Troglobionts + stygobionts
28
21
20
17
16
16
16
16
16
14
13
10
8
7
7
4
3
ASS: Assam (India); BRA: Brasil; CHI: Guangxi (China); HAL: Halmahera (Indonesia); LAO: Laos; MAL: Malaya (Malaysia); MYA: Myanmar; SAR: Sarawak (Malaysia); SUL: Sulawesi (Indonesia);
SUM: Sumatra (Indonesia); TAN: Tanzania; THA: Thailand.
DIVERSITY PATTERNS IN THE TROPICS
247
Salukkan Kallang-Tanette (SKT) system alone
(Deharveng and Bedos 2000). Why such a high
diversity in a lowland cave, whereas the best-
known surrounding regions ( e.g ., Java) are only
moderately rich? The unique conjunction of four
independent factors may explain this exceptional
richness: seasonal local climate, proximity to the sea
(several species present in Salukkan are considered
of marine origin), situation at the foot of relatively
high mountains (a potential source of cave
colonizers), and intensive sampling. Surveys of
other tropical karsts will tell us in the future
whether the latter was the predominant factor.
Troglomorphy along Environmental Gradients
Gradients that structure cave fauna operate at
different scales. At the local scale, inside a single
underground system or karst unit, physicochemical
factors (light, salinity, substrate granulometry, and
food resources) are associated with striking changes
in assemblage composition and troglomorphy level,
operating on very short ecotones (often a few meters
or less). There is strong evidence in particular that
troglomorphy is linked to scarcity of food resources.
Even more obviously than in temperate caves, high
degree of troglomorphy and high diversity of troglo-
morphic species are only observed in oligotrophic
habitats while high-energy habitats are only exploited
by non- or weakly troglomorphic species.
Elevation is a major structuring gradient for the
surface fauna. Because deep caves of the same system
have a very similar microclimate whatever their alti¬
tude, the expectation is that the cave fauna will be less
dependent on altitudinal gradient. In fact, the decline
of bats and guanobionts with elevation is paralleled by
important changes in the balance between ecological
categories of cave fauna, and an increase in troglo¬
morphic traits among troglobionts. Thus, outside sub¬
tropical China, troglomorphic beetles are exceptionally
encountered in lowland of tropical regions but regu¬
larly found in highland caves, such as Mayaphaenops
sbordonii at 3000 m in Guatemala, Speleodesmoides raveloi
at about 3200 m in Venezuela, and several species in
highlands of New Guinea.
Latitudinal gradients of seasonality and aridity are
the main drivers of changes in surface ecosystems.
Subterranean environments, because of their buffered
microclimate, seem to be relatively isolated from these
changes. Cross-checking of taxonomic and faunistic
information allows us to recognize that troglomorphy
increases with increasing seasonality, especially along
a latitudinal gradient. This pattern is observed in vari¬
ous zoological groups and concerns both the number
of troglomorphic species and the degree of modifica¬
tion of their morphological traits. A good example is
that of subterranean beetles in Southeast Asia, which
are few and weakly modified morphologically in the
lowland humid tropics (two species in Indonesia,
none in Malaysia), more numerous and more troglo¬
morphic in eastern and northern Thailand (four or
five species), and finally reaching the highest levels
of troglomorphy and diversity in southern China, just
north of the tropical belt. A similar pattern of increas¬
ing troglomorphy is seen among the most characteris¬
tic and frequent terrestrial cave groups of the region,
that is, Rhaphidophoridae crickets, Cambalopsidae
millipedes, and springtails. Interestingly, fish follow
the same gradient but stygobiotic invertebrates do
not: reduced-eyed and long-legged species of crabs
and shrimps are no more frequent in the subtropics
than near the equator (Guinot 1988).
Tropical versus Temperate Subterranean
Biodiversity
The tropics are known to have the highest biodiver¬
sity on Earth for surface ecosystems, but this might
not be true for subterranean habitats. Culver and Sket
(2000) compared the specific richness of 20 caves
and wells worldwide having 20 or more obligate subter¬
ranean species and came to the conclusion that "the scar¬
city of high diversity caves in the tropics is still
a puzzle" (p. 16). Their study included two tropical
caves, the SKT system in South Sulawesi, and the Bayliss
lava tube in northern Australia. These caves ranked 12
and 15 out of 20 for total species richness, far behind the
temperate cave systems of the Dinaric karst. The data
we provide here (Table 2) concern a larger array of trop¬
ical caves from various regions, including the best-stud¬
ied and richest ones, and it seems to confirm Culver and
Sket's observation. However, aquatic microinvertebrates
(Copepoda, Oligochaeta), which contribute the most bio¬
diversity in temperate caves, have not been studied in
SKT nor in most tropical caves. A more relevant com¬
parison, based on terrestrial fauna alone, places these
two caves much closer to the richest temperate caves
(ranking 4 and 5 out of 21 in Table 2). They would rank
even higher if the obligate subterranean guanobiotic spe¬
cies, which are considerably more diverse in the tropics
than in temperate areas, were taken into account.
RELICTUAL VERSUS
N ONRELICTU AL TAXA
Distributional relicts represent a precious component
of biodiversity for the paleobiogeographic and phyletic
ENCYCLOPEDIA OF CAVES
248
TABLE 2 Tropical
D
versus Temperate Cave Biodiversity
Bioclimate
Aquatic
Terrestrial
Total
Ranked on Terrestrial
Ranked on Total
Europe: Pyrenees
Baget system (FR)*
Temperate
24
9
33
18
8
Goueil di Herr (FR)*
Temperate
14
12
26
15
10
Europe: Italy
Busso del Rana*
Temperate
15
5
20
20
14
Grotta dell'Arena *
Temperate
6
14
20
13
14
Europe: Dinaric karst
Grad (SLO)*
Temperate
17
3
20
21
14
Jama Logarcec (SLO)*
Temperate
28
15
43
10
5
Krisna Jama (SLO)*
Temperate
29
16
45
8
4
Sica-Krka system (SLO)*
Temperate
27
7
34
19
7
Sistem Postojna-Planina (SLO)*
Temperate
48
36
84
1
1
Vjetrenica Jama (Bosnia)*
Temperate
39
21
60
5
2
Europe: Romania
Pestera de la Movile*
Temperate
18
29
47
2
3
USA
Mammoth Cave*
Temperate
15
26
41
3
6
Shelta Cave*
Temperate
12
12
24
15
11
China
Feihu Dong
Sub-temperate
4
16
20
8
14
Ma San Dong
Sub-tropical
2
15
17
10
19
Australia
Bayliss Cave*
Tropical
0
24
24
4
11
Southeast Asia
Sangki system (SUM)
Tropical
3
17
20
7
14
Air Jernih system (SAR)
Tropical
6
15
21
10
13
SKT system (SUL)
Tropical
7
21
28
5
9
Africa
Kulumuzi Cave (TAN)
Tropical
0
10
10
17
21
America
Areias system (BRA)
Tropical
3
13
16
14
20
Data are indicative, as ecological status of troglobiont may vary slightly between authors. Abbreviations are as in Table 1. FR: France; SLO: Slovenia.
Mata from Culver and Sket (2000).
information they carry. They are more numerous in
caves than in any surface habitat, as the buffered micro¬
climate of subterranean ecosystems is assumed to have
protected cave biota from environmental fluctuations
that often led surface fauna to extinction. Most cave
relicts are troglobiotic species and most are found in
temperate regions. From the observation that cave
species of Hawaii are derived from the surface fauna
present on the islands, Howarth (1973) hypothesized
that the absence of relicts might be characteristic of trop¬
ical caves. Hawaii, however, as a young archipelago,
did not experience drastic ecological changes that would
have eliminated the surface fauna, thus turning cave
species into relicts. Actually, evidence is accumulating
ENCYCLOPEDIA OF CAVES
DIVERSITY PATTERNS IN THE TROPICS
249
that climatic relicts are also present in tropical caves,
often in large numbers. The Cape Range of western
Australia, thoroughly studied by Humphreys, is well
known in this respect, but many other karsts host relic -
tual species, particularly but not exclusively in seasonal
regions of the tropics such as Laos or the Maros karst in
South Sulawesi.
There are various degrees and kinds of relicts. In
the tropics, marine relicts (i.e., species secondarily
adapted to freshwater) often represent the largest part
of freshwater cave fauna. For instance, the crab
Cancrocaeca xenomorpha of Maros (South Sulawesi) or
the isopod Cyathura (Stygocyathurn) chapmani of the
Mulu caves are likely to have arisen recently from
marine ancestors, but they are now totally discon¬
nected spatially and ecologically from marine habitats.
Among the terrestrial fauna, disjunct distributions
also point to a relictual status. Trechinae, for instance,
have extensively radiated in caves of southern China to
produce more than 50 species, but their potential surface
relatives are now hundreds of kilometers away, in the
Himalayas. More often, relict species belong to oligo-
and monospecific genera. Thus, a troglobiotic millipede
from a cave of southern Vietnam, Eostemmiulus caecus,
was shown to belong to Stemmiulidae, a large family
previously unknown from Southeast Asia, but well
diversified in the rest of the tropics. Another example is
the recent finding, in the large Kebang-Khammouane
karst across Laos and Vietnam, of two troglomorphic
species that belong to two monospecific genera of
Pseudochactidae, a family of scorpions described from
central Asia in 1998 based on a single species (Fig. 9). It
is actually no surprise that tropical karsts, affected by
bioclimatic changes and sea-level fluctuations, have gen¬
erated various relictual taxa. The overall picture is, how¬
ever, of a lower diversity of relicts than in temperate
caves, in line with the lower intensity of bioclimatic fluc¬
tuations in the tropics, as stressed by Chapman (1986).
FIGURE 9 Vietbocap canhi Lourenfo and Pham (Pseudochactidae),
a blind scorpion discovered in 2010 from a cave in the Phong
Nha-Ke Bang National Park (Vietnam). Photo courtesy of Elise-Anne
Leguin. Used with permission.
CONSERVATION ISSUES SPECIFIC
TO TROPICAL CAVES
Karst and cave fauna are seriously threatened by
human activities in the tropics, from pollution to
uncontrolled tourism and limestone exploitation. Like
other ecosystems, they have to face a growing demand
on natural resources generated by the rapid develop¬
ment of southern countries, in a context of less
stringent environmental regulations than in western
countries. These threats are aggravated by two charac¬
ters of tropical cave fauna: the importance of bats, and
the high proportion of narrow endemics.
Bats, Guano, and Biodiversity Conservation
It has long been more or less implicitly admitted that
guano hosted less biodiversity, or less interesting biodi¬
versity than low-energy habitats, and in particular a low
number of narrow endemics. This view is, however, chal¬
lenged by the recent discovery of a high Cambalopsidae
radiation in Southeast Asia caves (Fig. 8).
Mining, guano exploitation, hunting, and tourism
all have a detrimental effect on bat and swiftlet popu¬
lations. Indirectly, guano deposits, and hence their
associated fauna, are affected. As common bats such as
Tadarida plicata in tropical Asia are the largest provi¬
ders of cave guano in subterranean ecosystems, threats
on their colonies are paradoxically more worrying for
overall biodiversity than threats on rarer non-colonial
species.
When disturbance leads to bats abandoning a
roosting site, nonphoretic guanobionts will not survive
the disappearance of guano, as they cannot escape by
active dispersal, and contrary to troglobionts, they can¬
not find refuge in deeper cracks of the rocks because
there is no guano there. Given the vulnerability of bats
and the huge contribution of guano species to overall
cave diversity in the tropics, bat disturbance is proba¬
bly today the most worrying threat to tropical subter¬
ranean biodiversity.
Endemism and Vulnerability
Caves host a larger proportion of endemics than
any surface habitat, and cave endemics are on average
more narrowly distributed than surface endemics.
However, this remains poorly documented for tropical
karsts. Rapid weathering of limestone under tropical
climate often results in dramatic landscape of more
or less isolated karst blocks. During this process,
cave communities are fragmented and may genetically
diverge. Recent studies have shown that even very
small limestone blocks, when separated from each
ENCYCLOPEDIA OF CAVES
250
FIGURE 10 A critical issue in karst biodiversity conservation: the
Hon Chong hills in southern Vietnam (from Google map, photo
dated 2006). Limestone hills are outlined in white; non-limestone
hills are not outlined. Nui Khoe La and Nui Bai Voi will be quarried
flat, except in the extreme north of Nui Bai Voi which is being devel¬
oped as a tourist spot. Both hills host strictly endemic arthropods,
especially in their cave systems. Of the other hills to the south, only
Nui Hang Tien will remain unexploited.
other by a few kilometers of unfavorable habitat,
may host narrowly endemic species in their under¬
ground arthropod fauna and very local relicts differing
between karst blocks. More relictual and phyletically
isolated taxa of arthropods have been recorded in the
tiny hills of the Hon Chong karst in southern Vietnam,
for instance, than in any other large karst of the coun¬
try. Quarrying such small hills to the ground, as is
being done in several regions of Malaysia and Vietnam
(Fig. 10), and probably elsewhere, is the fastest way to
lead to the extinction of a large number of the rarest
subterranean microendemic species of invertebrates.
The selection of quarrying and mining sites ought first
to take into account the size and isolation of the eco¬
system that will be impacted, as no remediation will
be possible afterwards. This should be the cardinal
rule for a sustainable exploitation of limestone in the
tropics in terms of its impact on biodiversity.
See Also the Following Articles
Diversity Patterns in Australia
Diversity Patterns in the Dinaric Karst
Diversity Patterns in Europe
Diversity Patterns in the United States
Bibliography
Annandale, N., Brown, J. C., & Gravely, F. H. (1913). The limestone
caves of Burma and the Malay Peninsula. Asiatic Society of Bengal,
9, 391-423.
Chapman, P. (1980). The biology of caves in the Gunung Mulu
National Park, Sarawak. Transactions of the British Cave Research
Association, 7, 141—149.
Chapman, P. (1986). Non-relictual cavernicolous invertebrates in
tropical Asian and Australasian caves. In Ibynsa C./Badajoz
145—147 Barcelona (eds): Proc. 9th Int. Cong. Speleol., Barcelona
(Spain), 2: 161-163.
Culver, D. C., & Sket, B. (2000). Hotspots of subterranean biodiver¬
sity in caves and wells. Journal of Cave & Karst Studies, 62, 11—17.
Deharveng, L., & Bedos, A. (2000). The cave fauna of Southeast Asia.
Origin, evolution and ecology. In H. Wilkens, D. C. Culver &
W. F. Humphreys (Eds.), Ecosystems of the zvorld: 30 subterranean
ecosystems (pp. 603—632). Amsterdam: Elsevier.
Deharveng, L., Bedos, A., Le, C. K., Le, C. M., & Truong, Q. T. (2009).
Endemic arthropods of the Hon Chong lulls (Kien Giang), an unri¬
valed biodiversity heritage in Southeast Asia. In C. K. Le, Q. T.
Truong & N. S. Ly (Eds.), Beleaguered hills: Managing the biodiversity
of tire remaining karst hills of Kien Giang, Vietnam (pp. 31—57). Ho Chi
Minh City, Vietnam: Nha Xuat Ban Nong Nghiep.
Gnaspini, P., & Trajano, E. (1994). Brazilian cave invertebrates, with
a checklist of troglomorphic taxa. Reinsta Brasileira Entomologia,
38, 549-584.
Golovatch, S., Geoffrey, J. J., Mauries, J. P., & van den Spiegel, D.
(2009). Review of the millipede family Haplodesmidae Cook, 1895,
with descriptions of some new or poorly-known species
(Diplopoda, Polydesmida). In S. I. Golovatch & R. Mesibov (Eds.),
Advances in the Systematics of Diplopoda I. Zookeys 7, pp. 1—53.
Guinot D (1994). Decapoda Brachyura. In C. Juberthie & V. Decu
(Eds.), Encyclopaedia Biospeologica, voi. I, pp. 1991—2006. Socie'te'
de Biospe'ologie, Moulis, France.
Howarth, F. G. (1973). The cavernicolous fauna of Hawaiian lava
tubes. I. Introduction. Pacific Insects, 15, 139—151.
Humphreys, W. F. (2000). Relict faunas and their derivation. In
H. Wilkens, D. C. Culver & W. F. Humphreys (Eds.), Ecosystems
of the world: 30 subterranean ecosystems (pp. 417—432). Amsterdam:
Elsevier.
Iliffe, T. M. (2000). Anchialine cave ecology. In H. Wilkens, D. C.
Culver & W. F. Humphreys (Eds.), Ecosystems of the world: 30 sub¬
terranean ecosystems (pp. 59— 76). Amsterdam: Elsevier.
Jeannel, R., & Racovitza, E. G. (1914). Biospeologica XXXIII.
Enumeration des grottes visitees 1911—1913 (5 erne serie). Arch.
Zoologie Expe'rimentale et Generate, 53, 325—558.
Juberthie, C., & Decu, V. (Eds.), (1994—2001). Encyclopaedia biospeolo¬
gica (Vols. I—III). ). Moulis, France: Societe de Biospeologie.
Leleup, N. (1956). La faune cavernicole du Congo beige et considera¬
tions sur les coleopteres reliques d'Afrique intertropicale. Annates
du Muse'e du Congo Beige-Sciences Zoologiques, 46, 1 — 170.
Mitchell, R. W. (1969). A comparison of temperate and tropical cave
communities. Southwestern Naturalist, 14, 73—88.
Naruse, T., Ng, P. K. L., & Guinot, D. (2008). Two new genera and
two new species of troglobiotic false spider crabs (Crustacea:
Decapoda: Brachyura: Hymenosomatidae) from Indonesia, with
notes on Cancrocaeca Ng, 1991. Zootaxa, 1739, 21—40.
Reddell, J. R. (1981). Review of the cavernicole fauna of Mexico,
Guatemala, and Belize. Texas Memorial Museum Bulletin, 27,1—327.
ENCYCLOPEDIA OF CAVES
DIVERSITY PATTERNS IN THE UNITED STATES
251
DIVERSITY PATTERNS IN THE
UNITED STATES
Horton H. Hobbs III
Wittenberg University
INTRODUCTION
The scientific study of cave fauna (cavernicoles) in the
United States had its inception in 1842 with the
description by DeKay of the blind amblyopsid cavefish,
Amblyopsis spelaea, from Mammoth Cave, Kentucky. By
1888 the number of obligate cave-dwellers, that is, trog-
lobionts (terrestrial) and stygobionts (aquatic species),
had increased to more than 50, as reported by Alpheus
Packard in his compendium of North American cave
fauna (Packard, 1888); numerous additional descrip¬
tions of cave-restricted organisms occurred through the
1940s and 1950s. By 1960 Brother Nicholas listed 334
species (Nicholas, 1960), in 1998 Stewart Peck estimated
1353 cave species (including interstitial and unde¬
scribed species and subspecies; Peck, 1998), and in 2000
Culver et al. reported that a total of 973 species and
subspecies were described from caves in the 48 contig¬
uous states (Culver et al., 2000). Currently, 1138 species
and subspecies are described (inclusive of Alaska and
Hawaii) and are assigned to 239 genera and 112 families
(Table 1). Nearly all have evolved from the independent
invasion of surface organisms into the subterranean
realm where physical isolation and thus cessation of
gene flow with surface congeners has led to speciation.
In the stressed subterranean environment where there is
perpetual darkness and extremely limited food, selec¬
tive pressures have led to the evolution of parallel and
convergent regressive forms (profound morphological
alterations) that are characteristic of troglobionts and
stygobionts worldwide. This combination of characteris¬
tics, which includes reduction or loss of eyes and
pigments, gracilization and elongation of appendages,
increased chemical and tactile sensitivity, degeneration
of circadian rhythms, lowered fecundity and metabolic
rates, and increased longevity, is postulated to be adap¬
tive to life in such extreme ecosystems and is termed
troglomorphy.
This article summarizes the ecological, taxonomic, and
geographic patterns of biodiversity of cave-inhabiting
fauna in the 50 United States, focusing on troglobionts
and stygobionts (separated because the two environ¬
ments have very different environmental characteris¬
tics and faunas), but not ignoring cave visitors
(trogloxenes/stygoxenes) and "cave lovers" (troglo-
philes/stygophiles). Specific numbers of organisms
known from various cave areas (see below) are
TABLE 1 Summary of the Obligate Subterranean Phyla/Classes/
Orders, Families, Genera, and Species/Subspecies Described from
U.S. Caves
Phylum/Class/Order
Family
Genus
Species/
Subspecies
Turbellaria
4
6
28
Oligochaeta
4
6
13
Mollusca
5
16
29
ARACHNIDA
Acari
7
13
22
Aranaea
8
22
98
Opiliones
6
13
45
Pseudoscorpiones
9
29
152
Schizomida
1
1
1
Scorpiones
1
1
1
CRUSTACEA
Copepoda
2
3
7
Ostracoda
2
5
14
Bathynellacea
1
1
1
Thermosbaenacea
1
1
1
Amphipoda
11
18
137
Isopoda
6
15
101
DECAPODA
Shrimps
2
2
5
Crayfishes
1
4
35
Crabs
1
1
1
Chilopoda
3
4
4
Diplopoda
12
20
64
HEXAPODA
Thysanura
1
2
2
Diplura
2
3
8
Collembola
6
12
77
INSECTA
Orthoptera
1
2
3
Dermaptera
1
1
1
Hemiptera
2
2
2
Homoptera
1
1
7
Coleoptera
7
24
261
Diptera
1
1
1
Osteichtheys
2
5
6
Amphibia
1
5
11
Totals
112
239
1138
Encyclopedia of Caves.
© 2012 Elsevier Inc. All rights reserved.
252
D
derived mostly from a cave biota database that is
available on the World Wide Web at www.karstwaters.
org. No attempt is made herein to summarize the
protozoans or other microbial communities of caves
although clearly these are incredibly important in
cave ecosystems as sources of food (primary and sec¬
ondary production), in the deposition of minerals,
and in the process of speleogenesis. Additionally,
attention is directed toward the vulnerability as well
as to the challenges of conservation of this fauna and
associated habitats.
CAVE ECOLOGY
The obvious feature of caves (Hobbs, 1992) is perpetual
darkness, which results in the absence of green plants
(producers of carbon-based molecules and thus "food"
for organisms). Other characteristics of these food-poor
cavities are low variances in temperature and humidity
(usually near saturation), which make for a predictable,
less variable environment. Evolutionarily, as an organism
makes the transition to troglobiont or stygobiont, isolation
is another key feature of that environment. These are
unique combinations of characteristics that set caves
apart from other ecosystems and that make this extreme
environment inhospitable to most organisms.
The zonal variation in biological, chemical, and
physical properties influences the distribution and
abundance of fauna occurring in caves. The entrance
area of caves is the ecotone between the epigean and
hypogean worlds and has received only minimal
study. Clearly, here there is more diversity and greater
environmental variability than in any other cave zone.
On both horizontal and vertical scales, entrances pro¬
vide a transition of characteristics ( e.g ., temperature,
humidity, light) that may provide conditions for entry
or survival for preadapted species or for relictual spe¬
cies that otherwise are rare or have become extinct on
the surface due to climatic changes. They are impor¬
tant windows into the subterranean realm through
which pass migrating trogloxenes {e.g., bats, crickets)
and can be the point of entry of important organic
material. Farther into the cave but still within the lim¬
its of light penetration (the tivilight zone), the influence
of surface conditions is apparent and variation of
meteorological conditions is significantly less than at
the surface and entrance areas. The dark zone initially
demonstrates considerable influence of surface condi¬
tions but as distance increases from the entrance, it
grades into a much less variable environment that is
far from "constant" but significantly reduced in fluc¬
tuations of such parameters as atmospheric and water
temperatures. In the dark, deep interior of caves that is
characterized by more environmental constancy, virtu¬
ally no food is produced and organisms are thus
dependent on input of carbon from the surface that
energetically supports most cave ecosystems (plant
debris, bat and cricket guano). Exceptions to these gen¬
eralizations about food occur in caves where chemoau-
totrophic production by sulfur-oxidizing microbial
organisms {e.g., sulfur bacteria: Achromatium, Beggiatoa,
Thiothrix) occurs, resulting in sufficient energy to sup¬
port and sustain complex cave ecosystems. These types
of caves are dominated geochemically by reduced sul¬
fur compounds and are rare occurrences {e.g., Cesspool
Cave in Allegheny County, Virginia, and Lower Kane
Cave, Big Horn County, Wyoming).
Cave organisms can be separated on the basis of hab¬
itat and/or resource base: terrestrial riparian communities
found on stream banks with a resource base of allochtho¬
nous particulate organic matter deposited by stream
fluctuations; terrestrial transitory organic matter {dung)
communities usually living within a few hundred
meters of the surface with a resource base of organic
matter (often fecal material) that is derived from the
activities of animals {e.g., bats, crickets, raccoons) mov¬
ing in and out of caves; terrestrial epikarst communities
living primarily in the network of small, air-filled cavi¬
ties above the cave but below the surface; aquatic
stream communities living primarily in cave streams
ultimately dependent on dissolved and particulate
organic matter derived from the surface; aquatic phre¬
atic communities found in the permanent groundwater
at or below the cave itself, including the hyporheos;
and aquatic epikarst communities living primarily in the
network of small flooded or partially flooded cavities
above the cave but below the surface and most easily
sampled within the cave in drip pools.
DISTRIBUTION OF KARST, CAVES,
AND CAVERNICOLES
Within the continental United States, available and
appropriate habitat for subterranean fauna is not con¬
tinuous; thus cave-inhabiting organisms are found in
distinct areas (mostly karst), some being widely dis¬
tributed and isolated (Table 2, Figs. 1 and 2). Due to
these separations and a unique geology, history, and
climate, it should not be surprising that there are dis¬
tinct differences in regional fauna, variances
expressed in diversity, population densities, as well
as taxonomic groups. Isolation, due in part to folding
of strata, has resulted in high species richness (partic¬
ularly of troglobiotic beetles) in the Appalachians,
whereas in the Interior Lowlands this is not the case,
likely due to cave connectivity. Most cave-adapted
species are found south of the southern limits of the
ENCYCLOPEDIA OF CAVES
TABLE 2 Comparison of Species
DIVERSITY PATTERNS IN THE UNITED STATES
Richness and Characteristics of 10 U.S. Karst Regions 3
253
Major Karst Regions of the
United States
Number of
Troglobionts
Number of
Stygobionts
Total
Species
Number of
Caves
Area of Karst
(km 2 )
Appalachians
185 (2)
91 (1)
276 (2)
7,441 (2)
37,268 (5)
Black Hills
2(9)
0(10)
2(10)
160 (9)
7,272 (8)
Driftless Area
11 (8)
2(8)
13 (9)
615 (7)
25,222 (7)
Edwards / Balcones
105 (3)
58 (3)
163 (3)
2,011 (4)
65,586 (2)
Florida Lime Sinks
0 (10)
24(5)
24(6)
627 (6)
27,338 (6)
Guadalupes
13 (7)
1 (9)
14 (8)
1,379 (5)
43,522 (4)
Hawaii (lava)
35(4)
6(6)
41 (5)
?
?
Interior Lowland Plateau
256 (1)
63(2)
319 (1)
11,928 (1)
60,612 (3)
Mother Lode
20 (6)
3(7)
23 (7)
179 (8)
390 (9)
Ozarks
31 (5)
51 (4)
82 (4)
6,964 (3)
110,125 (1)
Total
658
299
957
31,304
377,335
a Ranks from highest to lowest are presented in parentheses (modified from Culver and Hobbs, 2002). An additional 180 species are from other cave and karst regions throughout
the contiguous United States and one described stygobiotic amphipod occurs in Alaska, bringing the total number of described obligate cavernicoles in the United States to 1138.
FIGURE 1 Map of major karst regions treated in this article (excluding the Hawaiian Islands, which are lava). The Interior Lowland
Plateau is stippled in order to differentiate it from the Appalachians
border) and the Ozarks. Modified from Culver et al., 2003.
Pleistocene glacial ice sheet, yet some were able to
survive subglacial conditions in the northern states (e.
g., New York, Wisconsin) as well as in Canada ( e.g
Alberta, British Columbia) and are represented
(extension into Canada truncated herein at Maryland—Pennsylvania
primarily by groundwater amphipods and isopods. In
the Driftless Area (parts of Illinois, Iowa, Minnesota,
and Wisconsin) and other smaller karst regions north
of the glacial boundary, low biodiversity in caves is
ENCYCLOPEDIA OF CAVES
254
attributed to the lack of time for colonization and iso¬
lation due to the effects of glaciers. Also, the
Guadalupe Mountains of New Mexico have a depau¬
perate cave fauna likely resulting from the aridity of
the region. Although high densities of lava tubes
occur in the western United States and limestone
caves have developed in numerous smaller karst
areas, this article focuses primarily on caves formed
in soluble rock, particularly limestone, and follows
the nine geologically defined cave regions discussed
by Culver and Hobbs (2002) (Fig. 1). These are the
Appalachians, Black Hills, Driftless Area, Edwards
Plateau and Balcones Escarpment, Florida Lime Sinks,
Guadalupe Mountains, Interior Lowland Plateaus,
Mother Lode, and the Ozarks. Also, a tenth, the Hawaiian
region, of volcanic origin (thus many islands are "young"
ecologically), is treated.
Caves of the coterminous United States are shown
in Figure 2 resulting from a plot of nearly 45,000 caves.
Clearly the distribution of cavernicoles is influenced
by the distribution and abundance of cavities available
for habitation. Indeed, although it is useful to examine
the number of cavernicoles in states (Fig. 3) or within
various karst areas (Fig. 4), Christman and Culver
(2001) and Culver et al. (2003) demonstrated that the
availability of habitat expressed by the number of
caves in a region (Table 2) is the best predictor of the
biodiversity of cave-dwelling organisms. Also, the dis¬
tribution of cavernicoles within caves is patchy and
nonrandom and is usually associated with concentra¬
tions of food sources. Whereas troglobiont diversity
within caves is likely affected not only by resource
availability but also by variety (organic plant debris,
bat or cricket guano, mammalian scat), stygobionts
tend to be less diverse in part due to the preponder¬
ance of feeding generalists and the lack of specializa¬
tion on food type.
TREATMENT OF THE FAUNA:
INVERTEBRATES
Class Turbellaria
This diverse group of stygobiotic flatworms (28
species) is assigned to four families and six genera.
The largest genus, Sphalloplana, is represented by 16
species found in subterranean settings primarily in
the Appalachians and Interior Lowlands major karst
regions.
Class Oligochaeta
This group of aquatic worms is represented by 13
species belonging to six genera and four families
ENCYCLOPEDIA OF CAVES
DIVERSITY PATTERNS IN THE UNITED STATES
255
State
« 70
•8 60
a. 50
2 40
2 30
■O
a
STYGOBITES
B
I I I
< u c
Z 1
< a o
~ o o
► < a
z S S
>
z
i i
> 5
State
State
FIGURE 3 Frequency histograms by state of number of described obligate subterranean species. (A) Troglobiotic biodiversity by state, (B)
stygobiont biodiversity by state, and (C) combined cavernicolous biodiversity by state.
mostly from the Appalachians and Interior Lowlands.
They are poorly known from cave soils and streams
as well as from interstitial waters, the latter habitat
only rarely sampled. Of interest, seven species of the
genus Cambarnicola (Branchiobdellidae) are restricted
as ectosymbionts to cambarid crayfishes occupying
subterranean waters. Tubificid worms are common to
subterranean streams containing much organic mate¬
rial (or that are receiving sewage effluent!), but no
stygobiotic forms are known.
Class Mollusca
Both aquatic and terrestrial snails inhabit caves, the
diversity of stygobionts being nearly six times greater
than that of troglobionts. Twenty-four stygobionts are
placed in 12 genera in two families and three families
contain four genera and five species of generally small,
translucent troglobiotic gastropods. Numerous species
of terrestrial epigean snails are deposited in caves on
debris transported in by floods but apparently are not
able to maintain populations for many generations.
Yet a few species are troglobiotic (mostly from
the Appalachians and Interior Lowlands karst regions,
e.g., Helicodiscus barri Hubricht, Carychium stygium
Call), feeding on guano and decaying allochthonous
material (mainly plant) and associated microbes.
Hydrobiid snails have been particularly successful in
subterranean waters with 23 stygobionts described
primarily from the Appalachians, Edwards Plateau
and Balcones Escarpment, Interior Lowlands (e.g.,
Antroselates spiralis Hubricht), and Ozarks. Some of
these snails are site endemics (e.g., Antrobia culveri
Hubricht, known from a stream in a single cave in
Taney County, Missouri) or are restricted to a single
drainage (e.g., Fontigens turritella Hubricht, from two
caves in the Greenbrier River drainage in Greenbrier
County, West Virginia). Physella spelunca Turner and
Clench is unique in that it is a site-endemic species
restricted to a thermal cave stream (Lower Kane Cave;
see above) draining into Big Horn River, Wyoming.
ENCYCLOPEDIA OF CAVES
256
o
u.
400
300
200
100
0
161
82
41
24
23
14
ca
N
HH
CD
hJ
<
w
O
Ph
S
u
o
SC
ca
Major Karst Region
FIGURE 4 Frequency histograms by major karst region of
described obligate cavernicoles. (A) Troglobionts, (B) stygobionts,
and (C) total. AP, Appalachians; BH, Black Hills; DL, Driftless Area;
EB, Edwards Plateau and Balcones Escarpment; FLS, Florida Lime
Sink; GUA, Guadalupes; HI, Hawaii; ILP, Interior Lowland Plateau;
ML, Mother Lode; OZ, Ozarks.
Class Arachnida
The cave-adapted arachnids are represented by a
very diverse group of mostly terrestrial arthropods that
are widely distributed throughout karst terrains of the
United States. Mites (Order Acari) are small arachnids
that are dominated by terrestrial species (11 genera and
19 species) with only three stygobionts placed in two
genera. Mites are undoubtedly underrepresented in
these totals due primarily to inadequate sampling.
Seven families are represented, with the rhagidiids
being the most diverse. Although most of the known
mites are predaceous and commonly observed mov¬
ing rapidly among organic debris in moist areas of
caves, some species are parasites on bats and har¬
vestmen. Stygobionts are known only from the
Interior Lowlands (Indiana), and the greatest diver¬
sity of troglobionts resides in the Appalachians and
Interior Lowlands; at least one species is endemic to
lava tubes on Hawaii. No troglobiotic ticks are
known but in some localities ticks are common, par¬
ticularly in bat caves.
Spiders (Araneae) have been quite successful in the
terrestrial cave environment with 98 species in 22 gen¬
era and 8 families. These troglobionts are known from
all the major karst areas except the Black Hills and the
Florida Lime Sinks. The genus Cicurina is particularly
diverse in Texas (49 species) and Phanetta subterranea
(Emerton) and Porhomma cavernicola (Keyserling) are
widely dispersed throughout the Appalachians and
Interior Lowlands, and less widely within the Ozarks.
Five species assigned to five genera and three families
are known from lava tubes in Hawaii. In addition to
the troglobionts, a significant number of facultative spe¬
cies {e.g., Gaucelmus augnstinus Keyserling, troglophile;
Meta ovalis (Latreille), trogloxene; Nesticus carteri Emerton,
troglophile) are common inhabitants of caves in the
major karst regions.
The Order Opiliones (harvestmen or "daddy long-
legs") are widely distributed throughout the karst
regions of the United States but are not overly diverse
except in the Mother Lode (13 species) and Edwards
Plateau and Balcones Escarpment (13 species). A total
of 44 species has been assigned to 13 genera and 6
families that are scattered over the Appalachians,
Interior Lowlands, and other lesser karst areas, as well
as those mentioned above. The genus Banksula is
diverse in the Mother Lode and environs (9 species)
and Texella is represented by 11 species in the
Edwards Plateau and Balcones Escarpment region as
well as an additional three species in California and
the Guadalupes. Large populations of trogloxenic har¬
vestmen are often encountered in ceiling pockets or
beneath ledges, out of the desiccating effects of air
movements where they form dense, undulating mats
{e.g., Leiobunum townsendii Weeds in caves in the
Edwards Plateau, Texas, and L. bicolor Wood in southern
Ohio caves). These trogloxenic species are typically
scavengers often associated with organic matter washed
into caves, while the troglobiotic forms are more com¬
monly predators feeding on microarthropods, such as
collembola.
Pseudoscorpions (Order Pseudoscorpiones) are mi¬
nute predators that have pincer-like pedipalps (as do
scorpions), but lack the tail and sting characteristic of
scorpions. Some are trogloxenic/troglophilic and are
often associated with bat guano communities, whereas
others are highly specialized troglobionts. This highly
diverse group of small arachnids (152 species within
29 genera and nine families) is particularly abundant
in Alabama caves (41 species) and also demonstrates
high diversity in Texas (25 species), and California and
Virginia (13 species each). They are found in all major
ENCYCLOPEDIA OF CAVES
DIVERSITY PATTERNS IN THE UNITED STATES
257
karst regions except the Florida Lime Sinks, the Black
Hills, and the Driftless Area. Two families are particu¬
larly widespread, Chernetidae and Chthoniidae, with
the latter being incredibly diverse (95 species). The
chernetids are dominated by the extensive genus
Hesperochernes, which is found in the Appalachians,
the Edwards Plateau and Balcones Escarpment,
Interior Lowlands, Mother Lode, and the Ozarks. The
chthoniids are heavily influenced by the genera
Apochthonius (15 species), Kleptochthonius (31 species),
and Tyrannochthonius (37 species) and are found in all
the major karst regions except the Black Hills, Florida
Lime Sinks, and the Driftless Area. Two species of the
genus Tyrannochthonius are found in lava tubes on the
Hawaiian Islands (Hawaii and Oahu).
The Order Schizomida is represented by a single
terrestrial species, Hubbardia shoshonensis Briggs and
Horn, from a cave in Inyo County, California.
Although similar to true whip scorpions, they have a
much shorter telson as well as other anatomical fea¬
tures that separate the two orders. They reside in leaf
litter and under stones and are predators on small
invertebrates.
Scorpions (Order Scorpiones) are believed to be one
of the most ancient terrestrial arthropods and the most
primitive arachnid. All are terrestrial predators charac¬
terized by two large pedipaps terminating in chelae as
well as by the stinging apparatus derived from the tel¬
son and bearing a sharp barb called the aculeus.
Although numerous pigmented species are found in
U.S. caves, only one species is troglobiotic: Uroctonus
grahami Gertsch and Soleglad, known only from Shasta
County, California. At least 11 species are troglobiotic
to the south in Mexican caves.
Subphylum Crustacea
Class Maxillopoda
The Subclass Copepoda (see "Crustacea" article, this
volume) is not well sampled in the cave environment
in the United States and this is reflected in the low spe¬
cies richness reported. Currently two families contain¬
ing three genera and seven species and subspecies of
these crustaceans are known from the Edwards
Plateau and Balcones Escarpment in Texas; the
Interior Lowlands in Illinois, Indiana, Kentucky,
and Tennessee; and from Orange County, North
Carolina. One obligate parasitic copepod, Cauloxenus
stygius Cope, is associated with the Northern
Cavefish, Amblyopsis spelaea, in southern Indiana
caves. Recent work in Slovenian caves and a few
caves in West Virginia has shown that copepods are
important constituents in the epikarst community.
Like the copepods, the Class Ostracoda has not been
studied thoroughly in U.S. caves. Even so, 14 species
belonging to 5 genera and 2 families are stygobionts in
the Appalachians, Edwards Plateau and Balcones
Escarpment, Florida Lime Sinks, Interior Lowlands, and
the Ozarks. Twelve species of the family Entocytheridae
are commensal on cambarid crayfishes inhabiting caves
and two species of Cyprididae are free living in cave
streams and lakes.
Class Malacostraca
The Order Bathynellacea (see "Crustacea" article,
this volume) is characterized by lacking maxillipeds
and a carapace. These small crustaceans are not
well studied, particularly in the United States, where
only one species, Iberobathynella bowmani Delamare
Deboutteville, Coineau, and Serban, is known from
two localities in Dickens and San Saba counties, Texas.
These reclusive organisms are interstitial or reside in
groundwater, those habitats not having been ade¬
quately sampled.
The Order Thermosbaenacea (see "Crustacea" arti¬
cle, this volume) is represented by a single species,
Monodella texana Maguire, which is known from a cave
and wells in Bexar, Hays, and Uvalde counties, Texas.
This is the sole North American representative of the
Thermosbaenacea and is endemic to the Edwards
Aquifer.
The Order Amphipoda is quite diverse with 134
stygobiotic species and subspecies in 18 genera and 10
families. An additional species is a talitrid troglobiont,
Spelaeorchestia koloana Bousfield and Howarth, which is
a rare endemic that feeds on plant roots and organic
debris in a few caves on the island of Kauai, Hawaiian
Islands. Most amphipods are scavenger/predators and
are laterally compressed, giving them a shrimplike
appearance. Amphipods occur in all major karst
regions except the Black Hills, Driftless Area, and
Guadalupes. Crangonyctidae is the most diverse and
widely distributed family, with the genera Bactrurus
and Crangonyx housing 7 and 12 stygobionts, respec¬
tively, yet the genus Stygobromus (98 species) is
approximately five times as diverse as those two gen¬
era combined. In addition to most of these being found
in unglaciated regions, S. quatsinensis Holsinger and
Shaw is endemic to caves and springs north of the
southern limits of Pleistocene glaciation in British
Columbia, Canada, and southeastern Alaska. Also,
eyed and pigmented stygophilic amphipods com¬
monly inhabit cave and karst spring waters of the con¬
tiguous states and are major players in the trophic
dynamics of some systems (e.g., Gammarus spp. and
Hyalella azteca Saussure).
ENCYCLOPEDIA OF CAVES
258
D
The Order Isopoda (see "Crustacea" article, this
volume) is represented in caves by 78 stygobionts in
11 genera and 4 families. The family Asellidae is
immense with all but 8 species of the order assigned
to this family of dorsoventrally flattened crustaceans.
Geographically, these stygobionts are widespread,
absent only from the Black Hills, Driftless Area,
Guadalupes, and the Mother Lode major karst
regions. The genus Caecidotea is represented by at least
60 species and subspecies and widely dispersed over
these regions, whereas other species are much more
restricted ( e.g., Remasellns parvus (Steeves) in Florida
and Salmasellus howarthi Lewis in Washington). The
genus Lirceus (also an asellid) has two stygobiotic spe¬
cies from southwestern Virginia (L. culveri Estes and
Holsinger and L. usdagalun Holsinger and Bowman
from Scott and Lee counties, respectively) and numer¬
ous stygophilic species {e.g., L. fontinalis Rafinesque)
that occupy springs and caves. The predominantly
marine family Cirolanidae is represented in the United
States by three dome-shaped and elliptical stygobitic
species, Cirolanides texensis Benedict and Speocirolana
hardeni Bowman, both from Texas, and the widely dis¬
junct Antrolana lira Bowman from Virginia and West
Virginia (Appalachians).
The troglobiotic isopods are represented by 19 spe¬
cies and subspecies in four genera. The 15 species in
the contiguous United States belong to three genera
(Amergoniscus, Brackenridgia, Miktoniscus ) within the
Trichoniscidae and these "pillbugs" are found in the
Appalachians and Interior Lowlands as well as in
Oklahoma, Oregon, and Texas. Four species of the
genus Hawaiioscia are endemic to the Hawaiian Islands
of Kauai, Maui, Molokai, and Oahu.
North American stygobiotic members of the Order
Decapoda (see "Crustacea" article, this volume) are
limited to 41 crayfishes, 5 shrimps, and 1 crab.
Crayfishes in the United States belong to the family
Cambaridae and the following 35 are stygobiotic: 11
species of Cambarus in the Appalachians, Interior
Lowlands, and the Ozarks; 9 species and subspecies of
Orconectes in the Interior Lowlands and Ozarks; 14
species and subspecies of Procambarus in the Florida
Lime Sinks and Interior Lowlands; and a single spe¬
cies of Troglocambarus in the Florida Lime Sinks.
Shrimps belong to two families: Atyidae, 2 species of
Palaemonias in the Interior Lowlands, and Palaemoni-
dae, three species of Palaemonetes in the Florida Lime
Sinks and Edwards Plateau and Balcones Escarpment.
The single crab (Grapsidae; Hemigrapsus oregonensis
Dana, formerly Hemigrapsus estellinensis Creed) is
endemic, and probably extinct, to a deep artesian
hypersaline spring in Hall County, Texas. Approxi¬
mately 50 stygophilic/stygoxenic crayfishes utilize
caves and, of special note and concern, is the impact
of introduced epigean species that utilize cave streams
(e.g., Procambarus (Scapulicambarus ) clarkii (Girard) in a
cave in San Diego County, California, and likely
introduced into the area from the southern Gulf
states).
Class Chilopoda
Troglobiotic centipedes are rare in caves, fast moving,
usually small, slender, and voracious predators. They
are found in organic debris, on clay, and burrowing in
silt. Four genera and species from three families are
known from the Guadalupes and the Appalachians as
well as from north-central California and the Edwards
Plateau and Balcones Escarpment.
Troglobiotic millipedes. Class Diplopoda, are slow-
moving grazers, feeding on organic detritus and asso¬
ciated microbes. Typically found in food-poor caves,
some 64 species and subspecies in 20 genera are
known primarily from the Appalachians, Driftless
Area, Guadalupes, Interior Lowlands, and Edwards
Plateau and Balcones Escarpment. Tetracion, Scoterpes,
and Pseudotremia are the three most important genera
in North American caves and the largest genus,
Pseudotremia (Cleidogonidae), has at least 33 troglobio¬
tic species in Appalachian and Interior Lowland caves.
In the west, the genus Amplaria is represented by at
least six species in California, Idaho, Oregon, and
Washington. Numerous troglophiles are noted in all of
these karst regions as well.
Class Hexapoda
The Class Hexapoda is restricted to the primitive
orders of insects lacking wings (apterygote) and, in
caves, is represented by three orders: Thysanura,
Diplura, and Collembola.
Troglobiotic thysanurans (bristletails) are unusual in
caves. A single species, Texoreddeiiia texensis (Ulrich), a
member of the Nicoletiidae family, is known from caves
in 18 counties in Texas and Speieonycta ozarkensis
Espinasa, Furst, Allen, and Slay, also a nicoletiid, occurs
in several caves on the Ozark Plateau in Oklahoma and
Arkansas. Trogloxenic species (e.g., Pedetontus sp.) are
not uncommon in entrance areas in many parts of the
United States.
Troglobiotic diplurans are small and lack a median
caudal filament but possess two cerci ("tails") that are
long, segmented filaments in the Campodeidae and
short, thick forceps-like appendages in the Japygidae
(see Reddell 1983). Eight species are assigned to three
genera and two families, with the campodeid Litocampa
being the most diverse genus with six species. This
poorly studied group is known from the Appalachians,
Interior Lowlands, the Edwards Plateau and Balcones
ENCYCLOPEDIA OF CAVES
DIVERSITY PATTERNS IN THE UNITED STATES
259
Escarpment, and a single species from Lincoln County,
Nevada ( Condeicampa langei Ferguson). These are
observed on wet flowstone, cricket guano, and on silty
substrata. Additionally, at least seven species belonging
to the genus Haplocampa are known from lava tubes in
Northern California, Idaho, Oregon, and Washington.
Troglobiotic springtails (Order Collembola) are small,
jumping insects that can occur in very high densities,
particularly where organic matter has accumulated.
Often they are observed moving about on the surface of
small drip pools as well as on larger stream pools. Not
only do they provide food for many small predators
but they also graze on bacteria and fungi. They are a
diverse group with 75 troglobiotic species assigned to
12 genera and are known from all the major karst
regions, except the Florida Lime Sinks, as well as other
minor karst areas. Entomobryidae is the largest family,
with the greatest diversity in the genus Pseudosinella (25
species); Sinella has 13 species. Fifteen species make up
the widely distributed genus Arrhopalites (Sminthuridae).
Numerous troglophilic species (eg., Sinella cavernarum
(Packard)) contribute to the terrestrial communities of
many caves. Six troglobiotic species assigned to three
genera and two families are found in lava tubes on the
islands of Hawaii, Maui, and Oahu.
Class Insecta
The Class Insecta is characterized by those insects
possessing wings (pterygote) or secondarily with
reduced wings and is represented in caves by troglo¬
biotic/sty gobiotic members of five orders. Troglobiotic
members of the Order Orthoptera are known only
from Hawaiian lava tubes. Three species of flightless,
mute crickets are assigned to two genera and belong to
the family Gryllidae. Caconemobius varius Gurney and
Rentz is known from five caves on Hawaii Island and
displays small eyes, reduced body pigment, and a trans¬
lucent exoskeleton. It is omnivorous or perhaps a scav¬
enger and appears to be most abundant in low, wet
caves. C. howarthi Gurney and Rentz has been collected
from four caves on the island of Maui. It has been
observed only in the dark zone of caves and appears to
be more cave-adapted than C. varius. Thanmatogryllus
cavicola Gurney and Rentz is known from approxi¬
mately 10 small, generally shallow caves on the island
of Hawaii. In some of these lava tubes it has been
observed on dangling roots of grasses and trees that
have penetrated the ceilings in characteristically wet
areas. In the contiguous states, pigmented rhaphido-
phorid crickets often are very numerous in caves, partic¬
ularly within the first 100 to 200 m of the entrance, and
these trogloxenes are represented by species of the gen¬
era Ceuthophilus, Hadenoecus, and Euhadmoecus. These
crickets not only migrate to the surface to forage for
food during spring, summer, and fall (and thus provide
food in the form of guano to caves), but they also use
caves as refugia during cold periods. Some crickets (eg.,
Ceuthophilus silvestris Bruner in small caves in glaciated
Wyandot County, Ohio) interact with caves only season¬
ally, entering them solely during the winter months.
A single endemic, troglobiont ( Anisolabis hozvarthi
Brindle) from Hawaii is the only known member of
the Order Dermaptera (earwigs) to inhabit caves.
Characterized by the heavily sclerotized posterior for¬
ceps (cerci), they use these for predation and are con¬
sidered to be scavenging omnivores as well.
Only two troglobiotic "bugs" (Order Hemiptera—
Heteroptera) are known, both endemic to lava tubes on
the island of Hawaii. The thread-legged bug, Nesidio-
lestes ana Gagne and Howarth (Reduviidae), is a preda¬
tor of arthropods, and the lava tube water-treader,
Speleovelia aaa Gagne and Howarth, sucks on rotting
fluids of deceased arthropods.
The Order Homoptera is represented by seven
troglobitic planthoppers endemic to the islands of
Hawaii, Maui, and Molokai. All belong to the Cixiidae
family and are placed in the genus Oliarus. Often they
are observed on roots that have penetrated ceiling
cracks of lava tubes.
The Order Coleoptera has been incredibly successful
in caves with 261 species in 21 genera and five families.
Only three species are aquatic, two of which are dryo-
pids (Oregon and Texas) and one a dytiscidae from
Texas ( Haideoporus texanus Young and Longley, the only
stygobiotic predaceous diving beetle known in the
United States). Beetles are major components of cave
communities particularly in the Appalachians and
Interior Lowlands. Troglobionts are dominated by the
carabids, particularly the genus Pseudanophthalmus
which is inordinately abundant in Kentucky and Texas
caves, 62 and 67 species, respectively. Two carabids are
endemic to lava tubes on the Hawaiian island of Maui.
Troglobiotic ground beetles are represented by the
genus Rhadine (17 species), which is restricted to Texas,
particularly dense in the Edwards Plateau and Balcones
Escarpment karst region. In certain Texas caves, Rhadine
s. subterraneanea (Van Dyke) and R. noctivaga Barr feed
exclusively on cave cricket eggs (Ceuthophilus spp.) that
have been buried in fine-grained, calcareous deposits.
This feeding behavior has also been demonstrated in
Mammoth and other caves in the Pennyroyal Plateau of
western Kentucky (Interior Lowlands) for the carabid
Neaphaenops tellkampfii (Erichson) preying on eggs and
nymphs of the cricket Hadenoecus subterraneus, and on
the Cumberland Plateau of eastern Kentucky the cricket
is Hadenoecus cumberlandicus and its predator is Darling-
tonea kentuckensis Valentine. Leiodids (Ptomaphagous
spp.) and pselaphids (predominantly Batrisodes spp.)
make up the remaining troglobionts. Numerous
troglophiles (eg., Carabidae, Leiodidae, Leptodiridae,
ENCYCLOPEDIA OF CAVES
260
D
Ptinidae, Staphylinidae, Tenebrionidae) also are com¬
mon members of cave communities. In addition to pred¬
ators, the cavernicolous beetles are scavengers and
opportunistic saprophiles.
A single, widely distributed troglobiotic fly (Order
Diptera) is known from caves in the Appalachians, the
Interior Lowlands, and the Ozarks: Spelobia tenebrarum
(Aldrich), Sphaeroceridae. A variety of flies utilize
caves periodically and include various culicids {e.g.,
Culex pipiens Linnaeus), heleomyzids {e.g., Amoebaleria
defessa (Osten-Sacken)), mycetophilids {e.g., Macrocera
nobilis), phorids {e.g., Megaselia cavernicola (Brues)), and
sphaerocerids {e.g., Leptocem tenebrarum (Aldrich)).
Additional insects are observed in caves but are
either troglo/stygoxenes or troglo/stygophiles and
include lepidopterans {e.g., Scoliopteryx libatrix
(Linnaeus)), psocopterans {e.g., Psyllipsocus ramburi
Selys-Longchamps), and siphonapterans {e.g., Myo-
dopsylla insignis (Rothschild)). A particularly dis¬
turbing example of surface insects entering the
subterranean world is the invasion of central Texas
caves during the summer months by the trogloxenic,
exotic red fire ant, Solenopsis invicta (Hymenoptea).
TREATMENT OF THE FAUNA:
VERTEBRATES
A number of vertebrates spend varying times of
their life in subterranean environments but most are
not considered cavernicoles {e.g., some snakes and
mammals are adapted to a fossorial mode of life in the
soil). Other vertebrates visit caves for shelter during
times of unfavorable surface environmental conditions
{e.g., amphibians, snakes), to avoid predation {e.g.,
pack rats, birds), or for breeding purposes {e.g., birds,
bats). Generally these cave-dwellers are classified as
accidentals or as trogloxenes/stygoxenes; fishes and
salamanders are the only vertebrates in which troglo-
philic/stygophilic or troglobiotic/stygobiotic species
have evolved.
Class Osteichthyes
Approximately 85 stygobiotic fishes are known
worldwide, however only six cave-obligate species are
recognized from North American subterranean waters
and belong to one of two families, Amblyopsidae and
Ictaluridae. The amblyopsid cavefishes are stygobiotic
except for one stygophilic species, Forbesichthys agassizii
(Putnam), which inhabits caves and springs of Illinois,
Kentucky, and Tennessee yet migrates to the surface at
night to exploit food resources. Four additional species
are confined to subterranean waters: Amblyopsis rosae
(Eigenmann) from the Ozarks, Amblyopsis spelaea from
the Interior Lowlands, Speoplatyrhinus poulsoni Cooper
and Kuehne from a single cave in the Interior
Lowlands, and Typhlichthys subterraneus Girard from
the Interior Lowlands and the Ozarks. The ictalurid
subterranean representatives, Satan eurystomus Hubbs
and Bailey and Trogloglanis pattersoni Eigenmann, are
known only from deep phreatic waters of the Edwards
Aquifer in Bexar County, southwest Texas, and have
never been observed in accessible caves. Members of
both of these families are predatory feeders primarily
of crustaceans except the toothless T. pattersoni, which
is likely a grazer taking up detritus.
Other fishes occupy subterranean waters (some reg¬
ularly demonstrate migratory patterns) and some of
the families represented are the following: Anguillidae
{e.g., Anguilla rostrata (LeSueur)), Centrarchidae
{Lepomis spp.), Cottidae {Cottus spp.), Cyprinidae
{Notropis spp.), Ictaluridae {Ictalurus catus (Linnaeus)),
Mugilidae {Mugil cephalus Linnaeus), and except for
Cottus sp., show no troglomorphic adaptations for liv¬
ing in this environment.
Class Amphibia
Oddly, no stygophilic or stygobiotic frogs are known
from caves (possibly because they cannot complete
their life history in the subterranean environment),
but they certainly utilize caves {e.g., Ram palustris
LeConte), primarily to retreat from surface climatic
extremes. Globally, only 13 stygobiotic salamanders are
known from hypogean waters and are assigned to two
families, Proteidae and Plethodontidae, the former
restricted to limestone caves and karst springs in Bosnia
and Herzegovina, Croatia, Italy, and Slovenia. The
plethodontids are represented by 11 species and sub¬
species in North America and are known from the
Appalachians, Edwards Plateau and Balcones
Escarpment, Interior Lowlands Plateau, and the Ozarks
as well as from caves in the Dougherty Plain of south¬
western Georgia and adjacent panhandle of Horida
(Tables 2 and 3). Numerous additional species of pletho-
dontid salamanders are stygophilic {e.g., Pseudotriton
ruber (Latreille)) and troglophilic {e.g., Eurycea lucifuga
(Rafenesque)).
Class Reptilia
Although there are no known obligate reptiles
from caves in the United States, these habitats are
used often by snakes for short durations or even sig¬
nificant amounts of time. Most tend to remain near
entrances and use the cave as a refuge from extreme
surface conditions. Some snakes prey on bats during
their flights out of the cave {e.g., the colubrid Elaphe
ENCYCLOPEDIA OF CAVES
DIVERSITY PATTERNS IN THE UNITED STATES 261
TABLE 3
Obligate Subterranean Salamanders (Plethodontidae) of the United States
Genus
Species/Subspecies
State
Global Conservation
Status (Nature Serve)
Eurycea
latitans
Texas
G3
rathbuni
Texas
G1
robusta
Texas
G1
spelaea
Arkansas, Kansas, Missouri, Oklahoma
G4
tridentifera
Texas
G3
troglodytes
Texas
G1
Gyrinophilus
gulolineatus
Tennessee
G1
palleucus palleucus
Alabama, Georgia, Tennessee
G2
palleucus necturoides
Tennessee
G2
subtermneus
West Virginia
G1
Haideotriton
wallacei
Florida, Georgia
G2
obsoleata (Say)) and certain lizards are observed in the
entrance areas of caves as well ( e.g ., the geckonid
Coleonyx variegatus brevis (Steineger) in Texas). In
Florida the American alligator ( Alligator mississipiensis
(Daudin)) and the cooter turtle ( Pseudemys floridana
(LeConte)) are seen in the entrances to many sub¬
merged caves and will venture short distances into
these cavities that serve often as the source of large
karst spring runs.
Class Aves
The most noted of all birds that utilize caves is
probably the troglophilic oilbird of Cueva del Guacharo
in Venezuela, South America ( Steatornis caripensis
Humboldt), for centuries prized by locals as a source of
cooking and lighting oil. Although the eyes of this bird
are large and well developed for use in dim light, they
utilize echolocation (as do bats) when in the complete
darkness of caves. In the United States, no such highly
specialized species occur in caves yet swallows (e.g.,
Hirundo spp.), the canyon wren ( Catherpes maxicanus
(Swanson), and the eastern phoebe ( Sayornis phoebe
(Latham)) are common inhabitants of cave entrances.
Various species inhabit the twilight zone of caves,
including the eastern (common) screech-owl (Otus asio
(Linnaeus)) and the turkey vulture ( Cathartes aura
(Linnaeus)) that nest in this semidark zone of caves.
Class Mammalia
The occurrence of mammals in caves is quite
common although none is obligate and most are
temporary residents/visitors, including numerous
rodents such as pack rats ( Neotoma spp.), mice
(Peromyscus spp.), beaver ( Castor canadensis Kuhl), the
wood-chuck ( Marmota monax (Linnaeus)), as well as
carnivores such as bears (Ursus spp.) and raccoons
(Procyon lotor Linnaeus). Clearly no bats are obligate
cave-dwellers yet 37 of the 50 + species and subspe¬
cies found in the United States occupy caves at least
occasionally. They are assigned to four families
(Table 4) and most are insectivorous (e.g., the endan¬
gered Indiana Bat, Myotis sodalis Miller and Allen), but
some feed on fruits, pollen, and nectar (e.g., Mexican
long-tongued bat, Choeronycteris mexicana Tschudi).
During the winter most hibernate in caves (or mines)
often in large, dense clusters of up to several thousand
individuals and the summers are spent in trees or build¬
ings. A few species (e.g., gray bat, Myotis grisescens
Howell) may live in caves throughout the year, although
different ones are utilized in winter and summer. Some
species, like the Brazilian free-tailed bat (Tadarida brasilien-
sis (Saussure)), occupy caves in very large numbers and
contribute an immense amount of organic material
(guano, i.e., bat droppings) to the cave ecosystem.
During the summer approximately 20,000,000 indivi¬
duals of this particular species occupy a single cave
near San Antonio, Texas (Harvey et al. 1999), and, at
that time, the population represents the largest concen¬
tration of mammals in the world. Guano produced by
bats living in caves not only supports very large com¬
munities of guanophiles (e.g., bat fleas, dermestid beetle
larvae, gnats, and pseudoscorpions), but also this nitro¬
gen-rich material has been mined for use as fertilizer as
well as for producing saltpeter (potassium nitrate),
which is an ingredient of gunpowder.
ENCYCLOPEDIA OF CAVES
262 d
TABLE 4 Cave-Dwelling Bats of the United States
Family
Genus
Species/Subspecies
Common Name
Global Status
Molossidae
TaAarida
brasiliensis
Brazil free-tailed bat
G5
Mormoopidae
Mormoops
megalophylla
Ghost-faced bat
G4
Phyllostomidae
Choeronycteris
mexicana
Mexican long-tongued bat
G4
Leptonycteris
nivalis
Mexican long-nosed bat
G3
Leptonycteris
yerbabunae
Lesser long-nosed bat
G4
Vespertilionidae
Antrozous
pallidus
Pallid bat
G5
Corynorhinus
rafinesquii rafinesquii
Rafinesque's big-eared bat
G3, G4
Corynorhinus
r. macrotis
Big-eared bat
G3, G4
Corynorhinus
townsentii ingens
Ozark big-eared bat
G4
Corynorhinus
townsendii townsendii
Townsend's Western big eared bat
G4
Corynorhinus
t. pallescens
Pale lump-nosed bat
G4
Corynorhinus
t. virginianus
Virginia big-eared bat
G4
Eptesicus
fuscus
Big brown bat
G5
Idionycteris
phyllotis
Allen's big-eared bat
G3, G4
Lasiurus
borealis
Eastern red bat
G5
Lasiurus
cinereus
Hoary bat
G5
Lasiurus
c. semotus
Hawaiian hoary bat
G5
Myotis
auriculus
Southwestern Myotis
G5
Myotis
austroriparius
Southeastern Myotis
G3, G4
Myotis
californicus
Californian Myotis
G5
Myotis
ciliolabrum
Western small-footed Myotis
G5
Myotis
evotis
Long-eared Myotis
G5
Myotis
grisescens
Gray Myotis
G3
Myotis
keenii
Keen's bat
G2
Myotis
leigii
Eastern small-footed Myotis
G3
Myotis
lucifugus
Little brown Myotis
G5
Myotis
septentrionalis
Northern Myotis
G4
Myotis
sodalis
Indiana Myotis
G2
Myotis
thysanodes
Fringed Myotis
G4, G5
Myotis
thysanodes vespertinus
Pacific fringe-tailed bat
G4, G5
Myotis
t. palmsapensis
Fringe-tailed Myotis
G4, G5
Myotis
velifer
Cave Myotis
G5
Myotis
v. brevis
Southwestern cave Myotis
G5
Myotis
volans
Long-legged Myotis
G5
Myotis
yamanensis
Yuma Myotis
G5
Parastrellus
hesperus
Western pipistrelle bat
G5
Perimyotis
subflavus
Eastern pipistrelle bat
G5
ENCYCLOPEDIA OF CAVES
DIVERSITY PATTERNS IN THE UNITED STATES
263
During the winter of 2006—07 disturbing observa¬
tions were made of hibernating bat species in eastern
New York. Affected bats demonstrated unusual behav¬
ior and displayed a white fungus on their muzzles and
wings and were subsequently diagnosed with "White
Nose Syndrome" (WNS) associated with an infection
of the fungus Geomyces destructans. The number of
mines and caves in which bats infected with the fun¬
gus has risen sharply and the estimates of losses
exceed one million individuals. Infected populations
with individual losses up to more than 90% have been
confirmed in Connecticut, Delaware, Massachusetts,
New Hampshire, New Jersey, Indiana, New York,
Ohio, Pennsylvania, Tennessee, Vermont, Virginia, and
West Virginia and is likely in Missouri and Oklahoma.
Currently the nature of the relationship of the fungus
to this widespread mortality is unclear.
DISCUSSION
The biodiversity of cavernicoles in the United States
is summarized in Table 1 and their distribution pat¬
terns are demonstrated in Tables 2 and 3 and Figs. 3
and 4. Clearly the more mobile troglobionts have been
very successful, with more than twice their numbers
invading subterranean habitats when compared to
stygobionts. Also, insects (361 species and subspecies
including apterygous and pterygous forms; 32% of all
obligate species), arachnids (319; 28%), and crustaceans
(302; 26%) dominate the species richness summaries,
with terrestrial arachnids making up nearly one-third
of all subterranean species. Much variation is shown
among major karst regions, with those to the north
(Black Hills and Driftless Area) having few obligate
species. The number of stygobionts ranges from zero
(Black Hills) to 91 (Appalachians), and the number of
troglobionts varies from zero (Horida Lime Sinks) to
256 (Interior Lowlands), and even at the generic level,
overlap among regions is low. Karst regions with the
greatest total biodiversity are the Interior Lowlands,
Appalachians, and the Edwards Plateau and Balcones
Escarpment (Fig. 4C), although troglobionts show some¬
what different patterns than do stygobionts (Figs. 4A,B).
States with the highest total species richness are Texas,
Tennessee, Alabama, Kentucky, and Virginia.
Stygobionts show a different pattern: Texas, Virginia,
Tennessee, Missouri, and West Virginia (Fig. 3). Most
assuredly, "hotspots" do exist where concentrations of
fauna occur and northeastern Alabama is one of these
centers of biodiversity for troglobionts and stygobionts
Culver et al. 1999b. Of particular note, Jackson County
has 1526 caves and 66 obligate cavernicoles, which
translates into more than three times the number of
caves and nearly twice as many species as any other
Alabama county. This shows the strong relationship
between the number of caves observed and the number
of species reported (r 2 = 0.81).
Troglobionts and stygobionts make up slightly more
than 50% of the imperiled U.S. fauna that is tracked in
the central databases of the Natural Heritage Program.
There are far too many potential (and realized) distur¬
bances to discuss herein that threaten these out-of¬
sight organisms and the reader is referred to Elliott
(2000) for an excellent review of them. Suffice it to say
that the ultimate long-term survival of subterranean
karst communities depends on appropriate manage¬
ment and protection of the cave, the groundwater, and
the entire catchment area.
It becomes clear that these subterranean species are
geographically concentrated in a small percentage of
the landscape, with more than 50% of cave-inhabiting
species occurring in less than 1% of the land. Hence, it
is much easier to preserve a large percentage of at-risk
species by focusing habitat conservation efforts in
those areas of high concentrations of obligate cave
fauna, or hotspots. Protecting and conserving karst
habitats and their biodiversity is a challenging but
most important task for modern and future speleolo¬
gists. The subterranean biodiversity of the United
States is globally significant but extremely vulnerable
and little is known of the long-term impacts of WNS
on bats or on entire cave ecosystems.
See Also the Following Articles
Diversity Patterns in Australia
Diversity Patterns in the Dinaric Karst
Diversity Patterns in Europe
Diversity Patterns in the Tropics
Bibliography
Christman, M. C., & Culver, D. C. (2001). The relationship between
cave biodiversity and available habitat. Journal of Biogeography, 28
(3), 367-380.
Culver, D. C., Christman, M. C., Elliott, W. R., Hobbs, H. H., Ill, &
Reddell, J. R. (2003). The North American obligate cave fauna:
Regional patterns. Biodiversity and Conservation, 12(3), 441—448.
Culver, D. C., & Hobbs, H. H., Ill (2002). Patterns of species richness
in the Florida stygobitic fauna. In J. B. Martin, C. M. Wicks &
I. D. Sasowsky (Eds.), Hydrogeology and biology of post-Paleozoic
carbonate aquifers (pp. 60—63). Charles Town, WV: Karst Waters
Institute (Special Publication 7).
Culver, D. C., Hobbs, H. H., Ill, Christman, M. C., & Master, L. L.
(1999a). Distribution map of caves and cave animals in the
United States. Journal of Cave and Karst Studies, 61, 139—140.
Culver, D. C., Hobbs, H. H., Ill, & Mylroie, J. E. (1999b). Alabama: A
subterranean biodiversity hotspot. Journal of the Alabama Academy
of Sciences, 70, 97—104.
ENCYCLOPEDIA OF CAVES
264
D
Culver, D. C., Master, L. L., Christman, M. C., & Hobbs, H. H., Ill
(2000). Obligate cave fauna of the 48 contiguous United States.
Conservation Biology, 24(2), 386—401.
Elliott, W. R. (2000). Conservation of the North American cave and
karst biota. In H. Wilkins, D. C. Culver & W. F. Humphreys
(Eds.), Ecosystems of the world 30: Subterranean ecosystems
(pp. 665—689). Amsterdam: Elsevier Press.
Harvey, M. ]., Altenbach, J. S., & Best, T. L. (1999). Bats of the United
States (Arkansas Game & Fish Commission). Asheville, NC: U.S. Fish
and Wildlife Service.
Hobbs, H. H., Ill (1992). Caves and springs. In C. T. Hackney, S. M.
Adams & W. H. Martin (Eds.), Biodiversity of the southeastern
United States: Aquatic communities (pp. 59—131). New York: John
Wiley & Sons.
Nicholas, B. G. (1960). Checklist of macroscopic troglobitic organisms
of the United States. American Midland Naturalist, 64(1), 123—160.
Packard, A. S. (1888). The cave fauna of North America, with
remarks on the anatomy of the brain and the origin of the blind
species. Memoirs of the National Academy of Science, 4, 1 — 156.
Peck, S. B. (1998). A summary of diversity and distribution of the
obligate cave-inhabiting faunas of the United States and Canada.
Journal of Cave and Karst Studies, 60(1), 18—26.
Reddell, J. R. (1983). A checklist and bibliography of the Japygoidea
(Insecta: Diplura) of North America, Central America, and the
West Indies. The Pearce-Sellards Series (Texas Memorial Museum),
37, 1-41.
DOCUMENTATION AND DATABASES
Peter Matthews
Informatics Commission, International Union of Speleology
INTRODUCTION
Where would we be without cave and karst documen¬
tation? Perhaps a cliche but it ranges from describing
the contents of one cave to plotting the karst areas of
Europe against existing protective reserves, and is the
foundation upon which much of cave exploration,
research, protection, and management is based.
This article offers a brief outline of what is involved
in cave documentation, its issues and techniques. The
reader is directed to the bibliography and to the
Informatics Commission [22] of the International
Union of Speleology's (UIS) website for further detail
and examples via the bracketed numbers within this
article. The plan for the UIS website is to continually
add more and updated examples and practical help
with cave and karst documentation. Readers who
know of further examples that should be on the web¬
site, please contact the author. Any reference in this
article to companies or products does not constitute
endorsement, merely examples.
This article first considers cave documentation in
general, and then expands on the database aspects. By
documentation, we mean the recording of descriptive
information in the broadest sense about specific caves
and karst areas. As scientific documentation can be
quite project-specific, we focus mainly on the more
general type generated by cavers and speleologists.
This is usually done on a volunteer basis and with a
very low budget. Although many cavers see their
activity primarily as a sport, they realize it is also use¬
ful to record what they find, especially as the informa¬
tion is otherwise not readily obtainable.
There are two broad categories of people passionate
about caves and karst: amateur cavers and professional
scientists. These quite different groups can be of signifi¬
cant benefit to each other. Caving groups over time
gather a large amount of information about a large num¬
ber of caves. While often not scientific experts, cavers
are usually able to mine their data to flag up caves that
may be of interest to scientists for deeper study but
which would otherwise remain unknown. Scientists, of
course, can help cavers understand the caves better and
perhaps even guide further exploration.
Three relatively recent technical innovations have
made a significant difference to the effectiveness of
cave documentation: computers, the Internet, and
global positioning systems (GPS). Computers and
databases have revolutionized data storage space,
retrieval, and presentation; the Internet and the web
have revolutionized communication and helped reveal
the existence of cave information from anywhere in
the world, and permitted it to be remotely accessed
and updated; and the GPS has simplified the position¬
ing of caves, even where there are no landmarks at all,
such as on the 1100-km Nullarbor Plain in Australia.
DOCUMENTATION
History
In the early days datasets traditionally consisted of
an index card box for summaries and cartons or multi¬
ple filing cabinets for the detailed reports. Impressive
datasets were built up by caving groups in many
countries. Based on these, printed cave lists started to
be published, such as Pennsylvania Caves in 1930 [1]
and the Australian Speleological Federation's (ASF)
national listing in Speleo Handbook in 1968 [2].
By the late 1960s, computerized cave databases had
started to appear and then were helped along by the
appearance of Digital Equipment's minicomputer and
the desktop personal computer, both of which brought
free computing within the reach of more cavers. Cave
surveying and mapping programs started to appear,
helped also by Hewlett Packard's small pen plotters.
In the 1980s came geographic information systems
(GIS), and in the 1990s the web started to take off.
Edited by Foxit Reader.
Copyright © by Foxit Corporation, 2005—2009. For evaluation only.
DOCUMENTATION AND DATABASES
265
enabling easy access to centralized databases. Today
we are working with GPS, laser cave profile mea¬
surers, hand held in-cave map plotters, and even
walk-through cave mapping using GPS principles with
underground radio [10]. And since 2007 the Karst
Information Portal (KIP) initiative [3] has provided a
major advance in web access to cave information from
around the world.
In 1986 UIS instituted its Informatics Commission
(UISIC), whose role was to "encourage and facilitate
the systematic collection and responsible use of cave,
karst and related data on an international basis." To
this end the Commission has operated a website since
1994 containing various aids to cave documentation,
and has various projects in the pipeline. The
Commission's Cave Survey Working Group, ably led by
Philipp Hauselmann, has established standards for
survey and mapping.
An example of a development from paper to com¬
puter database to publishing is ASF's Australian
Karst Index, a project by this article's author. In 1972
I was trying to work out for ASF how we could auto¬
mate storing, retrieving, and publishing cave infor¬
mation when I came across, in the United States, the
National Speleological Society's (NSS) then defunct 80-
column punch-card system, which had coded values
for the field associated with each card column.
Eureka! It sounds obvious now, but coded values
and a fixed vocabulary for each field enabled me to
design a tick-the-box-style 3-page form and a data¬
base with numerically coded values for most fields.
With the help of free computing from my employer
and a lot of hard work by caving clubs all around
Australia we created a comprehensive national cave
database which was then used to generate the
national cave list by translating the codes into
English and generating the masters for publication of
the Australian Karst Index 1985 [4]. In 1999 I con¬
verted this database to a fully relational one which
ran on PCs, and since then it has been further devel¬
oped by ASF for the web where it has been accessible
online since 2001 and updateable since 2005, using
ASF's open-source software [5].
Cave Documentation Issues
In managing cave documentation, the following
issues commonly arise:
• The difficulty of getting information at all. To overcome
human nature's inertia it is important to minimize
impediments to reporting. For example, the Karst
Institute in Slovenia was successful in receiving an
impressive amount of information over the years by
providing a post-paid, postcard-sized reporting
form which cavers could simply fill in and drop in a
postbox on their way home from a trip.
• Cave identification. To assign information to the
correct cave on an ongoing basis there needs to be a
method of identifying each cave and cave entrance.
To avoid having to devise a unique name for every
little cave, a cave numbering scheme is usually used
and is usually hierarchical by state, area, and serial
number, for example, 3B-5. This number is often
marked at the cave entrance to positively identify
the cave, and may also be the starting point for the
cave survey.
• Report types, filing, and retrieval. In terms of filing
and indexing paper documents there are basically
three types: cave reports, trip reports, and the rest
(books, scientific papers, newsletter and newspaper
articles, theses, websites, etc.). The question is how
to know that the information is there and how to
access it in these documents.
• Secrecy, trust, and publication. Cavers generally don't
like cave information to become available to the
public. Do they trust the custodian? What is safe to
publish? Get it wrong and few people will contribute
data.
• Safety of records. Are they safe against loss by
destruction and unauthorized dispersion?
• Access policy. To help ensure safety, some kind of
published Records Access Policy is needed, and
perhaps a Data Use Agreement for digital data.
• Credibility. If wrong or unreliable information gets
into a database its credibility suffers and it therefore
becomes much less useful. Data verification is
important.
• Standards. What standards will be used for recording,
indexing, storing?
• Data transfer. Invariably there comes a time when
digital data needs to be transferred into or out of a
system. What format will be used? Flow difficult
will it be for the recipient to understand what they
are receiving?
Recording
There are basically two scenarios: field recording of
raw data from one cave or one trip, and then the back-
office work consolidating, summarizing, and indexing
accumulated data.
Recording can be helped by the availability of
forms. For example, ASF has six forms which helped
with its KID project: three field forms (cave report, trip
report, sketch sheet) and three back-office summary
forms (cave summary, area summary, and map sum¬
mary); and the Victorian Speleological Association (VSA)
has developed a form for comprehensively indexing
ENCYCLOPEDIA OF CAVES
266
D
any cave-related document. UISIC has a project to make
forms available, including a cave summary form with
customizable fields; these will eventually appear on the
website. Recording directly into a database may be
possible in some situations, such as into a hand held
computer in the field for later uploading into a database,
but paper forms are usually still convenient for initial
recording and for consolidating information from vari¬
ous sources prior to data entry. It also helps others verify
the data after data entry.
It is best to encourage recording at least the cave
data while still on the trip. The longer the delay after
arriving home, the less likely the data will be supplied.
Where data are not going into a database, one useful
further recording step is at least the indexing of the
documents, so that you know which documents to
look in for details of a particular cave or speleological
aspect. The simplest index would be a list of the trip
reports or other documents with their date, title, and
which areas and caves were mentioned.
Expedition recording is a whole subject in itself.
Some of the issues are:
• Coordination, such as GPS data, camera
timestamps, recording methods, and interfacing to
members' instruments such as various GPS models.
It is recommended that before an expedition a GPS
datum be agreed on, all locations be recorded
against that datum, all cameras be synchronized
against local GPS time, and then the associated GPS
date/time screen be photographed to allow accurate
later auto-synchronizing of photos against GPS track
points
• Consolidating data from the various expedition
members' fieldbooks and instruments
• Issuing updated and consolidated data back to
members to aid the next stage of exploration and
mapping
• Publication of the expedition summary results
• The rational consolidation, storage, and distribution
of the mass of detailed expedition results, such
as cave descriptions, cave numbers, locations,
photographs, specimens, participants, trip journal,
etc.
• There are external issues as well, such as what
expeditions by others are planned or have already
been to the area, and the ethics to be observed when
exploring in a foreign country
• Things that may help [6]:
Cambridge University Caving Club's (CUCC)
Troggle expedition web database
The planned International Caving Expeditions (ICE)
database, a combined European Speleological
Federation (FSE) and UIS project
VSA is developing a method to make all data
from an expedition linked and accessible using
an offline web browser, and with "KML" files for
use with Google Earth
UIS Code of Ethics for Cave Exploration and
Science in Foreign Countries (available in seven
languages)
Information Storage
Datasets can take many physical forms, ranging
from paper files to spreadsheets to databases to GIS
systems.
The traditional method of paper files will be with us
for a while yet as that is how most of the data will be
provided. But they do present a storage and retrieval
problem. A lot of the information requires extraction
or indexing before filing, so this can easily get behind
in a volunteer environment. (By extraction is meant
extracting the report's actual data into a database, and
by indexing is meant cataloguing the report as to what
type it is, what caves and/or areas are mentioned, and
what kind and level of data are included.) The use of a
separate sheet per cave after a trip (cave reports) can
help, because then that sheet can simply be dropped
into that cave's folder; thus all the data for a particular
cave can be more readily located without any extra
work. Obviously also extracting that data would pro¬
vide even better access. Trip reports, on the other
hand, often contain data on several caves and therefore
some level of indexing is mandatory if the reference to
each cave is not to be buried.
Any form of data extraction is subject to human
error, so it can be an advantage to be able to easily
refer to the original document if necessary. Scanning
of the original documents is therefore a useful step,
as well as it creating a backup copy. If the scans are
checked carefully afterwards, the paper copy can
then be discarded, so saving considerable storage
space.
Today data extraction or indexing of original paper
documents is likely to be in digital form, such as listed
below. Because of the increasing level of skill and time
involved, not all data may get to the database or GIS
stage:
• Scans of original documents which are identified
only by their filename
• Indexes of original documents listing their existence
and preferably also their types of content
• Spreadsheets for simple data tables
• Databases allowing complex searching and
reporting
ENCYCLOPEDIA OF CAVES
DOCUMENTATION AND DATABASES
267
• Survey data, being the measurements made in
the cave or surface area, details of the people,
instruments, and methods used, and usually also
the resultant reduced data used for the subsequent
map
• Cave maps drawn using solely computer methods
and symbol libraries, some using the Scalable Vector
Graphics (SVG) standard [27]; example vector
software tools: CorelDraw, Adobe Illustrator,
Inkscape (free SVG editor)
• GIS systems allowing the searching, manipulation,
and presentation of combined spatial and nonspatial
data [26]
Summary data and scans can also be stored in hand
held computers (usually now with GPS) for taking
back into the field. This can help record keepers when
identifying and recording new caves.
To help the international use of data across a variety
of spoken languages, UISIC has been promoting the
storing of data as inherently language-independent
numeric codes rather than as text in a specific lan¬
guage. These codes are part of extendable fixed vocab¬
ularies for each field, for example, each rock type
would have a specific numeric code. When presented
on the screen or in a report, the numeric code would
be automatically translated into the chosen national
language. UISIC has so far defined almost 700 fields
and their codes on its website, some of them in
German as well as in English [7].
The use of standards and guidelines can be a real
benefit. Although not obligatory, their use can help
with how to document caves, and also how to inter¬
pret documentation done by others. Here are some
examples [8]:
• UISIC existing: Cave-mapping symbols, surface¬
mapping symbols, survey and mapping grades,
database table designs, speleological subject
classifications, multilingual word list [21] (340 speleo
concepts in 12 languages)
• UISIC in progress or planned: Cave data field
definitions and field value codes, extended subject
classifications, data transfer and archiving format,
minimum documentation standards
• ASF existing: Cave survey and mapping, cave
naming, terminology, cave and karst numbering,
map numbering
Retrieval
The only reason for storing data is to be able to
retrieve it and present it later, often in specific relation¬
ships to other data. So retrieval ability is just as
important as storage. For example, "show me the
name, location and phone number of all owners of
caves that contain protected biological species within
200m of this proposed highway route."
Paper-based retrieval of this kind of information can
obviously be very tedious if not impossible in practice,
hence the trend towards extracting as much information
as possible into digital databases. But there is more to it
than just the digitizing of the data if the retrieved data
are to be usefully understood outside the immediate
environment of its own database. For example:
• Differing understandings of apparently the same data
• Different formats and software for storing the data
which inhibits transferring or consolidating data
between systems
• Data spread over local differing databases, making the
analysis of data over a wider area much more tedious
• Language problems when interpreting data
internationally, for instance, when the data are in a
foreign language
UISIC is working to alleviate these problems by pro¬
moting standard definitions for cave and karst data
fields, numerically coded field values to help the lan¬
guage problem, and a common transfer format to
assist data consolidation, exchange, and archiving [19].
The KIP, an international consortium including UIS,
is working to permit international web-based access to
cave and karst data from around the world, both by
linking to existing databases and by creating new ones
where the need arises [3]. Using portal techniques it is
now possible to query and report on data across sev¬
eral independent databases in different locations in the
world, though the more dissimilar they are the more
set-up work is involved, of course.
A useful new way of displaying retrieved information
is to show it overlaid on Google Earth by embedding
the data in a KML file (Keyhole Markup Language). In
addition to positional and extent information, this can
also display other data values and photographs [25].
Surveying and Mapping
Surveying and mapping, both surface and under¬
ground, are major aspects of cave documentation.
Maps can be used to provide clues for further explora¬
tion; structural detail for geomorphologists; locational
detail for biological, cultural, or other cave contents
(cave inventories); to assist with landowner issues and
engineering works; and so on. They can also make
good wall decoration of course.
Maps could be on traditional paper sheets, multi¬
sheet map books, computer screens, solid models, or
GIS systems [24] (we're still waiting for full-size
ENCYCLOPEDIA OF CAVES
268
D
holograms). Computer screens could show them as
plans, rotatable perspective 3D diagrams, walk-throughs,
or deep, tiled zooms (such as on Google Earth) that
show passage fine detail as well as an overview of
hundreds of kilometers of passage. Laser stations can
allow cave wall detail to be recorded and shown [9]. GIS
systems allow the combining of spatial data with
nonspatial data, and the visual presentation of the results
of spatial queries on that data, for example, thematic
maps [26]. Maps drawn using a computer and symbol
libraries for wall and passage detail are also becoming
common. The surveying and mapping standards
mentioned above help with interpretation.
Traditionally, and also to a large extent today, sur¬
veying was done using a compass, clinometer, tape or
equivalent, and fieldbook because these were cheap,
robust, and readily available. But increasing use is
being made of modern equipment such as laser range¬
finders and even GPS-style radio location [10]. In-cave
processing and plotting are now possible using hand¬
held computers, for example, Auriga software [11]. The
advantage of in-cave plotting to scale, whether by
hand or computer drawn, is that measurement blun¬
ders are more immediately detected and so can be cor¬
rected on the spot. Something not often appreciated by
surveyors is that much more time is spent back in the
office producing the map than is spent doing the sur¬
vey, so taking the extra time to get the survey right
usually pays off—who wants to go back into a distant
cave just to sort out a survey problem?
A common approach is to survey accurately on the
surface by using precise survey instruments, existing
detailed maps, or GPS to accurately locate the entrance(s)
of a cave, and then adjust the typically less accurate
underground survey to fit the accurate entrance posi¬
tions. Where there is only one entrance, radio location
equipment can sometimes be used to relatively accu¬
rately locate on the surface where the end of the cave is.
These are part of the common technique of closing the
loop to distribute the random unavoidable measurement
errors around the survey.
Today several cave survey computer programs
are in common use [20]. However, one issue is the
different format each uses for storing the survey data,
though to some extent they can convert another
program's data. Another issue is the different data
structure used in Europe and the United States. No
commonly used data transfer format yet exists, though
several attempts have been made [12]. This issue
inhibits the consolidation of surveys done over a long
period, and the long-term archiving of survey data
independently of any particular software program.
A diagram on the UISIC data exchange website
shows the various data items involved in a cave sur¬
vey project and the relationships between them [13].
Data Safety
A significant issue with cave documentation is the
safety of the data:
• The physical safety of the data—is there only one copy?
• Preventing the release of sensitive data, which could
be to the detriment of the cave, its values, its
contents, or relations with the landowner, manager,
or data providers
Physically, there needs to be an up-to-date copy held
at a different site in order to guard against total loss by
fire, flood, burglary, or the like. Both locations must be
secure, also against the dispersal of sensitive data, as
there will be a reluctance to provide any data at all to a
central repository if it is not perceived to be safe. While
digital data are relatively easy to back up and store off¬
site, strict routines must still be established to ensure
smooth and effective management of regular backups,
and the restoration procedures in the event of loss.
A problem with digital data is that the whole data
store is more easily stolen as it is so compact and easy
to carry away, leaving no trace that it has been illicitly
copied. Encryption can be a help. Holding the location
data separately, and splitting it into several separately
encrypted files, say, one per cave area, can also reduce
the risk of complete dispersal.
Paper files are more difficult to back up, of course.
Photocopying is expensive, can be unclear, and doubles
the required storage space. Microfiche was a step for¬
ward in that the space taken was much more manage¬
able, but it was still not cheap. But now color scanning is
very cheap and practical, image resolutions can be cho¬
sen to show up obscure detail more clearly than the orig¬
inal, and the images can be used in day-to-day access as
well as being a second copy for offsite backups.
Map backups are sometimes overlooked. Modern
maps are often not a problem as they may already be
digital, or at least drawn on an A4 or A3 sheet that can
readily be scanned. However, sometimes there will be
only one copy of a large-sheet map, perhaps painstak¬
ingly drawn many years ago using only manual tech¬
niques. Some possibilities include a commercial print
or photocopy, a careful color photograph of the map
with a high-resolution quality-lens digital camera, per¬
haps in several partial images later stitched together,
or better, a commercial scan on a large scanner.
Sometimes government agencies will request data.
Despite any iron clad assurances that the data will not be
released to anyone else, experience has shown that this
cannot always be relied upon. With staff changeovers
and department reorganizations, the original agreements
can soon be forgotten. They may be in writing, but after
data escape it is too late: you gave your data to the
Hydrology Department but now the caves have shown
ENCYCLOPEDIA OF CAVES
DOCUMENTATION AND DATABASES
269
up on a public topo map by the Survey Department! It is
certainly a dilemma—the information is there to be used,
not hidden. One technique is not to provide bulk exact
location data, only coarse locations, and to provide exact
locations for specific caves only when actually required.
Publication
The cave and karst information collected is only use¬
ful if people know of its existence, so publication or at
least reference to it in some form is desirable. Three use¬
ful examples that list published speleological material
are shown in [14]. UIS also has a Documentation Centre
[23], hosted by the Swiss Speleological Society, and with
a network of co-operating libraries around the world.
Caves of ... books have always been popular with
cavers, researchers, and land managers, but have also
been contentious. Published locations have been the main
issue. Some people just will not supply any information if
it is to be published. A compromise that has worked is:
• Restrict sales to only the speleo and related
community
• Publish only coarse locations, for example, to the
nearest 10 km
• Provide information about whom to contact if more
detailed information is needed
The web has made the publishing of cave data
much more feasible and low cost. Web access can be
either open to the public or graded by means of user-
names and passwords. A web database can be particu¬
larly useful by allowing:
• Distributed updating of a central repository with
automatic backups
• Wide and graded access
• Complex queries of the data
• A range of presentation alternatives
• With suitable coded data storage and programming,
auto-translation of output into various national
languages
• Online linking with other systems, for example, via
the KIP
An example of a national web database is the KID
in Australia. It follows the proposals of the UISIC field
standards, and its software is open-source, free for
others to use and develop further [15].
DATABASES
Introduction
Cave and karst databases have been referred to
above but do require more elaboration. Basically, a
database is the computer-based systematic storage of
pieces of information in such a way that its contents
can be queried, and the pieces of information that
match the query can then be presented and consoli¬
dated in useful ways. For example, the results could be
presented as a sorted table with just the required col¬
umns, or even as a generated thematically-colored
map if the database was part of a GIS system.
Databases have become much more common in
recent years and have generally been produced as a
flow-on from earlier paper records, being a computer-
searchable extraction and consolidation from them.
But they are now also combining with the spatial capa¬
bilities of GIS systems so that we are seeing a conver¬
gence of factual data systems with mapping systems,
with all their data capable of being queried, analyzed,
and presented in much more useful ways. The UISIC
website contains a list of cave datasets around the
world, but is seriously out of date, so countries are
encouraged to update their entries [16].
A spreadsheet is a common example of a database
but a very rudimentary one: its columns represent the
fields containing the different types of data such as
cave length or depth ; the rows represent different sets of
those data for different items, for example, a row per
cave; and the whole spreadsheet is, of course, a
table representing a collection of one type, for exam¬
ple, caves. While a spreadsheet is convenient and eas¬
ily available, as soon as the data relationships start to
get more complex, such as further entities related to
the caves such as maps, people, or photographs, or
where fields need to be multivalued such as pitch
lengths, then the spreadsheet becomes inadequate and
a full database management system (DBMS) is needed.
Some examples of database software:
• Commercial DBMS: Access, Oracle, Filemaker,
Paradox
• Free DBMS: MySQL, OpenOffice Base
• Commercial GIS: ArcGIS, Maplnfo
• Free GIS: GRASS, QGIS
Components
When using databases it often proves handy to have
at least a basic idea of their main components:
• Entities: An entity is a thing or event which we are
collecting data about, such as a cave or a trip.
• Fields: Fields or elements are notionally labeled places
in the computer where a particular type of data is
stored, for example, the cave name, the cave length,
discovery date, owner type, and so on. Where a table is
being displayed, the fields are usually represented
by the columns.
ENCYCLOPEDIA OF CAVES
270
D
• Records: A record is like a row in a table, and
represents a collection of field values for just one
item, such as for one cave. The first field of each row
is usually a special type of field called a key field,
and serves to uniquely identify and distinguish that
row or instance of the entity from all the other rows.
• Tables: As expected, a table is a collection of rows
and columns, that is, a set of fields for a number of
instances of the particular entity, for example, for a
group of caves.
• Linkages: Sometimes the fields for one entity will be
spread across several separate tables, perhaps for
storage efficiency or perhaps to accommodate
multivalued fields. The corresponding rows in each
table are correlated by all having the same value for
their identifying key field. These linkages are stored
in the database and serve to assist in linking
tables during various queries and views of the data.
• Where the database contains more than one entity,
such as caves, areas, trips, maps, and people, these
will also invariably be related to each other via
linkages. For example, one of the fields in a cave's
data might be specifying a particular map by
containing the map's ID. This type of field in the
cave's table is known as a foreign key, and serves to
link the cave and map entities together.
• Forms: A form is a layout of fields displayed on the
computer screen, perhaps combining fields from
several tables and several entities to present a
particular topic in the data. Forms are typically used
to enter data into the database, as well as to display
its data.
• Queries and reports: A query is a search of the
database looking for records whose field values
satisfy particular criteria. The result would typically
be a table of matching records showing only the
wanted fields. A report is a specification of usually a
more formal layout for the presentation of the
results of a query, often for a printed output. Once a
query or report has been designed, and some are
quite complex, it is usually possible to store them in
the database for later reuse.
Database Issues
The following are some of the issues with using
cave databases. Some apply to paper datasets as well,
but they apply particularly to computerized databases.
• Data accuracy and credibility: When data are in a
computer there is a tendency to believe that they are
correct, yet there is plenty of scope for data to be
wrong. Great care must be taken to verify that the
data have been correctly copied or extracted from
paper records, and that they consist only of facts.
not opinions, guesses, or interpretations. If they do
contain opinions, then the status of this information
needs to be clearly evident to any user. Also: when in
doubt, leave it out. Once a database loses credibility
and cannot be relied on, its value will be much
decreased.
• Safety of the data: The compactness of digital data
makes unauthorized copying easier, even copying
of the whole data store. There is a case for
encryption and splitting up of the data.
• Access policy: As with paper files, there needs to be
an access policy, but it also needs to take into
account the ease of unauthorized copying. There
may also be need for a Data Use Agreement to cover
the use of the digital data.
• Ownership and attribution: This might become an
issue in some circumstances, such as when data are
being supplied to a third party for a license fee. If
several clubs have contributed data to a centralized
database, is the money to be split, and if so, how?
One method would be to implement data
attribution, that is, keep a tally of which
organization supplied what data. UISIC has
produced fields and example table structures to
allow this [17].
• Trust: There is an almost universal desire among
cavers not to let detailed cave information become
public, so for a database to be successful in getting
data contributions, cavers must trust that the
custodian will safeguard it in accordance with
agreed policies.
• Specifications: What fields will be included in the
database? The fields and definitions being used
need to be known by the users if they are to make
effective use of it. There is a good case for using
already defined standard fields to help
understanding and data transfer [7].
• Central repository: Many databases have been
initiated, designed, and operated by enthusiasts in
local clubs. This means that for a while at least, they
do have someone keen to look after them. However,
the disadvantages are that (1) it depends on one
person's continued availability, and (2) anyone
wanting to analyze data over a wider region needs
to separately negotiate and work with multiple
databases, probably all of different designs. So,
should the data instead be held in a central
repository such as for a state, or better, for a whole
country? This too can have disadvantages such as
loss of local control and local enthusiasm, the "all
eggs in one basket" syndrome, and harder staffing
problems. One solution used in Austria is for the
central national database to hold only summary
data, periodically uploaded from local databases,
and with the detailed data held only locally. This
ENCYCLOPEDIA OF CAVES
DOCUMENTATION AND DATABASES
271
also reduces the amount of central work needed.
A similar approach has been used in Australia,
though it is not yet working as well as in Austria.
• Language: Not everybody speaks English, of course,
so UISIC has been encouraging the storage of data
as language-independent numeric codes wherever
possible.
• Unique identifiers: Each record in a database should
have a globally unique identifier as its key field to
prevent clashes and to identify the data if they are
transferred to another database. In the case of cave
records, this should be distinct from the normal
public cave number as these are not necessarily
unique on a global scale or permanent.
• Data transfer: The chances are that at various times
data will need to be transferred into and / or out of a
database. Unless transfer systems are set up, this can
be a tedious, labor-intensive task. The obvious
commonly used mechanism is the export and import
of comma-separated-value files. However, when the
databases are different, this needs investigation and
manual manipulation, usually at the receiving end.
This issue is discussed in more detail below.
Data transfer and archiving
There is often a need to transfer data between data¬
bases such as copying data from a club database into a
state wide database, or supplying data to a government
agency. Also needed is the archiving of data indepen¬
dently of any particular database software.
For a valid data transfer we need three things: an
agreed meaning of the data, a unique identifier for
each record, and an agreed transfer format:
• Meaning: The data must mean the same thing at each
end of the transfer. For example, each end might have
a field called length, but at one end the value might
mean meters while at the other it might mean feet.
• Identifier: To avoid a possible clash at the target end,
each end needs to use globally unique record
identifiers. UISIC is recommending that these consist
simply of a unique organization code identifying the
original creating organization, followed by a serial
number. They do not also need to indicate the type
of record. The organization code consists of the ISO-
3166 two-letter country code followed by a unique
3-letter organization code within that country [18].
• Format: So that the recipient can interpret the data
received, it needs to be in an agreed format, for
example, a special transfer format or markup
language, or specified comma-separated-values.
UISIC has a project to define standards that permit
these three requirements [19]. Part of this is a cave and
karst data fields library. Although it currently contains
almost 700 field definitions, many more fields are
needed, so to be successful it will need the cooperation
of cavers and karst scientists from the various disciplines
to contribute any extra fields needed for their work.
The transfer format is planned to be a markup lan¬
guage called CaveXML [12] utilizing international
XML (Extensible Markup Language) standards. It is
expected to be compatible with the existing interna¬
tional standard Geographic Markup Language (GML).
Database Advantages and Disadvantages
Advantages
• A compact, convenient and systematic way of
storing a large amount of cave and karst data
• Easily updated over the years without increasing
the bulk
• Allows easy and tailored retrieval and presentation
of the data
• Allows easy backup of the data
• Facilitates consolidation of the data to allow queries
that could span a state, a country, or the whole
world
• Allows a range of output formats for retrieved data
• Allows easy interfacing to spatial data in GIS
systems
• A web database can allow easy remote access and
remote updating, such as local clubs updating a
state or national database
Disadvantages
• Requires a greater level of technical expertise to set
up and operate
• Can increase the risk of data dispersal
• Is a secondary source, a step removed from the
original documents, so is susceptible to transcription
errors
• Requires periodical upgrading of the system to
counteract the inevitable obsolescence of software
and hardware
WHERE TO FROM HERE?
We still have a long way to go:
• To enable easy access to information from around
the world we need further development and
support of the KIP.
• To help with better recording of cave and karst data
we need better communication and sharing of cave
documentation know-how.
• To make cave data more usable by more
organizations we need greater use of standards.
ENCYCLOPEDIA OF CAVES
272
D
• To enable easy transfer of data between different
organizations and research institutions we need our
library of standard field definitions, our
multilingual controlled vocabularies, and our
standard transfer format.
• To build up our library of standard field definitions
we need the cooperation of cavers and karst
scientists for the various disciplines.
• To help UISIC implement its many projects to aid
cave documentation we need more volunteers.
• We need accurate positions and altitudes for all our
caves to assist geomorphological, hydrological, and
other scientific studies.
• We need much greater uptake of GIS systems by
caving clubs to make their hard-won information
even more useful.
• We need GIS systems showing the karst areas in each
country superimposed on existing protective reserves
to see where further karst protection is needed.
• We need better access to information about
expedition data.
• We need a greater level of awareness, understanding,
and active cooperation between scientists and cavers
to get maximum benefit from the scientists' knowledge
and the vast amount of data that cavers collect.
These are just some of the challenges ahead.
Acknowledgments
My thanks first to Keith Wheeland who wrote the excellent first edi¬
tion of this article, focusing mainly on the database aspects—most of
Keith's points are included in this edition; he has also kindly perused
this expanded version. Thanks also to all the cave documentation
people in many countries who have shown me such hospitality and
taken the time to explain their systems since my first Cave
Documentation Safari in 1972. Thanks also to all the members of ASF
clubs whose hard work and cooperation enabled cave documentation
ideas to be turned into reality. And last, thanks to the editors of the
Encyclopedia of Caves for the opportunity to include this cave and
karst documentation article.
Bibliography
References to further detail are given by numbers cited in the text.
Because of the large number of website citations, this nonstandard
method of citation is used.
1. Stone, R. W. (1968). Pennsylvania caves. Commonwealth of
Pennsylvania Department of Internal Affairs, Topographic and
Geologic Survey, Bulletin G3, 1—63.
2. Matthews, P. (Ed.), (1968). Speleo handbook. Sydney: Australian
Speleological Federation.
3. Karst Information Portal (KIP), www.karstportal.org.
4. Matthews, P. (Ed.), (1985). Australian karst index 1985. Sydney:
Australian Speleological Federation.
5. Lake, M., & Matthews, P. (2005). An open-source web-based
national cave database. Proceedings of the 14th International
Congress of Speleology, 21—28 August 2005, Kalamos, Greece (Vol. 2,
pp. 473—475). Athens: Hellenic Speleological Society.
6. Expedition aids:
• The FSE/UIS International Caving Expeditions (ICE) database
project. www.eurospeleo.eu/index.php?option = com
content&view = article&id = 24&Itemid = 33.
• Curtis, A. (2009). Troggle: A novel system for cave
exploration information management. Proceedings of the 15th
International Speleological Congress, Kerrville, Texas, USA, July
19—26, 2009 (Vol. 1, pp. 431—436). International Union of
Speleology, www.code.google.com/p/troggle.
• Victorian Speleological Association's (AU) web browser method (in
progress). Contact: matthews@melbpc.org.au.
• UIS code of ethics for cave exploration and science in foreign
countries, www.uis-speleo.org/ethic-en.html.
7. UISICs draft field definitions and numeric codes, www.uisic.uis-spe-
leo.org/exchange/atenlist.html.
8. Cave documentation standards and guidelines:
• General: www.uis-speleo.org/guides.html.
• Survey: www.uisic.uis-speleo.org/wgsurmap.html.
9. Canevese, E. P., Tadeschi, R., & Forti, P. (2009). Laser scanning
use in cave contexts: The cases of Castellana (Italy) and Naica
(Mexico). Proceedings of the 15th International Speleological
Congress, Kerrville, Texas, USA, July 19—26, 2009 (Vol. 3, pp.
2061—2067). International Union of Speleology.
10. Wenger, R., & Jeannin, P.-Y. (2009). Towards a positioning system
for the subterraneous world (U-GPS). Proceedings of the 15th
International Speleological Congress, Kerrville, Texas, USA, July 19—26,
2009 (Vol. 1, pp. 612—617). International Union of Speleology.
11. LeBlanc, L. (2009). The Sierra Negra in a PDA: Expedition-wide
electronic cave surveying. Proceedings of the 15th International
Speleological Congress, Kerrville, Texas, USA, July 19—26, 2009
(Vol. 3, pp. 2095—2099). International Union of Speleology.
www.speleo.qc.ca/auriga.
12. Cave survey data transfer format attempt examples:
• Dotson, D., & Wookey (1993). HTO Survei/ Data Interchange
Format. www.chaos.org.uk/survex/cp/CP02/CPoint02
,htm#Art6.
• Lake, M. (2001). CaveScript. www.speleonics.com.au
cavescript.
• UISIC (2001). CaveXML. www.uisic.uis-speleo.org
wgsurmap.html#exchange.
• van Ieperen, T., & Petrie, G. (1997). RosettaStal (a format
convertor), www.resurgentsoftware.com/rosettastal.htm.
13. Survey data model, www.cavexml.uis-speleo.org/task-sdm
,html#diagram.
14. Some indexes to published material:
• UIS/SSS Speleological Abstracts/Bulletin Bibliographique
Spele'ologique (SA/BBS). www.ssslib.ch/bbs.
• Karst Information Portal, www.karstportal.org.
• Northup, D. E., Diana E. Northup, Emily Davis Mobley,
Kenneth L. Ingham III, & William W. Mixon. (Eds.), (1998).
A guide to speleological literature of the English language:
1794—1996. St. Louis: Cave Books.
15. Example national web database: Australian Karst Index Database
(KID), http://kid.caves.org.au/kid.
16. Country datasets: www.uisic.uis-speleo.org/contacts.html.
17. Example fields and table structure for data attribution: www
.uisic.uis-speleo.org/exchange/histables.html.
18. UISIC proposal for unique record identifiers: wrww.uisic.uis-speleo
.org/exchange/exchprop.html#rident & www.uisic.uis-speleo
.org / exchange / exchprop. html#dident.
19. UISIC data exchange project: www.uisic.uis-speleo.org/exchange/
exchprop.html.
20. Lists of cave survey programs:
• www.resurgentsoftware.eom/winkarst.html#Software%20Links/
• wwrw.survex.com/related.html.
ENCYCLOPEDIA OF CAVES
DOCUMENTATION AND DATABASES
273
21. Multilingual word list and cave/karst glossaries:
• Multilingual word list: www.uisic.uis-speleo.org/lexuni.html.
• Glossary list: www.uisic.uis-speleo.org/lexintro.html#refs.
22. UIS Informatics Commission: www.uisic.uis-speleo.org.
23. UIS Documentation Centre: www.ssslib.ch/bbs/public/anglais/
centreDoc.htm.
24. Example of mapping a long cave: Horrocks, R. D., & Austin, D. C.
(2009). Making a digital map of Wind Cave, Wind Cave National
Park, South Dakota. Proceedings of the 15th International Speleological
Congress, Kerrville, Texas, USA, July 19—26, 2009 (Vol. 3, pp.
2091—2094). International Union of Speleology.
25. KML for Google Earth: Wernecke, J. (2009). The KML Handbook:
Geographic visualization for the web. Boston, MA: Addison-Wesley,
Pearson Education Inc. Also: www.earth.google.com/support/
bin/answer.py?hl = en&Also:answer = 148119.
26. Some CIS references:
• General GIS introduction: www.rockyweb.cr.usgs.gov/
outreach/articles/nss_gis_article.pdf.
• Hose, L. (Ed.) (2002). Journal of Cave and Karst Studies (Special
issue: Intro to cave and karst GIS), 64(1). Huntsville, AL:
National Speleological Society.
• Examples of cave/karst GIS using ArcGIS: www.esri.com/
industries / cavekarst.
• Maplnfo Professional: www.pbinsight.com/welcome/
mapinfo.
• GRASS GIS: www.grass.osgeo.org/ (free, open-source).
• Quantum GIS (QGIS): www.qgis.org/ (free, open-source).
27. Sample cave maps using SVG: www.carto.net/neumann/caving/svg.
ENCYCLOPEDIA OF CAVES
ECOLOGICAL CLASSILICATION OL
SUBTERRANEAN ORGANISMS
Eleonora Trajano
Universidade de Sdo Paulo, Brasil
Classifying, that is, grouping entities according to
shared characteristics, is part of the human nature,
constituting a first step to understand the surrounding
world at different intellectual and emotional levels.
Therefore, it is no surprise that the early researchers
on cave faunas were concerned about classifying those
organisms, not only into general systematic systems
(reflecting their evolutionary relationships), but also
into ecological systems, in view of the particularities of
the selective regime in subterranean habitats, reflected
in the morphology, distribution, and biology of organ¬
isms found there.
The hypogean (subterranean) domain contrasts
greatly with the epigean one because of its permanent
absence of light and, thus, of both photoperiodicity
and photoautotrophy, tending toward food scarcity
and environmental stability. Two different processes
representing distinct steps in the evolution of subterra¬
nean lineages are colonization, which leads to the estab¬
lishment of hypogean populations that spend at least
part of their life cycles in this habitat, and genetic isola¬
tion, which may give rise to exclusively subterranean
taxa. These two processes may be concurrent, but
more often isolation postdates colonization and derives
from independent events. A minimum requirement for
successful colonization of the hypogean habitat is that
colonizers not be dependent on light as a primary
source of energy or for orientation cues. If the energy
sources available in the hypogean habitat are sufficient
for the physiological requirements of the species and
the individuals are able to find food and mates for
reproduction nonvisually, then they can complete their
life cycle without leaving the subterranean habitat.
Subterranean populations are subject to selective
regimes more or less distinct from those in the epigean
environment, including the cessation of typically epi¬
gean selective pressures such as those related to light.
If genetically isolated, such populations may differenti¬
ate due to the interruption of genetic flow from epigean
individuals, causing cessation of indirect influence from
surface conditions.
As soon as humans became aware of cave organ¬
isms, they noticed the existence of very different,
"bizarre" animals living there together with "normal"
(generally similar to epigean) animals. This led to the
first classification of cave animals, published by J.C.
Schiodte in 1849, with four categories based on habitat,
namely, lighting and substrate: animals, respectively,
in shadow, twilight, obscure zones, and on flowstone
in the obscure zone ("animaux des ombres, crepuscu-
laires, des regions obscures, et des regions obscures a
concretions stalagmitiques"). A similar classification
based on habitat was proposed in 1882 by G. Joseph,
including: inhabitants of cave entrances, illuminated
regions, and with variable temperature; inhabitants of
middle, twilight regions; and, inhabitants of deep,
aphotic regions with constant temperature (Racovitza,
2006).
In 1854, J.R. Schiner proposed a classification based
on distribution and ecology: (1) occasional visitors Qiotes
occasioned )—animals found in caves, but also at surface,
everywhere "where they encounter conditions proper to
their way of living"; (2) troglophiles (troglophiles )—ani¬
mals living in regions where daylight still penetrates,
which exceptionally may be found at surface or which
have only light-loving representative forms; (3) troglo-
bites (troglobies )—exclusively cavernicoles, never found
in epigean habitats besides exceptional events such as
floods (Racovitza, 2006).
This classification was modified in 1907 by E.
Racovitza, who substituted "occasional visitor" with
"trogloxene," redefining these categories: (1) troglox-
enes (trogloxenes ) are either lost or occasional hosts, the
latter attracted by humidity or by food, but they do
not live here always and not reproduce here; (2)
Encyclopedia of Caves.
275
© 2012 Elsevier Inc. All rights reserved.
276
E
troglophiles (troglophiles) inhabit continuously the sub¬
terranean environment, but preferentially in superficial
regions; they reproduce here often, but can be encoun¬
tered also outside; (3) troglobites (troglobies ) have as
habitat exclusively the subterranean environment and
stay preferentially in its deepest parts; they are very
modified and offer the most profound modifications to
life in dark. Troglophiles provide in all periods the
main source of troglobites and they are the first coloni¬
zers when a new region of the subterranean environ¬
ment is offered (Racovitza, 2006).
The Schiner—Racovitza classification is widely used
by speleobiologists up to the present, being the basis
for new proposals in the past century, which differed
mainly in terminology and conceptualization of the
third category. In fact, trogloxenes have been the main
point of disagreement, including from animals with a
well-defined ecological relationship with subterranean
habitats, such as bats, to accidentals surviving there
for variable periods. Some of the twentieth-century
classifications were quite complicated, such as that by
T. Pavan, in 1944 (Sket, 2008), which considered seven
classes (eu- and subtrogloxenes, aphyletic and phyletic
trogloxenes, sub- and eutroglophiles, and troglobites).
Sket (2008) clearly stated the minimum criteria for an
acceptable, biologically meaningful classification as
being ecologically and evolutionarily adequate; appli¬
cable in association with the current state of biological
knowledge; close to the traditionally used vocabulary;
and having the greatest concordance with the historical
priorities.
Some of the early classifications included the notion
of "true" versus "false" cavernicoles, the former with
features presenting a certain adaptation to the life in
darkness and the latter without such characteristics,
not differing from closely related epigean forms.
A first step to apply any classification is to define
subterranean animals (encompassing populations in
non-cave habitats). Trajano (2005) considered subterra¬
nean organisms those regularly found in the subterra¬
nean biotope and for which this is part of or the whole
natural habitat. In contrast, accidentals may be intro¬
duced into caves by mishap (by being washed into
caves or falling through upper openings, for instance)
or when entering in search of a mild climate; although
accidentals can survive temporarily, their inability to
properly orient and to find food leads to their eventual
demise.
From the evolutionary point of view, subterranean
organisms (cavernicoles sensu latu) may be defined as
evolutionary units responding to subterranean selec¬
tive regimens. Subterranean habitats would provide
resources—for example, food, shelter, substrate, cli¬
mate—which affect survival and reproductive rates.
Such units have a historical connectivity and therefore
may be classified as systematically meaningful biological
systems.
In contrast, accidentals are evolutionary culs-de-sac.
From an ecological point of view, accidentals are
potential resources for subterranean organisms (food,
substrate, etc.). Resources per se have no historical con¬
nectivity; therefore, when an organism becomes a
resource, it makes no sense to classify it further, identi¬
fying species or higher categories (a biologically mean¬
ingful classification of food would be according to its
accessibility, such as availability, protection against
predation, etc.), nutrition value, and so on. Moreover,
accidentals are grouped by a negative trait (they are
not subterranean organisms, as herein defined); there¬
fore they do not constitute a biologically meaningful
group. In conclusion, it is clear that the "accidental"
concept has a different nature and therefore should
not be included in the same Schiner—Racovitza
system.
The North American speleobiologist T.C. Barr pro¬
vided objective and clear definitions, followed by
many other specialists such as Holsinger and Culver
(1988): (1) trogloxenes are regularly found in the subter¬
ranean habitat, but must leave it during some period(s)
in order to complete their life cycles (usually because
hypogean food sources are insufficient for the species
requirements); (2) troglophiles are able to complete their
life cycles both in the hypogean and in the epigean
environment, forming populations in both habitats,
with individuals commuting between them and main¬
taining genetic flow between these populations; and
(3) troglobites are species restricted to the subterranean
domain and are usually characterized by apomorphic
character states related to the hypogean life ( troglo-
morphisms) such as reduction until loss of eyes and
dark pigmentation. A fourth category, that of acciden¬
tals, was also mentioned. Sket (2008) suggests, for
semantic reasons, that troglobiont is preferable to
troglobite.
Some authors use distinct sets of words for subterra¬
nean aquatic species— stygoxene, stygophiles, and stygo-
bites (stygobionts). This terminology has been proposed
by Gibert et al. (1994) as descriptive of groundwater
fauna, because it includes organisms living in nonkarst
subterranean habitats, such as porous aquifers.
Another version, used by some Europeans, is based
on Pavan—Ruff o's classification (Sket, 2008), which dif¬
fers basically in terminology: eutroglophile instead of
troglophile in Barr's classification, subtroglophile instead
of trogloxene, and trogloxene instead of accidental. These
terms are misleading because they tend to suggest an
evolutionary progression or incomplete adaptation.
Subtroglophiles are not quasi-eutroglophiles. At a
given moment, each population is either a troglophile
or a trogloxene (in Barr's sense). Because food is
ENCYCLOPEDIA OF CAVES
ENTRANCELESS CAVES, DISCOVERY OF
277
frequently the limiting factor, individuals may leave
hypogean habitats periodically to eat (thus becoming
trogloxenes), under special circumstances, and in
highly eutrophic caves (e.g., bat caves), individuals of
such species may form troglophilic populations.
However, this corresponds to a very particular ecologi¬
cal situation, not representing an evolutional progres¬
sion. In fact, as stated by Sket (2008), the boundary
between epigean and hypogean habitats and between
competing surface and cavernicolous animals may
change (move) if trophic resources are enhanced. Other
ecological traits, such as differences in population den¬
sities, are also irrelevant, because they represent
responses to specific environmental conditions. It is
noteworthy that, probably due to higher competition
pressures, for a given species the densities of epigean
populations are frequently lower than those of the
troglophilic ones.
One of the most interesting and useful recent con¬
cepts in speleobiology is the distinction between
source and sink populations: a sink population, if cut
off from all other migrants, eventually becomes extinct,
whereas a source population has excess production
and continues to grow if isolated (Fong, 2004). Sink
populations are habitat-level phenomena, correspond¬
ing to stranded groups of individuals in habitats less
than suitable (in terms of space, food, and other
resources necessary for self-sustained, source popula¬
tions). Therefore, their presence in such habitats is
unpredictable.
When applied to the Schiner—Racovitza classifica¬
tion, this conceptualization leads to clearer and biolog¬
ically meaningful definitions of the three categories
classically recognized: (1) troglobites (stygobites) corre¬
spond to exclusively subterranean source populations;
sink populations may be found in surface habitats; (2)
troglophiles (stygophiles) include source populations
both in hypogean and epigean habitats, with indi¬
viduals regularly commuting between these habitats,
promoting the introgression of genes selected under
epigean regimes into subterranean populations (and
vice versa)- (3) trogloxenes (stygoxenes) are instances of
source populations in epigean habitats, but using sub¬
terranean resources (in the so-called obligatory troglox¬
enes, all individuals are dependent on both surface
and subterranean resources). It is noteworthy that sink
populations do not fit the Schiner—Racovitza scheme,
unless one considers stranded troglobitic or stygobitic
individuals as part of the subterranean source popula¬
tion from which they originated.
Bibliography
Fong, D. W. (2004). Intermittent pools at headwaters of subterranean
drainage basins as sampling sites for epikarst fauna. In W. K.
Jones, W. V. Coord, D. C. Culver & J. S. Herman, Epikarst (pp.
114—188). Proceedings of the symposium held October 1 through
4, 2003, Shepherdstown, Special Publication 9. Charles Town,
WV: Karst Waters Institute.
Gibert, J., Stanford, J. A., Dole-Olivier, M. J., & Ward, J. V. (1994).
Basic attributes of groundwater ecosystems and prospects for
research. In J. Gibert, D. A. Danielopol & J. A. Stanford (Eds.),
Groundwater ecology. San Diego, CA: Academic Press.
Holsinger, J. R., & Culver, D. C. (1988). The invertebrate cave fauna
of Virginia and a part of eastern Tennessee: Zoogeography and
ecology. Brimleyana, 14, 1—162.
Racovitza, E. G. (2006). Essay on biospeleological problems: French,
English, Romanian. [Facsimile of the publication Essai sur les pro-
blemes biospe'ologiques (1907), translated by D. C. Culver &
O. Moldovan.] Cluj-Napoca, Romania: Institul de Speologie "Emil
Racovitza."
Sket, B. (2008). Can we agree on an ecological classification of subter¬
ranean animals? journal of Natural Elistory, 42(21—22), 1549—1563.
Trajano, E. (2005). Evolution of lineages. In D. C. Culver & W. B.
White (Eds.), The encyclopedia of caves (pp. 230—234). San Diego,
CA: Academic Press.
ENTRANCELESS CAVES,
DISCOVERY OF
Nevin W. Davis
Butler Cave Conservation Society, Inc.
INTRODUCTION: HUMANKIND’S NEED
FOR EXPLORATION AND DISCOVERY
Humankind has always had the need to explore
and use caves. Early Americans entered caves in
the Virginias, Kentucky, and elsewhere. Polynesian
Hawaiians have explored lava tubes on the Big Island
of Hawaii from early times to the present. The
Hawaiians used the tubes for ceremonial purposes,
collecting water, and burials, but from the evidence
left by charcoal from cane torches in remote places,
they also must have enjoyed exploration for its own
sake. Today there is a large group of cavers whose
exploration instinct is satisfied with the traversing of
known passages and the mapping or remapping of
passages that are already "in the book" or "previously
enjoyed" (slang terms for cave passages that are well
known by the caving public), but there is also a select
group of cavers who are satisfied only with explora¬
tion and mapping of "virgin" passages. So the ques¬
tion arises "What is a virgin cave passage?" Purists
would say that a virgin cave passage is a passage that
has never been seen by humans. A looser definition
would be a passage that has never been recorded or
described even though it may have been viewed years
before or by aboriginal man. In this article we use the
first definition.
Encyclopedia of Caves.
© 2012 Elsevier Inc. All rights reserved.
278
E
A virgin cave passage can be found in known caves.
Many times exploration will yield a passage that
appears to have never before been entered. A passage
that has no footprints or scrape marks is a good candi¬
date for being virgin but there is no assurance in
today's world of well-explored caves unless some
obstacle is passed that would preclude others from
having been there. These include but are not limited to
the following:
1. Water-filled sumps that have no written or oral
record of being dived
2. Very remote passages in a project cave that no
casual explorer would have gotten to
3. Passages that are blocked by breakdown or dirt fill
that, when removed, reveals extensions
4. Deep pits
Virgin cave passages can also be found in another
interesting way. Suppose the cave is totally unknown
and has no entrance. (An entrance is an opening large
enough for a human to enter.) Absent an entrance a
cave is very likely to be virgin. In Virginia the number
of entrances per cave varies widely. One cave has
16 entrances, Clark's Cave has 8 entrances, most caves
have 1 entrance, but some of the best known caves
originally had no passable entrance. They were dug
open. One of the best known of these is Butler
Cave—Sinking Creek System with 26.8 km of passage
that was dug open in May 1958. Coincidentally, in the
same year Rane Curl published a paper entitled "A
Statistical Theory of Cave Entrance Evolution" (Curl,
1958). In it he came up with a statistical way to predict
the number of entranceless caves in a region with a
distribution of caves with 0 through 4 entrances. For
example, in West Virginia at that time, out of a popula¬
tion of 257 caves with lengths greater than 30 m,
228 caves had 1 entrance. His theory gave the number
of entranceless caves in this area as 2405. This is an
impressive number, but the theory also leads to a dis¬
tribution function of lengths of caves within this set
with 0 through 4 entrances. Unfortunately, 50% or
1200 of the entranceless caves would have passage
lengths of less than 55 m (1200 would also have
lengths greater than 55 m), while 200 of the single¬
entrance caves would have a length greater than 55 m.
This means that any successful effort to dig open a
cave would be met with a large number of short caves!
Fortunately, the search for entranceless caves in the
real world is quite different from the theoretical
approach. In the real world, some intelligence can be
applied to the effort to better the odds of finding
something significant. Here are some recent numbers
from the Virginia Speleological Survey database. In
Virginia there are 2601 caves with a length greater
than 9 m. Of this population, 2066 caves have one
entrance and 251 caves were dug open. The average
length of the single-entrance caves is 239 m, while the
average length of the formerly entranceless caves is
809 m. Clearly something is happening here! Cavers
are not digging for short caves or are only reporting
their successes.
ENVIRONMENTAL SETTINGS OF
ENTRANCELESS CAVES
The discussion that follows is based on experience
in the Appalachians in limestone terrains. Insight that
applies to the Appalachian Mountains in the eastern
United States should also apply to many other lime¬
stone terrains with significant rainfall. The environ¬
mental variables that produce significant long and
deep known caves obviously likewise apply to en¬
tranceless caves. These include the following:
1. Large upland catchment areas directing the runoff
into sinkholes and losing streams. Catchment areas
can vary from under 260 hectares to more than
26,000 hectares.
2. Entrenched rivers with large karst springs.
3. Thick, contiguous beds of limestone.
4. Insoluble caprock that overlies some portion of the
limestone, protecting it from surface solution. This
condition occurs only in small areas in the folded
Valley and Ridge Province of the Appalachians but
is found over large areas in the Cumberland Plateau
and in Central Kentucky near Mammoth Cave.
5. Moderately high relief. Relief of 30—60 m, as is
found in the valleys of central Pennsylvania, leads
to much shorter caves than are found in the
Burnsville Cove in west central Virginia, where the
relief from the highest cave entrance to the spring
exceeds 244 m.
All of these environmental variables need not be in
their most desirable range to produce long and deep
caves. Hosterman Pit in central Pennsylvania sits
under a hill in the center of Penns Valley. Nearby Pine
Creek is only entrenched 60 m below the hilltop, the
karst springs run at less than 30 liters/sec, and there is
no caprock or large catchment area; however, the cave
has more than 1.6 km of passage and is 59 m deep.
Another consideration in searching for caves is
entrance lifetime. Caves are long-term features under
the landscape with lifetimes measured in millions of
years, whereas entrances to them are fleeting features
with lifetimes measured in millennia. One factor affect¬
ing entrance lifetime in a temperate climate is airflow.
Bedrock in any climate will assume the mean tempera¬
ture of the region. This, of course, means that the air
temperature in the depths of the caves surrounded by
ENCYCLOPEDIA OF CAVES
ENTRANCELESS CAVES, DISCOVERY OF
279
that bedrock also has this temperature. In the winter,
outside air temperatures are often below freezing in
the Appalachians while the cave temperatures are near
10°C. Almost all air movement in caves is driven by
the chimney effect, that is, warm air rises. This means
that warmed cave air exits the upper entrances of
caves to draw the cold outside air into lower entrances.
Eventually, the temperature of the rock at lower
entrances drops below freezing causing any water in
the rock to freeze, expand, and fracture the rock. This
can also occur in caves with only one entrance because
significant quantities of air can penetrate the soil, cov¬
ering potential high entrances including domes and
high ceiling fissures. The large surface area over a cave
means that air velocities through the soil can be insig¬
nificant, while the velocities in the smaller cave pas¬
sages can be meters per second. The upshot of all of
this is that lower entrances to caves have very short
lifetimes.
Another consideration in the Appalachians where
most of the ridges are covered with deciduous trees is
that leaves and twigs will soon cover and block small
vertical entrances. Pits less than a meter in diameter
can be totally blocked in one season. Leaves blocking a
small entrance are soon followed by roots and more
leaves and it is not long before all traces of an entrance
are gone. Humankind does not help this condition
much since it seems that a natural instinct is to cover
holes with poles (for safety?) and the cycle of covering
is repeated in an unnatural way, unless of course
someone steps on the hidden, rotten poles in a few
years and falls into the pit.
Humankind also has had an effect on valley and
ridge margin sinks, all potential cave entrances. Any
depression extends an invitation to be filled with trash.
Many cave entrances have been lost this way. In a
sinkhole dig in Burnsville Cove, Virginia, a domestic
pig skull was found under 2 m of soil with some 50- to
70-year-old domestic trash near the surface. Because
this sinkhole receives no drainage from nearby fields,
the 2 m of compacted soil covering the skull took no
more than 200 years to accumulate from the hill above.
This gives an idea of rates of infilling of small sinks.
STALKING THE ELUSIVE
ENTRANCELESS CAVE
The same moving air that can destroy a lower
entrance to a cave can help find an upper entrance. In
the winter when the ground is covered with snow,
potential cave entrances can be found by looking for
melted spots. Remember that air does not necessarily
need a large opening to move through. It can move
through a soil covering as well as loose rocks. The
most sensitivity to this effect can be had with early
season snows before the ground freezes. This will also
lead to more false positives because rotting stumps
and other such sources of heat can also melt the snow.
The best condition to search for snow melt is with a
new snowfall in midwinter with an overcast sky, since
sunlight can also give false positives by shining
through shallow snow cover onto rocks and melting
the snow. This is a tried-and-true method that has led
to countless new caves.
Stick to looking in smaller sinks. The smaller the
better because larger sinkholes have more material
blocking a potential entrance and will require a more
protracted dig. Give preference to sinks with a bedrock
outcrop. These are more likely to lead to a cave if for
no other reason than that there is at least one solid
wall to follow when digging!
A straight 1.5-m digging bar can be a helpful tool to
probe suspected sites. In a small sinkhole, if the bar
sinks easily into loose, organic material this is a poten¬
tial dig site. Many times a 1.5-m bar will punch
through into space at the top of a dome pit or fissure.
If the bar indicates that the fill in a sink is consolidated
material without organic content, it is probably a sign
to look elsewhere.
The topography in the Valley and Ridge Province of
the Appalachians features major parallel ridges of clastic
rocks and oftentimes karstic valleys. This means that the
clastic ridges serve as the catchment areas to direct
recharge into the valley margin sinkholes. These sinks,
although they are potential entrances to the caves that
underlie the valleys, are poor choices for dig sites unless
they are known to have been entrances in the past. The
reason for this is that the bed load in the sinking streams
tends to choke any entrances. An effort to open entrances
in these locations may require equipment such as a
trackhoe to get anywhere. At the base of Jack Mountain
near Burnsville Cove in Virginia, Cycle Sink provides an
example of this. This sink was a collapse in the bed of a
sinking stream that reached 6 m depth without digging.
Digging efforts there resulted in very little progress until
a storm completely removed the collapse, filling it and
leaving no sign that it ever existed. The stream now
flows past this area before it sinks.
In a similar situation where the surface is composed
of limestone outcrops, cavers did not dig at the stream
sink. Instead, at a Giles County site in Virginia, the dig
was 9 m from the stream sink and above it. This was a
place where some air exited. The dig is now the
entrance to a significant Virginia cave.
Traditionally, the way to find caves is to talk to
landowners, hunters, ginseng hunters, and other out¬
door people. The problem often is asking the right
questions, particularly when trying to find entranceless
caves. Ask about snow melt spots or places that may
ENCYCLOPEDIA OF CAVES
280
E
have entrances which are now covered over. If all else
fails in the search for the elusive entranceless cave, go
where there are already known long and/or deep
caves. Search nearby or in an adjacent drainage and
success will surely follow.
EXAMPLES
The Burnsville Cove in west central Virginia consists
of a synclinal valley surrounded by ridges composed of
clastic rocks. The valley is bisected by Chestnut Ridge, an
anticlinal structure. Butler Cave—Sinking Creek System
and its drainage lie on the western side of this ridge. The
entrance to Bobcat Cave, a significant cave at 15 km in
length, lies high on the western side of the ridge but the
cave has passage on both sides of the ridge. Of special
interest to cavers is the drainage on the eastern side of the
ridge, especially to the southwest where the recharge
area and the majority of cave passage must be. Bobcat
Cave ends in breakdown and tight fissures to the south¬
west. Ridge-walking by the author over the years has led
to the discovery of innumerable snow melt spots on
Chestnut Ridge. In March 1991 a crew dug at one of
these, a 1.2-m-diameter sink high on the eastern side. In
20 minutes a narrow fissure had been opened. Below was
a 9-m drop followed by a second drop into a room. Two
more drops lead to Strychnine Canyon, a tight, crawly
passage over 274 m long that eventually ends in a large
phreatic passage at Ghost Hall. This is the beginning
of 7.5 km of passage that eventually connects to Bobcat
Cave. The ultimate goal of finding a series of passages
upstream to the recharge area has not been met, but
20 minutes of work to open access to more than 7.5 km of
cave passage isn't bad!
Buckwheat Cave is a small, pleasant, 670-m-long
cave in a ravine on the western side at the base of
Chestnut Ridge. The entrance is 4.6 m above the base of
the ravine on the north side. An ephemeral stream col¬
lects in the ravine and only flows past the entrance in
the wettest of times. This cave is part of the population
of entranceless caves that were opened by digging.
With this knowledge cavers looked at the next ravine to
the north. Topographically, it is very similar and even
has a small sink 1.5 m above the bottom of the ravine
on the north side. Stratigraphically, the sink is located
in the same limestone unit as Buckwheat Cave. The
sink is 2.4 m in diameter and only 0.6 m deep. It had no
snow melt and the fill was rather tight and inorganic.
Optimistic cavers spent one day digging there in March
1999 and after sinking a 4-m shaft, opened aptly named
Blind Faith Cave. The cave is now 1019 m long and
48 m deep and comes very close to connecting to
another excavated cave. Battered Bar Cave.
The first example described here is an illustration of
the benefits of following the air. The use of snow melt
in finding new caves is a powerful tool that can lead to
minimum effort in opening them. The second example
shows that intimate knowledge of a particular area can
lead to discovery even if the signs are not otherwise
favorable.
See Also the Following Articles
Entrances
Geophysics of Locating Karst and Caves
Bibliography
Curl, R. L. (1958). A statistical theory of cave entrance evolution.
National Speleological Society Bulletin, 20, 9—21.
ENTRANCES
William B. White
The Pennsylvania State University
If a cave is to be accessible to human exploration it
must have an entrance. Entrances, however, are rarer
than might be expected. For the most part, the pro¬
cesses that form caves do not demand that they have
entrances. If a cave entrance was created as part of the
original cave-forming process, the entrance can easily
be destroyed by later collapse and slumping. Most
entrances are formed late in the history of caves by
random processes of erosion and collapse. Some caves
may have no entrances at all (White, 1988).
LOCATIONS OF ENTRANCES
The search for caves by cave explorers is a search for
entrances. Some entrances, of course, are obvious—the
storybook picture of a gaping hole in the hillside. These
caves have generally been known for centuries. Other
entrances are small, obscure, and often go unrecognized
until the systematic combing of an area is undertaken.
Still other entrances have been opened by human activi¬
ties. They are found in road cuts, rock quarries, and by
deliberate excavation. The search for cave entrances has
become more intense as the more obvious openings have
been discovered and the caves explored, surveyed, and
described. Much exploration is now focused on finding
concealed entrances. These are places where cave pas¬
sages intersect the Earth's surface but where collapses,
slumping, and other occlusions have erased all traces on
the land surface. At such locations, a relatively modest
amount of excavation may expose the cave passage.
Encyclopedia of Caves.
© 2012 Elsevier Inc. All rights reserved.
ENTRANCES
281
TYPES OF CAVE ENTRANCES
Cave entrances come in three broad categories:
(1) those that are an intrinsic part of the cave-forming
process so they are the same age as the cave; (2) those
that formed by later processes of truncation and col¬
lapse and thus are younger than the cave; and (3) those
that formed by human activities such as quarrying
and road building—these entrances are the youngest
of all.
1. Entrances formed as an intrinsic part of the cave-forming
process:
Entrances at sinking streams
Entrances at springs
Pit and fissure entrances
2. Entrances formed by later processes of truncation and
collapse:
Cave passages truncated by valley deepening
Entrances formed by sinkhole collapse
3. Entrances created by human activity:
Entrances in quarries and road cuts
Excavated entrances
Whether or not a sinking stream results in a cave
entrance depends to some extent on the magnitude of
the sinking stream. Large streams tend to produce
large entrances. However, large streams can also close
up cave entrances. Many stream entrances are blocked
with stream sediment as well as logs and debris that
have been washed into the entrance by floods. The
stream flows through the permeable debris pile but
explorers cannot follow. Stream sinks are often
regarded as possible sites for digging but there may be
tens of meters of flood debris between the surface
stream and the underlying cave passage. Of the very
large number of sinking streams, only a few are open
cave entrances.
Some springs issue from open cave mouths, providing
an entrance that is easy to enter (Fig. 1). Other springs
are flooded but may be penetrated by divers. Often the
passage flooding occurs only near the spring and a short
distance upstream is the open cave passage. Other
springs are blocked with rubble, permeable to the water
but not to would-be explorers. Some dry entrances to
caves are, in fact, abandoned spring mouths.
Pits and shafts are sometimes an integral part of the
underlying cave system and sometimes are younger
features that formed from infiltrating surface water
more recently than the primary portion of the cave.
Although pits and shafts are dissolved by seepage
water clinging to the walls of the shaft and do not
require streams of water for their development, some
shafts are the sink points for small surface streams.
Some shafts are open to the surface and some are not.
FIGURE 1 Overholt Blowing Cave, Pocahontas County, West
Virginia. A cave entrance at a spring mouth. Photo by the author.
Open shafts can provide a natural entrance to the
underlying cave system. However, because the devel¬
opment of the shaft is by independent water sources,
the shaft may be connected to the underlying cave
only by a small and sometimes impenetrable drain.
Other drains may be blocked by loose rock and other
debris that have fallen down the shaft.
Caves tend to form as near-horizontal passages
directed toward major surface valleys. As the valleys
deepen and widen, the caves are gradually shortened
as valley walls retreat. With time, tributary valleys
deepen and may cut directly through underlying
cave passages. Rubble associated with the collapse
of the tributary valley floor will cut the cave into
two segments and may completely block both ends.
Continued downcutting of the tributary valley can
remove some or all of the rubble so that one or both of
the cave passage segments become exposed on the val¬
ley wall.
Once cave passages are formed, they remain in
place beneath the landscape. The landscape does not
remain in place. Continuous erosion and valley deep¬
ening lowers the land surface so that the amount of
rock above the cave passage gradually decreases. The
roof of the cave loses the ability to support its own
weight. Descending water from the land surface can
enlarge joints and further weaken the roof until the
cave ceiling collapses. The collapse also truncates the
underlying passage. One or both ends may remain
open on the sides of the sinkhole (Fig. 2).
Cave entrances are found at many size scales from
the very large (Fig. 3) to moderate human-size to the
very small (Fig. 4). There is no relationship between
the size of the entrance and the size of the cave.
For this reason it is necessary to check out every
opening with a size sufficient to admit a human being
(Fig. 5).
ENCYCLOPEDIA OF CAVES
282
FIGURE 2 Tytoona Cave, central Pennsylvania. The entrance is
formed by the collapse of the roof of a large passage carrying an
underground stream. Photo by the author.
FIGURE 3 A large entrance: Fortress of Ponor, Apuseni
Mountains, Romania. Note human figures in lower left of the photo¬
graph. Photo by the author.
Many cave passages are blocked, and thus do not
provide entrances, because of the soil slumping and
collapse that is a normal part of the weathering and
retreat of valley walls. Such blockages are often quite
shallow. Human activities that scrape away the weath¬
ered zone are, therefore, quite likely to produce
entrances and expose the underlying cave passages. It
is for this reason that so many cave entrances occur in
quarries, road cuts, and other excavations. Figure 4
FIGURE 4 A small entrance: Roadside Pit, Pocahontas County,
West Virginia. A small crevice in a road ditch opens into the top of a
20-m pit, which in turn opens into a large cave. Photo by the author.
FIGURE 5 Solution conduit in a quarry wall. Every opening must
be checked. Photo by the author.
shows an entrance in the ditch of a country road that
was opened by building the road. This innocuous
opening is the top of a 20-meter pit that opens into a
cave of substantial size.
STATISTICS OF CAVE ENTRANCES
Most caves have only one entrance but some have
two entrances and a few have more than two entrances.
Because entrance formation is primarily a stochastic
process, one might expect a statistical relationship
between the number of caves with a given number of
entrances and the number of entrances and, indeed,
this is the case (Fig. 6). For three populations of caves
in Pennsylvania, West Virginia, and Alabama, the num¬
ber of caves falls off systematically with the number of
entrances. This leads to very interesting speculation.
Because the number of caves with one entrance is
ENCYCLOPEDIA OF CAVES
ENTRANCES
283
FIGURE 6 Number of caves as a function of the number of
entrances. Data for Alabama are based on 1973 data of the Alabama
Cave Survey. Data for Pennsylvania and West Virginia are from Curl
(1958). The curves are extrapolated to predict the number of caves
with zero entrances.
much larger than the number of caves with more than
one entrance, the curves can be extrapolated to predict
the number of caves that have zero entrances. There is
predicted to exist a very large number of caves with no
human access from the surface. A few of the previously
entranceless caves have been exposed by human exca¬
vations. Consideration of the large distances between
stream sinks and their resurgent springs suggest that
many caves must exist that do not have entrances.
Because the most important mechanism for the for¬
mation of entrances is random truncation and collapse
of cave passages, larger caves are more likely to have
entrances than smaller caves. This hypothesis is sup¬
ported by the distribution of caves by length (Fig. 7).
Caves with larger numbers of entrances tend to be the
larger caves. The hypothetical population of entrance¬
less caves is likely to consist mostly of small caves.
Cave length (m)
FIGURE 7 Distribution of caves in West Virginia as a function of
number of entrances and cave length. The number of all caves
decreases as their length increases (that is, there are many more small
caves than there are large caves, regardless of number of entrances).
Overall, caves with multiple entrances tend to be longer. Caves with
zero entrances are predicted to be, on the average, smaller but also
with fewer and fewer caves of greater length. After Curl (1958).
transitional. Ecologists call such transitions ecotones.
Outside is the surface environment—a sun-drenched
hillside, a forest, a deep sheltered valley, or a desert.
Inside is the cave environment—complete darkness,
high humidity, and constant temperature. The bound¬
ary between inside and outside is often taken by cave
surveyors as the drip line, that is, the line marking the
last spot where rainfall reaches the ground. In the
outer transitional zone, cool, moist cave air creates a
microclimate that is cooler in the summer and warmer
in the winter than the surrounding terrain. Ferns,
mosses, and flowering plants, some of them rare, grow
in these sheltered locations. The inner transitional zone
is a twilight zone where enough daylight penetrates to
permit the growth of some plants. The width of the
transitional zone depends on the local setting. It may
vary from a meter or less for a crawlway entrance on a
cliff face to hundreds of meters for a large cave
entrance at the head of a narrow valley.
The cave entrance zone is, in effect, a continuous
sequence of microclimates, each segment of which can
provide habitat for organisms. Organisms can migrate
from the surface environment through the outer and
inner transitional zones into the deep cave environment.
ENTRANCES AS_HABITAT
Although entrances might appear to be a sharp
boundary between the outside surface environment
and the inside cave environment, they are, in fact.
ENTRANCES AS PALEONTOLOGICAL
AND ARCHAEOLOGICAL SITES
Cave entrance areas provide shelter for animals
and provided shelter for early humans, who made
ENCYCLOPEDIA OF CAVES
284
E
their camps, either migratory or semipermanent, in
appropriately shaped entrances. The best sites were
entrances on south-facing slopes that were large
enough for easy movement. Early humans typically
left their debris—remnants of meals, scraps of cloth¬
ing and tools, campfires, and sometimes burials.
Predatory animals often dragged their prey into cave
entrances. The bones remained behind.
Cave entrances are not stable. Because of freeze/
thaw cycles and other hillside erosion, there is a con¬
tinuous, if sparse, rain of soil and rock fragments into
the sheltered area of the cave entrance. This pile of
weathered material is known as entrance talus. Because
of movement of material downslope, the entrance not
only migrates deeper into the hill but also migrates
upward as rocks break away above to contribute to the
talus pile below. Entrance talus piles can vary from
very small to depths approaching 100 m. The continu¬
ous accumulation of entrance talus buries animal
remains and human debris, which are entombed in
rough stratigraphic sequence as the talus pile enlarges.
The sediments in cave entrance areas are, therefore,
often excellent sites for paleontological and archaeo¬
logical investigations.
See Also the Following Article
Entranceless Caves, Discovery of
Bibliography
Curl, R. L. (1958). Statistical theory of cave entrance evolution.
National Speleological Society Bulletin, 20, 9—22.
White, W. B. (1988). Geomorphology and hydrology of karst terrains.
New York: Oxford University Press.
EPIKARST
Michel Bakalowicz
Hydrosciences, Universite Montpellier 2, France
DEFINITION
The word epikarst (in French: epikarst) gradually came
into use during the 1990s following the definition by
Mangin (1973) of the epikarstic aquifer (in French: aqui-
fere epikarstique). According to Mangin, the epikarstic
aquifer is the perched saturated zone within the
superficial part of the karst that stores a part of the
infiltrated water. The term epikarst is a generalization
of the concept of the epikarstic aquifer. It is the shal¬
low, superficial part of karst areas, in which stress
release, climate, tree roots, and karst processes fracture
and enlarge rock joints and cracks, creating a more
permeable and porous zone over the massive carbon¬
ate rock in which only few open vertical joints and fine
cracks occur (Fig. 1). The epikarst overlies the infiltration
zone itself, which is intersected by occasional enlarged
vertical fractures and karst conduits, so that the base of
the epikarst acts as an aquitard and may contain a local
perched water table, the so-called epikarstic aquifer.
Epikarst generally corresponds to the karren (or lapiaz )
zone of geomorphologists and speleologists (White,
1988). According to Williams (1983), dolines should be
considered as part of it.
Instead of epikarst, Williams (1983) preferred subcu¬
taneous karst, which, according to him, was an expres¬
sion translated from the French karst sous-cutane and
proposed by Birot to describe the shallow part of karst
morphology in tropical regions. However, Birot did
not use that expression, but instead used either superfi¬
cial, subsuperficial, or subepidermic. Prior to this, Ciry
employed the term karst cutane (cutaneous karst), refer¬
ring specifically to the shallow caves and karst features
of Burgundy, supposedly developed during glacial
periods, when the permafrost limited karst processes
during summer melting. It is now known that no karst
processes occurred during glacial periods and that the
karst cutane in Burgundy is what remains of ancient
karst phases after the reduction of thickness through
surface erosion.
The expression subcutaneous karst is no longer used,
and epikarstic zone, epikarstic aquifer, and epikarst are
now widely accepted.
EPIKARST, A NECESSARY CON CEPT
In Ecology
The concept of epikarst was first proposed by
groundwater ecologists (Rouch, 1968), who considered
that perched saturated zones must exist within the
temporary percolation zone, that is, the infiltration
zone. They observed aquatic microfauna, mainly
Copepoda, in water dripping from stalactites, in shal¬
low caves. According to Rouch, the observed diversity
and abundance of the fauna imply the existence of a
local and permanent perched phreatic zone, a few
meters below ground surface. In nonkarstic rocks, ecol¬
ogists had already described such a zone, rich in
groundwater fauna, which was named by Mestrov the
hypotelminorheic zone. The concept was applied to most
rock types and was considered to refer to the "skin" of
the subsurface.
Later, Juberthie et al. (1980) extended the concept
to terrestrial underground fauna, by defining the
Encyclopedia of Caves.
© 2012 Elsevier Inc. All rights reserved.
EPIKARST
285
FIGURE I The epikarst is represented on its right
part without a significant soil cover, showing a karren
at its surface. The local saturated zone is shown.
Arrows indicate (1) the percolation seepage, slow and
dispersed in the fine cracks; (2) the rapid and concen¬
trated infiltration flow through enlarged fractures and
vertical conduits; and (3) the overflow of the local satu¬
rated zone into the doline—shaft system. Source: From
Mangin (1975), modified.
underground shalloiv medium (in French milieu souterrain
superficiel or MSS). The shallow zone of the karst is
expressed as the "reservoir" of life, the place where
exchanges occur between surface and subsurface, where
surface animals discover and colonize the underground
environment. A general paradigm (Rouch, 1986)
explains the existence of fauna adapted to living in the
karst and the underground environment by a permanent
and active colonization from the ground surface of the
epikarst, that essential medium for life, which is distinct
from the cavernous environment.
In Hydrogeology
Drogue, in his unpublished doctorate thesis, men¬
tions for the first time the hydrogeological role played
by the epikarst and named it the superficial karst. He
observed a perched, discontinuous, and temporary
water table in the karstified limestones of Languedoc
(southern France), drained by a fracture network to the
saturated zone.
To explain the observed homogenization of isotopic
characteristics in three karst aquifers of the French
Pyrenees, Bakalowicz et al. (1974) invoked the exis¬
tence of shallow storage in karst aquifers and sug¬
gested its generalization to all karst aquifers. The role
of the epikarst is also shown in the process of evapo-
transpiration, that is, the consumption of groundwater
by plants. Numerous karst scientists, mainly in France,
were to show though hydrodynamic and hydrogeo¬
chemical studies that the epikarst is necessary to
explain the functioning of the karst aquifer.
Moreover, in one of the earliest attempts to model a
karst aquifer by a grid approach, Kiraly (1975) showed
that the successful simulation of karst spring hydro¬
graphs is dependent on the introduction of a shallow
layer of elevated porosity and permeability representing
the epikarst.
In Karstology
The epikarst represents the vertical extension of the
soil, and as such can act as a reservoir for the accumu¬
lation of organic matter. The decomposition of organic
matter within this layer produces carbon dioxide, CO 2 ,
which is the main agent of carbonate rock solution,
and of karst processes, when dissolved in groundwa¬
ter. The epikarst is the key site for carbonate rock solu¬
tion. Karst depressions (dolines) are initiated by
drainage of epikarst storage via vertical conduits.
Karren areas develop in poorly drained areas of the
epikarst.
EPIKARST, A NOT YET FULLY AGREED
CONCEPT
Some authors (see Klimchouk, 2004) consider that
dolines and more usually the whole shallow part of karst
belong to epikarst. Bakalowicz (2004) and others (see the
final discussion in Jones et al., 2004) restricts the epikarst
to the parts where shallow water storage may occur, con¬
sidering that epikarst should be defined not only by the
surface morphology, that is, karren fields and their possi¬
ble soil or sediment cover, but overall by its hydrological
ENCYCLOPEDIA OF CAVES
286
E
functioning and the typical processes occurring in it. The
main processes are related to water stored in the upper¬
most part of karst, because of a vadose seepage at its
base: infiltration is delayed, soil CO 2 is directly used for
carbonate solution, organics are dissolved and weathered,
and dissolved salts are concentrated because of water con¬
sumption by evapotranspiration.
HOW DOES THE EPIKARST WORK?
The first approaches to studying the epikarst were
indirect, in that its behavior was deduced from that of
the whole karst aquifer. It soon became evident that a
more direct approach to investigating the shallow zone
of the karst aquifer was required. Several field test
sites were designed for analyzing the epikarst. The use
of tracer tests and artificial rainfall complemented the
classical hydrodynamic, chemical, and isotopic inves¬
tigations (Bakalowicz, 1995). According to the first
interpretations by Mangin and by Bakalowicz, these
investigations showed the existence of various infiltra¬
tion modes through the shallow part of karst. The rain¬
water recharges the karst aquifer and emerges at its
springs by the following pathways:
1. A part of the water infiltrates directly and quickly
through wide fractures and vertical conduits, from
dispersed infiltration at the karst surface, or from
point infiltration through sinkholes.
2. The other part is stored in the epikarst where it
contributes to different processes: a part is consumed
by plants (evapotranspiration), another part percolates
slowly through the fine cracks and rock porosity (slow
infiltration), and, during the heavy rains, the last part
is flushed away into the vertical conduits of the
infiltration zone (the doline—shaft system; Fig. 2), from
epikarst overflow (delayed infiltration).
Point and direct fast infiltration waters do not flow
through the epikarst. Dolines and shafts allow direct
vadose flow to rapidly reach the conduit system, act¬
ing as holes breaking the epikarst, and simultaneously
working as overflow drains of epikarst when it is filled
up with water. Figure 2 describes how infiltration
water is distributed in karst aquifers according to the
different types of underground flows.
PLACE OF EPIKARST IN KARST
EVOLUTION AND MORPHOLOGY
The epikarst is completely involved in CO 2 produc¬
tion, storage and dispersion in karst aquifers. The epi¬
karst works as a CO 2 reservoir, gently recharging the
infiltration zone by means of the slow infiltration. As a
Recharge from
another system
Sp vl/
Runoff on non
karstic terrains
reaching
swallow holes
Recharge
^ ^ ^ ^ ▼
(effective rainfall)
(D
C
0
N Concentrated
c infiltration
0 Doline-shaft
system
Direct -^T Epikarst
infiltration Epikarst
overflow Storage delaying
infiltration
1
/Fast
infiltration
Slow
infiltration
Underground
runoff
Conduit
system
t
Annex-to-drain
systems (ADS)
storage Fissured blocks
matrix
Overflow m Leakage towards
springs Spring(s) another aquifer
outlet of the system
FIGURE 2 Functional scheme of a karstic system showing the dis¬
tribution of infiltration water according to the different types of
underground flows. The functions are written in italic; the "organs"
are written in bold.
matter of fact, this slow infiltration induces mixing of
air and water during the slow percolation in fine
cracks and pores. This process drives the CC^-rich air
present in soil down into the system and disperses it
through the whole infiltration zone. In that way, the
fast infiltration water, from direct dispersed and point
recharge, may dissolve CO 2 and consequently carbon¬
ate rock in the depth of the karst, at the top of the
phreatic zone. Consequently, storage in the epikarst is
an essential mechanism for karst development, at the
surface as well as at depth.
The functioning of epikarst determines the spatial
distribution of carbonate rock solution, not only at the
ground surface, but also between the surface and
the karst at depth. Vertical conduits are enlarged near
the surface as closed depressions, then constituting the
doline—shaft system. They allow for the introduction
at depth of fast-infiltrated aggressive waters, which
dissolve the rock at various depths, in such a way that
fractures may be enlarged into conduits, which is a
prerequisite for the development of caves.
At the top of the system, water stored in the epi¬
karst gently dissolves the rock in the few meters below
the soil, deepening fractures and cracks. The enlarged
ENCYCLOPEDIA OF CAVES
EPIKARST
287
fractures can stock detritic sediments, residue of the
dissolution of carbonate rocks or from displacement by
rain, wind, and surface runoff—this is the karren field,
which may or may not be covered with soil and plants.
When climate and plant cover conditions are stable for
a long time, karren, and then epikarst, may develop
into spectacular landforms, such as in the Stone Forest
in Yunnan, China, or the tsingy in Madagascar.
Recent developments in geophysics have unveiled the
existence of the epikarst at depth. Two methods are par¬
ticularly effective: ground penetrating radar (GPR) which
shows the vertical structure of epikarst (Al-Fares et al.,
2002), and magnetic resonance sounding (MRS), which
demonstrates the presence of water stored in the epikarst.
GPR has shown the existence of an extensive fractured
zone within the first 10 m below the ground surface of a
Mediterranean karst. The zone approximately follows the
ground surface, independently of strata dipping.
Occasional major vertical fractures are the origin of verti¬
cal conduits, which are in turn the origin of the conduit
network which characterizes the karst aquifer.
THE EPIKARST, THE SKIN
OF THE KARST
Finally, epikarst is an essential interface between the
biosphere and the karst itself. Karst develops and
evolves as a result of the epikarst, which distributes the
infiltrated water and the rock solvent (CO 2 + water) in
such a way that a characteristic landscape is created at
the surface (closed depressions, karren) as well as at
depth (the conduit network and caves). Therefore epi¬
karst can be compared to the skin of a living being.
In the same way, the epikarst is a very fragile, sensi¬
tive medium. It may easily be eroded. During glacial
periods of the Quaternary, the freeze/thaw alternation
and the ice- and snow-cover movements totally
destroyed the epikarst of cold regions of Europe and
North America. Most of the slope debris that has accu¬
mulated at the foot of limestone cliffs originated from
such a process. In Mediterranean areas, heavy rains
and flash floods act in the same way, but the erosion is
limited to steep slopes, and the accumulation occurs in
dejection cones. Karst of the high mountains of
Turkey, Lebanon, and Syria are often coated with a
thick cover of carbonate debris which were not yet
washed out. This debris cover works as a continuous
buffer, filtering seasonal variations in infiltration flow,
and is what may regulate the spring discharge.
Humans are also an efficient erosive agent. We
deforest and cultivate the rich but generally thin
soils. Some well-known examples in Ireland and
England show preserved soil and epikarst under dol¬
men in areas otherwise devoid of such features. As
observed in Mediterranean karst areas and in southern
China, even when soils were once thick, they have pro¬
gressively disappeared, revealing karren with rounded
shapes, typical of formations more normally found
under soil or a sediment cover. These examples show
how much the epikarst is fragile and may be easily
eroded. Moreover, during the construction of high¬
ways or ski stations the epikarst is systematically
destroyed, either because from a geotechnical point of
view it provides too fragile of a foundation for build¬
ings, or because it provides a useful source of rubble.
Once the epikarst is eroded, a uniform rock surface,
the pavement, may be observed, interspersed with
only a few cracks and occasional large vertical con¬
duits. To recreate the necessary conditions for karst
evolution, soil and a plant cover must develop, but the
process requires several thousand years in a temperate,
humid climate.
See Also the Following Article
Epikarst Communities
Bibliography
Al-Fares, W., Bakalowicz, M., Guerin, R., & Dukhan, M. (2002).
Analysis of the karst aquifer structure by means of a Ground
Penetrating Radar (GPR). Example of the Lamalou area (Herault,
France). Journal of Applied Geophysics, 51, 97—106.
Bakalowicz, M. (1995). La zone d'infiltration des aquiferes karstiques.
Methodes d'etude. Structure et fonctionnement. Hydroge'ologie, 4,
3—21 (in French).
Bakalowicz, M. (2004). The epikarst, the skin of karst. In W. K. Jones,
D. C. Culver, & J. S. Herman (Eds.), Epikarst (9, pp. 16—22).
Leesburg, VA: Karst Water Institute. Special Publication.
Bakalowicz, M., Blavoux, B., & Mangin, A. (1974). Apports du traf-
age isotopique naturel a la connaissance du fonctionnement d'un
systeme karstique. Teneurs en oxygene 18 de trois systemes des
Pyrenees, France. Journal of Hydrology, 23, 141—158 (in French).
Jones, W.K., Culver, D. & Herman, J.S., ed. (2004). Epikarst. Special
Publication 9, Karst Water Institute, Leesburg, VA.
Juberthie, C., Delay, B., & Bouillon, M. (1980). Extension du milieu
souterrain en zone non calcaire: Description d'un nouveau milieu
et de son peuplement par les Coleopteres troglobies. Me'moires de
Biospe'ologie, 7, 19—52 (in French).
Kiraly, L. (1975). Rapport sur l'etat actuel des connaissances dans
le domaine des caracteres physiques des roches karstiques.
In A. Burger & L. Dubertret (Eds.), Hydrogeology of karstic terrains
(No. 3, pp. 53—67). Paris, France: International Union of
Geological Sciences. Series B.
Klimchouk, A. (2004). Towards defining, delimiting and classifying
epikarst: Its origin, processes and variants of geomorphic evolution.
In W. K. Jones, D. C. Culver & J. S. Herman (Eds.), Epikarst (9,
pp. 23—35). Leesburg, VA: Karst Water Institute. Special Publication.
Mangin, A. (1973). Sur la dynamique des transferts en aquifere kars¬
tique. Proceedings of the 6tli International Congress of Speleology,
Olomouc, III, Canada, 157—162 (in French).
Rouch, R. (1968). Contribution a la connaissance des Harpacticides
hypoges (Crustaces, Copepodes). Annales de Speleologie, 23(1),
5—167 (in French).
ENCYCLOPEDIA OF CAVES
288
E
Rouch, R. (1986). Sur l'ecologie des eaux souterraines dans le karst.
Stygologia, 2(4), 352—398 (in French).
White, W. B. (1988). Geomorphology and hydrology of karst terrains.
New York: Oxford University Press.
Williams, P. W. (1983). The role of subcutaneous zone in karst
hydrology. Journal of Hydrology, 61(1), 45—67.
EPIKARST COMMUNITIES
David. C. Culver, Anton Brancelj,' and Tanja Pipan
American University, U.S.A-f National Institute of Biology,
Slovenia; Karst Research Institute at ZRC SAZL7, Slovenia
INTRODUCTION
The epikarst, or subcutaneous zone, is the uppermost
part of karstified rock, a perched aquifer, and an ecotone
between surface and subterranean environments. It has
been variously defined but in general it is the boundary
between soil and rock in karst, honeycombed with small
fractures and solution pockets. Its vertical extent varies
from nearly zero to a few meters. It is a zone where
environmental parameters (e.g., temperature) can change
significantly over short distances and where there are
seasonal and even daily oscillations. The cavities and
cracks in the rock in the epikarst zone may or may not
be well integrated in the horizontal direction. Sometimes
water percolating through the epikarst moves laterally
at the base of it for substantial distances before finding a
pathway deeper into the vadose zone. The epikarst con¬
sists of a series of small cavities and crevices, some of
which are water-filled, some of which are filled with
organic material, humus, and insoluble material, and
some of which are air-filled. Whereas hydrogeologists
often stress the water storage capacity of the epikarst
zone, biologists usually stress the vertical movement of
water in the zone, and the term percolating water is often
used in this connection. Epikarst is one of the major
aquatic shallow subterranean habitats.
SAMPLING TECHNIQUES
The habitat has rarely, if ever, been sampled
directly. Instead, biologists have had to rely on indirect
samples (Pipan, 2005). Direct methods for sampling
fauna in the epikarst are still being developed. In the¬
ory, drilling vertically into the epikarst zone and col¬
lecting water from voids there is possible. In
principle, a metal or plastic tube, closed at the bottom
and with a series of holes some centimeters above the
sealed end that is left in position for some time
should act as a pitfall trap for fauna there—but that's
in theory! In practice, there are two possibilities for
sampling the fauna. The most direct is to sample per¬
colating water from the trickles, especially in perma¬
nent trickles (Fig. 1). By means of a funnel, water
from a trickle is directed into a plastic container with
one or several holes on the sides. The hole is covered
with a mesh (mesh size of 60—100 pm; Fig. 1).
Containers need to be emptied periodically, perhaps
once a month, to minimize losses due to predation
and other factors.
The second method is to collect and filter water
from small pools on calcareous slopes or on the bot¬
tom of galleries, which are filled with water from
the trickles. The volume of the pools can vary from
a few milliliters to a liter or more. Water from the
pools can be collected by means of a pump (Fig. 2),
which is very efficient in small rimstone pools or in
the deep and narrow cracks on stalagmites that are
filled with water. During the sampling, vigorous
agitation of the water is recommended to collect
particles from the bottom of the pools where most
animals are attached. While sampling drip pools is
much easier than sampling drips, the sampling
from pools is a biased sample of drip water, and
typically has fewer stygobionts, species specialized
for subterranean life (Pipan et al. r 2010). They are
sink populations in the ecological terminology of
source—sink populations. Nonetheless, they are
important collection sites for the epikarst fauna,
especially when the first survey of fauna is made in
a particular cave, and can yield a very diverse fauna
(e.g., Brancelj, 2002).
ENVIRONMENTAL CONDITIONS
Epikarst water is highly heterogeneous in fluxes.
Measurement of discharge of 35 drips over one year in
six Slovenian caves (Pipan, 2005) revealed that fluxes
are extremely dependent on precipitation. They range
from less than 0.5 mL min 1 to nearly 1400 mL min
Measurements in the Slovenian cave Velika Pasjica
showed that discharge can increase from 0.2 mL min -1
to more than 870 mL min -1 in less than 6 hours and
drop back to minimal flow again in 24 to 48 hours. The
relationship of discharge in a cave to precipitation
depends on saturation of soil, cracks, and voids with
water. If they are saturated with water, reaction is
quick (within an hour). After a long dry period, drips
react with long delays or even do not react at all, as
rainwater is retained in soil, cracks, and voids and per¬
colates downward very slowly. In general, drip water
shows the same seasonal pattern as surface precipita¬
tion, but with a delayed response.
Encyclopedia of Caves.
© 2012 Elsevier Inc. All rights reserved.
EPIKARST COMMUNITIES
289
4
mesh
water
FIGURE 1 Device for collecting fauna from percolating water. By
means of a funnel, water is collected in a container with mesh-cov-
ered sides. The lower rim of the holes in the side of the container is
positioned some centimeters above the bottom, allowing a small pool
of water to form in the bottom of the container.
Temperature of water in the epikarst is relatively
constant with some oscillations related to seasons and
weather events. In principle, the water temperature in
the epikarst equals the mean air temperature of the
location, however it rarely, if ever, reaches this equilib¬
rium. In her six study caves in Slovenia, Pipan found
that temperature in dripping water varied from 0.4°C
to 13.7°C in a region where the mean annual tempera¬
ture was approximately 9°C.
Chemically, epikarst water is highly enriched in cal¬
cium ions with a concentration 10 to 50 times that of
rainwater, indicating that it is actively dissolving cal¬
cium carbonate. Measurements of calcium concentra¬
tions in percolation water from six caves in Slovenia
indicate high values. On average, the concentration
of Ca was about 38 mg Ca L 1 (ranging from 10 to
100 mg Ca IT 1 ). Conductivity was also high—the
average value was 355 p,S cm" 1 (ranging from 170 to
730 |iS cnT 1 ), the result of the length of time the water
has been in contact with bedrock, which may be weeks
or even months. Measured values of oxygen concentra¬
tion are usually between 60 and 100% of saturation,
but this may be misleading since these measurements
were based on epikarst water exposed to the air, which
may replenish the oxygen.
Epikarst water contains varying amounts of coarse
and fine particulate organic matter (POC) and dissolved
organic carbon (DOC) as a result of feeding and meta¬
bolic activities in the soil and the epikarst itself. Epikarst
is a source of organic input as well as the location of the
exchange of surface and subsurface fauna. Input of
DOC and POC as a source of nutrients for bacteria and
invertebrates may arrive in the karst basin via two dif¬
ferent pathways: openings (sinks and pits) or infiltration
through soils. Water percolating through soils into the
epikarst carries both DOC and POC, but POC is mostly
filtered by soils. Once transported into the karst under¬
ground, POC and DOC are used or processed to differ¬
ent forms. An important element for DOC and POC
transformation are microbial films on rocks (epilithion).
The DOC concentration in the epikarst drips in the
Postojna-Planina Cave System (Slovenia) and Organ
Cave (West Virginia, U.S.A.) was around lmgCL 1
(Simon et ah, 2007). Organic carbon concentrations in
sinking streams, cave streams, and the resurgence of the
Postojna-Planina Cave System were higher, but in Organ
Cave organic carbon concentrations of epikarst, cave
stream, and resurgence waters were similar. In other
caves, organic carbon concentrations of epikarst water
can be much higher than 1 mg CL -1 . This highlights the
need for careful measurement of organic carbon in dif¬
ferent systems. Compared to cave stream water, the
organic carbon in epikarst is low in aromaticity and
humification, indicating that it was more readily assimi¬
lated into the biotic community (Simon et al., 2010).
ENCYCLOPEDIA OF CAVES
290
E
FIGURE 2 Device for collecting fauna from small amounts of percolating water in pools and cracks.
TAXONOMIC COMPOSITION
A wide variety of invertebrate species are known
primarily or exclusively from epikarst. In the compen¬
dium of subsurface aquatic species Stygofauna mundi
(Botosaneanu, 1986), over 10 species of oligochaetes,
280 species of crustaceans, 2 species of beetles, and 30
species of flatworms are reported from percolating
waters of caves. Eight genera of crustaceans have 10 or
more species that are found in the epikarst (Table 1).
In North America, the amphipod Stygobrovius and the
isopod Caecidotea are common epikarst species,
whereas in Europe the isopod Proasellns, the syncarid
Iberobathynella, the cyclopoid copepods Speocyclops and
Diacyclops, and the harpacticoid copepods Bryocamptus,
Elaphoidella, and Parastenocaris predominate.
Intensive sampling of epikarst since 2000 in different
countries all over the world (Brazil, China, Romania,
Slovenia, Spain, Thailand, U.S.A.) has revealed rich
fauna there. A rather long list of species from epikarst
exists, with Copepoda (Cyclopoida and Harpacticoida),
Ostracoda, and Syncarida (Bathynellaceace) as the most
common representatives, followed by Amphipoda and
Oligochaeta. Other groups, including Gastropoda, are
TABLE 1 Genera of Crustaceans with Ten or More
Epikarst Species
Order
Genus
Total No. of
Subterranean
Species
No. of
Epikarst
Species
Cyclopoida
Speocyclops
41
28
Diacyclops
34
10
Harpacticoida
Elaphoidella
91
34
Parastenocaris
167
15
Isopoda
Caecidotea
57
13
Proasellus
115
29
Amphipoda
Stygobromus
100
45
Syncarida
Iberobathynella
18
10
(Data from various sources).
less common in samples, although their empty shells
are rather frequent in the pools filled by dripping water.
Only eight genera have been found in the epikarst
(Table 2). They include the harpacticoid copepod
from genera Morariopsis (the Dinaric Mountains), and
ENCYCLOPEDIA OF CAVES
EPIKARST COMMUNITIES
291
TABLE 2 Aquatic Genera Likely Found Only in Epikarst
Class
Order
Genus
Number of
Species
Crustacea
Harpacticoida
Moraricipsis
3
Harpacticoida
Paramorariopsis
3
Syncarida
Batubathynella
1
Amphipoda
Niphargobates
2
Insecta
Coleoptera
Trogloguignotus
1
Coleoptera
T roglelmis
1
(Data from various sources)
Par amorariopsis (the Alps and Dinaric Mountains), the
amphipod Niphargobates (the Balkans), the syncarid
crustacean Batubathynella (Malaysia), the beetle Trog-
loguignotus (Venezuela), and the beetle Troglelmis (the
Republic of Congo). Several of these species are shown
in Figure 3. Since the publication of Stygofauna mundi
in 1986, many more aquatic epikarst species have
been described but new exclusively epikarst genera
were recorded only within Copepoda (Harpacticoida),
which is the most common and abundant stygobiotic
group in epikarst.
Many epikarst species are found in genera where
most of the other species are from interstitial habitats
such as shallow alluvial aquifers and the underflow
of streams (z’.e., hyporheic zone). Examples include
the amphipod genus Bogidiella, the isopod genus
Microcharon, and the Syncarida in general. These tax¬
onomic distribution patterns suggest an affinity
between interstitial and epikarst habitats. At the same
time there are very few species which inhabit both
habitats, interstitial and epikarst. In North America,
the exclusively subterranean genus Stygobromus is
found in a wide variety of subterranean habitats,
including cave streams, interstitial habitats, deep
groundwater, and seepage springs. In eastern North
America, 28 of 56 species are found exclusively in the
epikarst, and three others were found in epikarst as
well as in other subterranean habitats.
Because part of the epikarst is air-filled, one would
expect terrestrial species as well. To our knowledge,
no one has sampled these species directly, but Gibert
(1986) and Pipan et al. (2008) did collect considerable
numbers of terrestrial species in nets under ceiling
drips. They found a total of 12 and 18 taxa, respec¬
tively, including one representative of Collembola, one
Diplura, and one Coleoptera, known only from caves.
Pipan et al. (2008) found a snout beetle (weevil),
Troglorhynchus anophtalmus, a very specialized coleop-
teran, feeding on the tree roots that penetrate from soil
into epikarst and even caves. It is interesting that the
most common terrestrial species Gibert found was the
collembolan Arrhopalites secondarius, a globular-shaped
species with morphology typical of deep soil species.
Indeed, it would be surprising if the terrestrial epikarst
fauna did not share affinities with the deep soil fauna.
Little else is known about terrestrial epikarst species
except for tantalizing hints scattered in the literature.
For example, there is a very unusual carabid beetle
(Coleoptera) in its own genus ( Horologion ), found only
once, in 1938 in a small, shallow cave in West Virginia.
In spite of repeated visits to the cave and nearby caves,
it has never been collected again. Given that it was
found in a very shallow cave and that it has not been
seen since, it seems likely that this species and others
collected in similar circumstances are in fact epikarst
species. The terrestrial snail Zospeum, whose shells are
common in many caves of the Dinaric karst, is likely
an epikarst specialist. Although reported from other
parts of the caves, especially from rotten wood near
the entrances, its main habitat is small crevices in the
ceiling and walls of the caves. Other parts of the caves
are probably inhabited by "sink" populations.
ESTIMATING TOTAL SPECIES RICHNESS
OF EPIKARST COPEPODS
Because of the possibility of continuous sampling,
the epikarst fauna can be more thoroughly and accu¬
rately studied than fauna in other subterranean habi¬
tats. The epikarst copepod fauna is a significant part of
the aquatic cave fauna, contributing about 20% of the
species at the regional level. An example is from caves
in Dinaric Karst in Slovenia, where 35 drips in six
caves were continuously sampled for one year. Species
accumulation curves and estimates of total copepod
species richness were calculated. At the scale of indi¬
vidual drips, three to four months' sampling appeared
to be sufficient to collect most of the species. At the
scale of the individual cave, five drips sampled for one
year appeared to be sufficient to collect most of the
species, and at the regional scale, five caves were suffi¬
cient to collect 90% of all epikarst species of Copepoda
(Pipan and Culver, 2007a). In the case of epikarst
Copepoda, the relatively rapid saturation of accumula¬
tion curves and the relative completeness of samples is
the result of the fine scale of heterogeneity and possi¬
bly intensive movement of individuals compared to
other subterranean habitats. The epikarst is a finely
structured habitat at a small scale (meters or tenths of
meters) in which a relatively high number of species
coexist because of habitats' patchiness and different
environmental conditions.
Some copepod species in epikarst have very local
distributions. In a series of caves in southwest Slovenia
and West Virginia a significant difference in fauna
ENCYCLOPEDIA OF CAVES
FIGURE 3 Examples of species from five genera endemic to epikarst (see Table 2). (A) Trogloguignotus concii, (B) Troglelmis leleupi,
(C) Morariopsis dumonti, (D) Niphargobates orophobata, and (E) Allobathynella malaya (scale bar: 1 mm for (A), (B), (D), and (E); 0.1 mm for (C)).
ENCYCLOPEDIA OF CAVES
EPIKARST COMMUNITIES
293
composition of Copepoda over a distance of 100 m
appeared (Pipan and Culver, 2007b). In general, the
pattern of distribution of different taxa within a cave
follows the same pattern as for distribution between
the caves. This suggests a pattern of extremely
restricted lateral flow under normal hydrological con¬
ditions. Under high flow conditions lateral flow may
increase what is often the case during a toxic spill.
Frequent occurrence of copepods in dripping water,
knowledge of their local distribution within a cave, as
well as their small size make them a useful tool in trac¬
ing the potential route of pollutants and potential
water tracers within a small karst system.
MORPHOLOGICAL FEATURES
Compared to obligate cave-dwelling species, epi-
karst species tend to be smaller in size. Groups with
larger sized species that are absent in the epikarst
include fish, salamanders, crayfish, and shrimp. On
the other hand, groups with very small species, usu¬
ally less than 2 mm, predominate, both in terms of
abundance and number of species. Primarily, these are
cyclopoid and harpacticoid copepods. Even within a
genus, epikarst species tend to be smaller. The North
American amphipod genus Stygobromus is exclusively
subterranean, and in the eastern United States occurs
in epikarst, seepage springs, cave streams, and deep
groundwater. Body sizes were smallest for epikarst
species, averaging 5.3 mm for 28 species. In cave
streams, sizes averaged 9.0 mm for 9 species (Culver
et ah, 2010). On the other hand, there were no differ¬
ences in antennal lengths (corrected for body size) in
the different habitats.
Epikarst copepods have some specific adaptations
that make them easily recognizable as epikarst spe¬
cialists. In Cyclopoida specific adaptations are, apart
from reduction of segments on legs PI to P4, the pres¬
ence of very strong and robust spines on exopodites of
the first pair of legs. In addition, they have very strong
spines at the base of each endopodite on the first pair
of legs oriented backward. Those spines prevent them
from being washed away by strong current, as they act
as anchors. In Harpacticoida the most characteristic are
modifications in body shape, which is elongated and
cylindrical. Furcal rami are divergent (up to 45°) and
equipped with robust and stiff terminal setae. There
are also short and strong or (more commonly) robust
setae on the baseoendopodite of leg P5. Spines on exo¬
podites, especially on the first pair of legs, are very
robust and also curved. Short and robust spines and
setae enable specimens to obtain better grip against
water current not to be washed out from the epikarst
(Brancelj, 2009).
BIOGEOGRAPHY OF EPIKARST SPECIES
Local species richness in the epikarst can be quite
high compared to other subterranean environments.
Gibert (1986) lists 12 species as occurring in the perco¬
lating waters of Cormoran Cave in France, among
them four obligate subterranean species, including two
harpacticoid and one cyclopoid crustacean species. In
the 100-m-long Slovenian cave Velika Pasjica, Brancelj
(2002) found an amazingly rich epikarst copepod
fauna: 11 harpacticoids and 1 cyclopoid in drip pools.
Pipan (2005), sampling both drips and pools in
Zupanova Jama in Slovenia, found even more spe¬
cies—14 harpacticoids and 2 cyclopoids. In both cases,
all species but one were stygobionts. In addition, inten¬
sive sampling by Pipan in the epikarst zone in six caves
in Slovenia revealed a total of 37 taxa of Copepoda.
Twelve of them are ubiquitous, the rest are stygobiotic,
and 14 of them are endemic to Slovenia. The high spe¬
cies richness of Copepoda in Slovenian epikarst is in
accord with the high richness of Copepoda also in
other subterranean habitats. At the moment there are
51 cyclopoid and 56 harpacticoid species known from
Slovenia of which 18 cyclopoid and 27 harpacticoid
species are stygobionts and 2 cyclopoid and 13 harpac¬
ticoid species are restricted to epikarst only (i.e., 15% of
all recorded species in Slovenia).
Most epikarst species have been found in caves
throughout the world very recently, although the very
first ones, such as Speocyclops infernus and Morariopsis
scotenophila, were described in 1930 and were collected
from pools filled with drip water. The best studied
karst region is the Karst region of Slovenia where
seven caves have been intensively sampled by Pipan
and Brancelj. The caves are situated in a rectangle of
about 2400 km 2 and the distance between caves varied
from about 1 km to approximately 60 km. Several pat¬
terns characteristic of epikarst species emerged. First,
endemism is high. Among 37 taxa collected in the
seven caves, 14 taxa are endemic to Slovenia. Second,
within the 7 caves studied, distribution of species was
patchy. Only one species, Speocyclops infernus, a cyclo¬
poid copepod, was collected in all 7 caves. In contrast,
one-third of taxa were present only in 2 caves (11 taxa)
or in 1 cave (13 taxa). This indicates that the majority
of stygofauna in epikarst zone are (1) very localized on
a microscale, and (2) very diverse on a macroscale.
This supports the hypothesis that communication
between voids in the horizontal direction is limited
and thus distribution/migration of individuals, includ¬
ing genetic material, is limited.
In spite this high level of heterogeneity and local
endemism, some patterns did emerge. First, and most
remarkably, the frequency of occupation of habitats at
ENCYCLOPEDIA OF CAVES
294
E
one scale (drips within a cave) was a good predictor of
frequency of occurrence in different caves (Culver
et al., 2009). This in turn suggests that the distribution
of epikarst copepods is a very dynamic process, and
that dispersal ability of different copepod species is an
important determinant of distribution pattern. It also
suggests that there are some scale invariant (fractal)
characteristics of the distribution of epikarst copepods.
Second, the contiguous distribution of many species
of epikarst copepods within a cave is on the order of
100 m, suggesting that they can be used a tracers of
the lateral movement of epikarst water (Pipan and
Culver, 2007b).
ECOLOGY OF EPIKARST SPECIES
Pipan et al. (2006) analyzed the relationship between
the epikarst copepod fauna collected from drips in five
Slovenian caves and a variety of physical (drip rate,
temperature, conductivity, ceiling thickness, and surface
precipitation in the preceding month) and chemical (chlo¬
ride, nitrate, sulfate, sodium, potassium, calcium, and
magnesium) parameters. Canonical Correspondence
Analysis (CCA) allows the simultaneous representation of
environmental variables and species' preferences (Fig. 4).
The axes of the resulting two-dimensional graph show the
linear combination of environmental variables that explain
the greatest possible amount of the variance of all the
environmental variables taken together. Species lying close
to one of the lines for the environmental variables is
strongly associated with that variable. For example,
Moraria varica is found in waters with higher concen¬
trations of NO 3 , and Bryocamptus balcanicus is associ¬
ated with drips with higher flow rates (Fig. 4). Each
species niche can be similarly defined by its position
with respect to the canonical axes. Species with nearby
positions have similar niches. Overall differences
among the caves with respect to the environmental
variables were found, and these differences help
explain faunal differences among the caves.
The variables that best explained the variation in the
29 copepod species were thickness of the cave ceiling,
and concentrations of Na + , NO 3 , and K 1 . Nitrate is a
macronutrient and the reason for a correlation with
sodium and potassium is not clear, but it may be
FIGURE 4 Ordination diagram based on species composition and abundance data of copepods in drips in relation to 12 environmental
variables in five caves (Crna Jama (Cr), Dimnice (Di), Pivka Jama (Pi), Skocjanske Jame (Sk), and Zupanova Jama (Zu)). CC1 and CC2 are the
first two canonical axes and together account for 51.9% of the total variance. See Pipan (2005) for further details.
ENCYCLOPEDIA OF CAVES
EVOLUTION OF LINEAGES
295
connected with surface pollution. The negative correla¬
tion with ceiling thickness, resulting in the richest
fauna from the thinnest ceilings, suggests that the ceil¬
ing, the zone of percolation, below the epikarst acts as
a filter for fauna falling out of the epikarst.
The fauna of Copepoda in trickles and pools in "fos¬
sil" galleries (and probably all stygofauna collected
there) is a result of accidental downward drift. The
abundance of copepods collected under drips is posi¬
tively correlated with flow rates, which is to be
expected if the appearance of animals is the result of
accidental downward drift. The extreme heterogeneity
in the numbers of animals collected in drips is consis¬
tent with the hypothesis that occasionally all of the
copepods get washed out from a small solution pocket.
The presence of all life stages in the drip collections is
also consistent with this hypothesis.
Based on the reduced frequency of nauplii and
immature individuals, and the relative rarity of stygo-
bionts in drip pools relative to drips, Pipan et al. (2010)
conclude that drip pools are sink habitats for epikarst
copepods. That is, if the continuing flow of copepods
from drips were cut off, the copepod populations
would mostly disappear. The flux of copepods can be
surprisingly large. In Organ Cave, West Virginia, an
average of one copepod per day was found in dripping
water, and the yearly flux of copepods into pools often
exceeds the number of copepods in the pool. Of
course, very large pools may harbor source popula¬
tions that are self-sustaining. This seems to be the case
for an artificially enlarged drip pool in Skocjanske
jame. Some species may also thrive in pools.
Considering the high number of specimens of
Speocyclops infernus collected in drip pools in Slovenian
caves, it appears that the species either reproduces in
pools or is easily washed out from epikarst. It com¬
monly appears as the only species in some pools,
articularly in pools with clay bottoms, where it fre¬
quently coexists with Niphargus.
See Also the Following Articles
Epikarst
Shallozv Subterranean Habitats
Bibliography
Botosaneanu, L. (Ed.), (1986). Stygofauna mundi Leiden, The
Netherlands: E. J. Brill.
Brancelj, A. (2002). Microdistribution and diversity of Copepoda
(Crustacea) in a small karstic cave in central Slovenia.
Hydrobiologia, 477(1-3), 59-72.
Brancelj, A. (2009). Fauna of unsaturated karstic zone in Central
Slovenia: Two new species of Harpacticoida (Crustacea:
Copepoda), Elaphoidella millennia n.sp., & E. tarmani n.sp., their ecol¬
ogy and morphological adaptations. Hydrobiologia, 621(1), 85—104.
Culver, D. C., Holsinger, J. R., Christman, M. C., Ik Pipan, T. (2010).
Morphological differences among eyeless amphipods in the
genus Stygobromus dwelling in different subterranean habitats.
Journal of Crustacean Biology, 301(1), 68—74.
Culver, D. C., Pipan, T., & Schneider, K. (2009). Vicariance, dispersal,
and scale in the subterranean aquatic fauna of karst regions.
Freshwater Biology, 54(4), 918—929.
Gibert, J. (1986). Ecologie d'un systeme karstique Jurassien.
Hydrogeologie, derive animale, transits de matieres, dynamique
de la population de Niphargus (Crustace, Amphipode). Me'moires
Biospe'ologie, 13(40), 1—380 (in French).
Pipan, T. (2005). Epikarst—A promising habitat. Ljubljana: ZRC
Publishing.
Pipan, T., & Culver, D. C. (2007a). Regional species richness in an
obligate subterranean dwelling fauna—Epikarst copepods. Journal
of Biogeography, 34(5), 854—861.
Pipan, T., & Culver, D. C. (2007b). Copepod distribution as an indica¬
tor of epikarst system connectivity. Hydrogeology Journal, 15(4),
817-822.
Pipan, T., Blejec, A., & Brancelj, A. (2006). Multivariate analysis of
copepod assemblages in epikarstic waters of some Slovenian
caves. Hydrobiologia, 55(1), 213—223.
Pipan, T., Holt, N., & Culver, D. C. (2010). How to protect a diverse,
poorly known, inaccessible fauna: Identification and protection of
source and sink habitats in the epikarst. Aquatic Conservation:
Marine and Freshwater Ecosystems, 20(7), 748— 755.
Pipan, T., Navodnik, V., Janzekovic, F., & Novak, T. (2008). Studies
of the fauna of percolation water of Huda Luknja, a cave in iso¬
lated karst in northeast Slovenia. Acta Carsologica, 37(1), 33—43.
Simon, K. S., Pipan, T., & Culver, D. C. (2007). A conceptual model
of the flow and distribution of organic carbon in caves. Journal of
Cave and Karst Studies, 69, 279—284.
Simon, K. S., Pipan, T., Ohno, T., & Culver, D. C. (2010). Spatial and
temporal patterns in abundance and character of dissolved
organic matter in two karst aquifers. Fundamental and Applied
Limnology, 177, 81—92.
EVOLUTION OF LINEAGES
Eleonora Trajano and Marina Cobolli 1
University of Sdo Paulo, Brazil, ^University of Rome
“La Sapienza", Italy
Subterranean ecosystems are distinguished by the pres¬
ence of species that may considerably differ from their
epigean (surface) relatives in terms of morphology,
physiology, behavior, and ecology. To investigate which
organisms colonize the subterranean realm to establish
more or less self-sustained populations (and why others
do not succeed in this), how some of these lineages dif¬
ferentiate, and what mechanisms are causing this evolu¬
tionary change, are the main challenges and sources of
debate for speleobiologists throughout the world.
COLONIZATION OF SUBTERRANEAN
HABITATS
Subterranean organisms illustrate well the impor¬
tance that preadaptations (there are authors who
Encyclopedia of Caves.
© 2012 Elsevier Inc. All rights reserved.
296
E
prefer "exaptations") may have for the adoption of a
new way of life. ( Preadaptations , as defined here, are
character states conferring performance advantage
in a given selective regime, but which have been
selected in another, independent previous regime.)
Classical preadaptations to the subterranean life
include nocturnal activity, selecting for development
of mechanosensory and/or chemosensory traits, and
opportunistic, generalist feeding, allowing survival
in a food-poor environment that has scattered, vari¬
able, and frequently unpredictable food sources.
Therefore, subterranean communities represent a
subsample of the epigean ones that is strongly biased
toward nocturnal, ecologically generalist taxa, includ¬
ing those currently living in the area as well as those
that had lived there, colonized the hypogean habitat,
and then become extinct in the surface (see discus¬
sion of relicts below).
There are several possible colonization routes to
hypogean habitats. In the case of aquatic organisms,
marine faunas are a source of colonizers for both
marine and freshwater hypogean habitats. The latter
may be reached through invasion of aquifers along
interstitial and other pathways (e.g., anchihaline
habitats, submarine karst springs) by marine organ¬
isms progressively adapting to freshwater condi¬
tions. Subterranean freshwater species may also
evolve directly from marine ancestors by stranding
of founder populations in gradually freshening
groundwaters during marine regressions. Among
hypogean populations derived directly from epigean
freshwater ancestors, stream-dwellers dependent on
lotic conditions (fast-flowing, well-oxygenated waters)
may access the subterranean environment mainly
through stream sinkholes and resurgences. Those
adapted or at least tolerant of lentic conditions (slow-
moving, low-oxygen waters), and especially small¬
sized organisms, may also penetrate deep into the
water table, colonizing phreatic habitats both laterally
and vertically (for instance, through hyporheic zones),
and reaching caves through this route. In the first case,
the boundary between epigean and hypogean habitats
is clearer and, due to the fragmented nature of
streams, disruption of populations is easier to achieve.
Phreatic habitats, on the other hand, may be continu¬
ous throughout wide areas and encompass different
kinds of subterranean habitats where troglomorphic
taxa may be found. Routes of colonization of the sub¬
terranean realm by terrestrial organisms are basically
limited by their size. A vertical colonization of caves
through deep soil and spaces (cracks, crevices) in the
overlying bedrock is probable for small organisms;
large animals require larger contacts between the
subterranean habitat and the surface, and large cave
entrances may be a frequently used route. The impor¬
tance of forest-floor litter fauna as a source of coloni¬
zers for hypogean habitats has long been noted for
many karst areas in the world.
The colonization of subterranean habitats by pre¬
adapted taxa forming troglophilic populations could
occur at any time, but it is predicted that it happens
mainly during the periods most favorable to these
preadapted species in the epigean environment,
when reproductive success and survival rates are at
their highest. Therefore, the generality of the notion
of colonization constrained by unfavorable climatic
conditions and immediately followed by speciation
resulting in troglobites, which is central to the paleo-
climatic model (see below) according to several
modern authors, is highly questionable. Dispersion is
a natural tendency of all organisms, which tend to
occupy all suitable and ecologically accessible habi¬
tats. The gradient between epigean and hypogean
environments is generally restricted to the proximity
of their contacts. Thus, except perhaps for organisms
living deep in a progressively drying soil or sediment
over karst bedrock, epigean animals cannot be
"forced" to enter subterranean habitats during stress¬
ful periods because they would not know in which
direction to move in order to find access to the shel¬
tered hypogean environment. Animals do not present
purpose-oriented behaviors—they just survive where
they are already living, to which they are adapted.
On the other hand, organisms may be washed or
dragged into the subterranean habitat at any time
(by floods, for instance, in a case of passive coloniza¬
tion) and survive or not depending on the presence
of preadaptations or simply by chance.
In conclusion, animals colonize subterranean habitats
through cave entrances in the human sense (openings
large enough for humans to cross) to much smaller con¬
tacts between the epigean and the hypogean realms,
such as the MSS (mesovoid shallow substratum), epi-
karst, deep soil, and so on. Colonization may be hori¬
zontal, as in the case of animals following water
courses until they reach sinkholes or resurgences, to
vertical, for example, animals passing through soil
and/or fractured rocky layers and reaching larger sub¬
terranean spaces as caves. The latter may be accelerated
by climate change, for instance, due to progressive dry¬
ing of the surface: individuals able to survive in deeper
layers would follow changing gradients. However, this
colonization by constraint is restricted to a few situa¬
tions, where gradient changes follow a continuous,
undisrupted homogeneous route. In the many cases,
habitat heterogeneities (such as variations in depth
ENCYCLOPEDIA OF CAVES
EVOLUTION OF LINEAGES
297
along streams resulting in isolated pools during dry
seasons or epochs) would prevent this mode of
colonization.
MODELS OF GENETIC
DIFFERENTIATION AND ORIGIN
OF TROGLOBITES
As for epigean taxa, genetic differentiation is prob¬
ably more frequently achieved by geographic isola¬
tion leading to allopatric speciation of troglobitic
species. In view of the vertical spatial continuity,
which allows permeability between the epigean and
the hypogean compartments of the biosphere and
which is higher for smaller organisms, the most plau¬
sible mechanism of geographic isolation for most sub¬
terranean populations is through local extinction of
related epigean populations. A classical model for
temperate regions, which also applies to the tropical
ones, is based on local extinction due to paleoclimatic
fluctuations, rendering the epigean environment
unsuitable for many taxa. Because the buffered condi¬
tion of the subterranean environment results in the
maintenance of environmental variables within the tol¬
erance range of several species, troglophilic populations
in their hypogean habitats could survive periods of cli¬
matic stress above and could differentiate in the absence
of genetic flow from their (locally extinct) epigean
conspecifics.
The majority of hypogean terrestrial populations
belongs to hydrophilic taxa that are preadapted to
life in the generally damp subterranean environment.
Thus, the colonization of hypogean habitats and estab¬
lishment of troglophilic populations would take place
mainly during humid periods (glacial or interglacial,
depending on the region), and the geographic isolation
in drier periods, when humid vegetation types are
replaced by drier ones and the epigean drainage may
be disrupted. During these drier periods, which are
climatically stressful for hydrophilic and aquatic taxa,
a decrease in food input is expected, causing a decline
in subterranean population sizes. This would result
in bottleneck effects and favor genetic drift leading
to rapid differentiation. Under such conditions, one
can predict strong selection for improved efficiency
in finding food and mates (as indeed is observed in
many troglobites), which, in turn, could allow a new
increase in population sizes.
On the other hand, geological and hydrological bar¬
riers, such as the presence of insoluble, impermeable
rocks interposed between the surface and habitable
subterranean habitats, obliteration of subterranean
conduits by collapse, siltation or chemical deposition,
karst drainage divides, and stream capture, could also
provide isolation events for specific areas and taxa with¬
out extinction of epigean populations.
In addition to the derivation of troglobitic species
from well-established troglophilic populations, a possi¬
ble instance of concomitant colonization and isolation
is a consequence of stream capture. This is a noncycli-
cal, geomorphologic model proposed for the inten¬
sively studied Mexican tetra characin fishes, genus
Astyanax. Stream capture occurs when a surface stream
running on an impervious stratum beneath which
there is soluble rock (usually limestone) meets at an
intersection with a vertical joint in the latter, which is
gradually enlarged by dissolution and erosion to the
point that all of the stream water sinks into the captur¬
ing pit. At this point, the stream reach downstream
of the sinkhole and its fauna would be completely
subterranean, isolated from the reach upstream by
waterfalls. Nevertheless, it has been observed that
surface fish have constant access to the caves from
both above and below; thus the applicability of this
model to the origin of the cave Astyanax species is
questionable.
Models of parapatric and sympatric speciation
have also been proposed and may account for the
origin of some troglobitic species. These models postu¬
late the occurrence of genetic differentiation without
geographic isolation, dependent on intrinsic biological
properties of the taxa. A particular explanation pro¬
posed initially for tropical terrestrial troglobites is the
adaptive-shift model (Howarth, 1993), which is based
on the notion that the hypogean habitat is so stressful
for epigean organisms that those managing to colonize
it are prone to rapid genetic reorganization and differ¬
entiation. The adaptive-shift model is not in accor¬
dance with the observation that most troglobites
belong to preadapted taxa, for which the subterranean
life would not represent such a new, highly stressful
ecological challenge. As a matter of fact, troglophilic
populations, especially in tropical areas, are far more
frequent than predicted by this model, according to
which every subterranean terrestrial population would
tend to become troglobitic with time.
It is important to note that troglophiles, which by
definition are as well adapted to the subterranean as to
the epigean way of life, are not merely an "intermediate"
or "preparatory" step in the evolution of troglobites.
Only when extrinsic (climatic, geological, hydrological)
or intrinsic (genetic) barriers interrupt or greatly
decrease the gene flow between hypogean and epigean
individuals does a troglophilic population have the
potential to originate troglobitic species. Otherwise,
ENCYCLOPEDIA OF CAVES
298
E
troglophilic populations may remain as such forever,
and the high proportion of troglophiles, at least in tropi¬
cal caves, indicates that this is frequently the case.
SINGLE VERSUS MULTIPLE ORIGIN
OF TROGLOBITES
By definition, troglobites cannot survive long or com¬
plete their life cycles in the epigean environment, either
because conditions in the latter became unsuitable for
the species (first step in the climatic model) or, further
on, because they acquired specializations that preclude
the epigean life, even when the original environmental
conditions are restored. Therefore, their dispersion is
basically subterranean, and most troglobitic species
have small geographic distributions, restricted to lim¬
ited areas or even to a single cave or cave system.
Nevertheless, troglobites have been assigned to single
species that present wide distributions and that may
encompass superficially discontinuous karst areas.
In any case, but especially for the latter, the question
arises of whether these species originated after a single
(Fig. 1) or multiple (Fig. 2) independent colonization
event by the same epigean ancestor, followed or not
followed by subterranean dispersion. Because selective
forces and constraints are generally similar throughout
the subterranean environment, different populations
of the same ancestral species isolated independently
in the hypogean habitat tend to evolve in the same
direction, acquiring similar apomorphic traits ( e.g .,
reduced eyes and pigmentation) and frequently result¬
ing in very similar, morphologically indistinguishable
taxa—a classical case of parallel evolution. In many
cases, such cryptic species resulting from independent
speciation events can only be recognized by genetic
methods.
Troglobitic species distributed in superficially discon¬
tinuous karst areas are strong candidates to represent
examples of multiple colonization and independent
speciation. However, keep in mind the possibility of
a continuity through deep karst or that a previously
continuous karst area (for instance, limestone on top of
an anticlinal structure), harboring a single troglobitic
species, was fragmented by erosion and karst denuda¬
tion. This illustrates well the need for integration of
s x
s x
I
UrnUr 5 nUruiL^
EHI
C 3
C 2
C 4
FIGURE 1 Generalized scheme for a
single colonization event of cave system Ci
by an epigean ancestor (S x ), resulting, after
genetic isolation, in a troglobitic species (Si)
restricted to that cave system (as shown in
(A)) or also occurring in other cave systems
(S 2 - 5 ) due to subterranean dispersal of Si
(as shown in (B)). EHI, epigean/hypogean
interface. Arrows indicate gene flow between
populations. Extracted from Holsinger, 2000.
J x -► -► °X -► 5>X
EHI
Ci C 2 C3 C4 C5
u
s 2
u
S3
u
s 4
UuuU
EHI
Ci C 2 C3 C4 C5
FIGURE 2 Generalized scheme for multiple
colonization events by a single, widespread
epigean ancestor (S x ), establishing independent
populations in cave systems Ci_ 4 , each one origi¬
nating a troglobitic species (Si_ 4 ) after genetic
isolation; subsequently, S 4 expanded its range to
cave system C 5 by subterranean dispersal. EHI,
epigean/hypogean interface. Arrows indicate
gene flow between populations. Extracted from
Holsinger, 2000.
ENCYCLOPEDIA OF CAVES
EVOLUTION OF LINEAGES
299
morphological, ecological, genetic, and geological data
for the elucidation of this and many other relevant
questions in biospeleology.
TR OGLOM ORPHIC TRAITS
Variation and Time of Isolation: “Recent”
versus “Ancient” Troglobites
Since Racovitza's time and even before, degree of
troglomorphism, usually reduction of eyes and pig¬
mentation and elongation of appendices, has been
used as a main measure of phylogenetic age, lead¬
ing to the distinction between "ancient" and "recent"
troglobites. Within the paradigm of gradual evolution,
troglobitic species which are totally anophthalmic and
depigmented, with long appendages, especially with¬
out known epigean relatives, have been considered
as ancient troglobites, living in subterranean habitats
much longer than those with partially regressed eyes
and pigmentation and/or showing intrapopulation
variation in such characters. For the latter, the isolation
time would not be enough for fixation of transformed
character states throughout the whole population,
whereas species isolated for a long time and/or subject
to fast differentiation conditions might differ consider¬
ably from epigean-related taxa, to the point that such
relatives are not recognized anymore, either because the
subterranean taxon has diverged too much or because
the surface ones had became extinct. Subterranean taxa
belonging to lineages extinct in the epigean environ¬
ment are referred as phylogenetic relicts ( e.g ., the crusta¬
cean order Spelaeogriphacea, with only three extant
species, all troglobites showing a Gondwanaland distri¬
bution). Those belonging to epigean groups that have
disappeared from a particular area but which still exist
somewhere else are geographic or distributional relicts.
However, a straight correlation between degree of
troglomorphism and phylogenetic age may be ques¬
tioned, as pointed out by Trajano (2007) and exemplified
for cixiid planthoppers by Wessel et al. (2007), who did
not find a correlation between troglomorphy and age
of caves. First, it is based on the assumption of fairly
constant evolutionary (differentiation) rates among char¬
acters in more or less related taxa. The occurrence of
different mosaic of character states in closely related
species, as observed in several Brazilian cavefishes (e.g.,
Ancistrus cryptophthalmus), suggests different mechan¬
isms acting at different rates in each population. Such
mosaics may encompass many characters, morphologi¬
cal, physiological, and/or behavioral, as seen in North
American amblyopsid (Niemiller and Poulson, 2010)
and Thai balitorid fishes (Parzefall and Trajano, 2010),
hampering any attempt to rank such species according
to their degree of specialization to the subterranean life
( = troglomorphy). Moreover, intravariable populations
may be a result of secondary introgression between
fully differentiated troglobites and closely related epi¬
gean, eyed, and pigmented populations, as proposed for
Mexican cavefishes.
Genetic studies, associated with morphological and
physiological data, brought into light the complexity
of these systems. There are two independent develop¬
mental units within the eyes, which may regress
independently in closely related populations, as
observed for Mexican cavefishes (Wilkens, 2010), and
depigmentation may be achieved either morphologi¬
cally, due to progressive loss of melanophores (muta¬
tions in a polygenic system), and / or physiologically,
by loss of ability to synthesize melanin (through at
least two different, apparently monogenic systems;
Felice et al., 2008) —these two modes of pigmentation
loss may coexist in a population, but degenerative
processes affecting monogenic systems are likely to
proceed faster.
In addition to the specific mechanisms of character
transformation, two factors may influence rates of diver¬
gence: population sizes and life-cycle strategies. Small
populations tend to differentiate faster due to phenom¬
ena such as genetic drift. Population sizes are extremely
influenced by ecological factors, and may vary among
closely related species due to differences in habitat.
Progressive specialization to the subterranean life may
also affect population sizes, through an increasing effi¬
ciency in the use of resources allowing population incre¬
ments, which would slow down differentiation rates.
Moreover, K-selected life strategies imply lower differ¬
entiation rates due to delayed ages for first maturation
and low reproductive rates. In conclusion, there is a
complex balance between different genetic, ecological
and biological factors producing the actual divergence
rates, which may differ among related taxa and even
among different characters (Trajano, 2007).
Constructive versus Regressive Traits:
Definitions and Mechanisms
Two kinds of apomorphic traits of troglobites related
to the hypogean life (troglomorphisms) can be distin¬
guished: constructive and regressive traits. Constructive
character states correspond to developed traits in rela¬
tion to those observed in epigean relatives (e.g., elonga¬
tion of antennae in arthropods), for which natural
selection, usually related to feeding and reproductive
efficiency, may be evoked. Classical examples are the
developed mechanosensory and chemosensory organs
of many troglobites. Regressive characters are transfor¬
mation series in which derived states correspond to
ENCYCLOPEDIA OF CAVES
300
E
a decrease until loss (character reversal) of structures,
physiological, and behavioral components selected in
the epigean environment occurs. Regressive traits,
particularly reduced eyes and pigmentation, are the
most conspicuous and distinguishing features of troglo-
bites. These were understandably the first aspects
of subterranean biology to attract the attention of
researchers and are the central focus of most studies on
hypogean organisms.
One of the most debated questions about subterra¬
nean evolution refers to the mechanisms underlying
character regression, and dozens of hypotheses, more or
less distinct, have been proposed since the nineteenth
century, which included neo-Lamarckian (role of disuse)
and orthogenetic (age of phyletic lines) theories. The
main models elaborated within the neo-Darwinian para¬
digm fall into two groups: those directly or indirectly
involving natural selection (adaptationists) and those
involving adaptively neutral mutations. All of them,
however, are based on the observation that regressing
structures and behaviors are generally those that
become useless in the subterranean habitat, that is, those
directly or indirectly related to light.
Because nutrient scarcity is one of the most striking
features of subterranean ecosystems and seems to
be the logical explanation for development of many
structures in troglobites, it is not surprising that
economy of energy was one of the first Darwinian
explanations evoked for character regression as well.
According to this adaptationist model, mutant indi¬
viduals for underdeveloped eyes and pigmentation,
normally excluded from epigean populations by
stabilizing selection, would have an advantage over
normally eyed and pigmented individuals in the
subterranean habitat because they could save extra
energy (otherwise spent in useless structures) for
survival and reproduction. Although appealing, this
model finds no good support from empirical data
and lacks generality—not every troglobite showing
regressive traits lives in a food-poor environment,
and some caves may be quite rich in nutrients, espe¬
cially those harboring large populations of troglox-
enes such as bats. Moreover, regression of eyes and
pigmentation is also observed in some epigean organ¬
isms such as those living in deep soil, turbid large
rivers, and also endoparasites, for which food is not a
limiting factor. On the other hand, at least in theory,
economy of energy could be in some cases associated
with other genetic models, accelerating differentiation.
Currently, there are two main theories aiming to
explain character regression in troglobites, based respec¬
tively on neutralism, that is, the accumulation of neutral
mutations, and on pleiotropic effects of selection for
constructive characters, as developmental tradeoffs. In
the 1940s, Kosswig made a case for the accumulation of
neutral mutations as the basis of regression of eyes and
pigmentation in troglobites, as characters that became
useless in a new environment. From that time on, many
genetic studies, especially on the Mexican cavefish,
genus Astyanax, were designed to put forward evidence
for the neutral model, such as those by Wilkens (2010).
The neutralism theory postulates the accumulation of
mutations that become neutral after a change in way of
life as a major cause of regression of any trait that loses
its function under the new selective regime. There is
genetic evidence that such mutations, excluded by stabi¬
lizing selection under the previous selective regime,
would no longer be eliminated and, because mutations
are more frequently deleterious than constructive,
would accumulate and cause disorganization of struc¬
tures and loss of physiological and behavioral
components.
More recently, pleiotropic effects of development
of antennae causing regression of eyes due to compe¬
tition for neural centers have been hypothesized for
North American freshwater amphipods Gammarus
minus (Culver et ah, 1995). Jeffery and colleagues,
after experimental studies of eye development in epi¬
gean and troglobitic Mexican characins, Astyanax
spp., have shown that the Hh midline signaling is a
positive regulator of many developmental events,
such as taste buds, whose number is increased in the
cavefish. However, the enhanced signaling has a
negative effect on eye development. Thus, selection
for traits positively controlled by the Hh midline
pathway would result in eye regression by default
(Jeffery, 2005).
As a general model for eye regression, the pleiot-
ropy hypothesis requires a particular association
between a regressive and a constructive character in
every troglomorphic species, which is highly unlikely
in view of its wide occurrence among so different,
distantly related taxa, from planarians, gastropods,
and arthropods, to salamanders and fishes. It is
improbable, for instance, that different mechanisms of
sensorial compensation, affecting basically the mecha-
nosensorial system in amblyopsid fishes, but the chemo-
sensorial one in siluriforms such as Rhamdia catfishes,
have the same collateral output, namely, the regression
of eyes. It is relevant that different brain centers and
embryonic placodes are involved in these sensorial
systems in fish. The model proposed for Astyanax
is also not consistent with the late regression of
eyes observed for siluriforms, such as the armored cat¬
fish Ancistrus cryptophthalmus, among other studied
Brazilian catfishes. As predominantly chemoriented
fishes, it is expected that sensory compensation
involves an increment in the taste bud number as in
Astyanax, and that eyes should also regress at an early
ontogenetic stage. However, this is contrary to the
ENCYCLOPEDIA OF CAVES
EVOLUTION OF LINEAGES
301
observed. Wilkens (2010) has shown that the neutral
model does apply to the Mexican cavefish and that
constructive and regressive developmental modules
are inherited independently in this fish; thus eye
regression would not be a spin-off effect of the
improvement of constructive characters through
pleiotropy.
Whatever the mechanism(s) leading to character
regression, it does not operate in the same way and
rate for all characters in each troglobitic species or
population, as discussed above.
It is noteworthy that autapomorphies of troglobites
do not necessarily encompass reproductive isolation.
Reproductive characters are conservative and it is
predicted that several other traits may change before
isolation reproductive mechanisms are fixed through¬
out the population. Such is the case with the Mexican
tetra characins, genus Astyanax, which encompass
at least 19 cave troglomorphic populations, some
of which have been shown to introgress with epi-
gean, eyed, and pigmented tetras to produce fertile
hybrids.
Within the paradigm of irreversibility of evolution
(Dollo's Law), regressed characters have been viewed
as the last possible state in the character transforma¬
tion series, and troglobites as evolutionary dead-ends,
unable to (re)colonize epigean habitats. Nevertheless,
some authors question this notion. Based on genetic
evidence. Culver et al. (1995) hypothesized that eyed
amphipods, Gammarus minus, living in karst windows
are derived from troglomorphic cave populations.
According to these authors, the reacquisition of large
eyes indicates that eye reduction in the original cave
populations resulted from selection for degenerative
mutations occurring at regulatory loci governing eye
expression while the structural loci remain relatively
intact. More recently (Dillman et al., 2011) applied par¬
simony to a phylogenetic hypothesis for amblyopsid
fishes, with epigean and troglophilic species as apical
taxa inside the cluster of troglomorphic species. This
implies accepting the origin of these species from trog¬
lobitic populations, therefore supporting the hypothesis
of reevolution of eyes and melanic pigmentation, associ¬
ated with the recolonization of epigean habitats.
The same reasoning is valid for the black proteus,
P. anguinus parkelj, which is restricted to a few caves
in a small area within the wide range of highly
troglomorphic P. anguinus; in a recent molecular
phylogeny, the black proteus was positioned as an
apical taxa within several distinct populations of
P. anguinus anguinus. Since vertebrates are born
with eyes and melanic pigmentation, which regress
along ontogeny, a truncation of the degenerative
process due to blocking of genes responsible by trig¬
gering such process could result in the so-called
character reversal. Clearly, this would not apply to
invertebrates.
AN OVERVIEW OF MOLECULAR DATA
Since Darwin, cave organisms have stimulated the
interest of biologists as they offer unique opportunities
to comparatively unravel evolutionary processes in an
ecologically simplified and temporally stable environ¬
ment. Pioneer studies based on morphology have
helped in gaining a better understanding of evolution in
caves revealing, at the same time, that convergence is
way more frequent in subterranean organisms than in
their surface counterparts. Hence, it became evident that
molecules would have provided an alternative, indepen¬
dent source of information, less prone to convergence
and exhibiting variation appropriate to the different
evolutionary questions posed.
In the early 1960s, molecular methods for examining
genetic variation at the protein and nucleic acid levels
started being adopted by evolutionary biologists. Since
then an increasing amount of genetic data has been
amassed for terrestrial and aquatic cave taxa and used
to test a variety of hypotheses. The next paragraphs
present an overview of the different methods and
markers that have been placed at the disposal of cave
biologists and briefly describe, through a few selected
examples, how these markers have been used to
describe evolutionary processes in caves.
Historically, assaying protein polymorphisms via
electrophoresis has been the first and most widespread
molecular method to describe allele frequency changes
in species. Protein data have broadened our under¬
standing of genetic variability of populations, gene
flow, recognition of species boundaries, and phyloge¬
netic relationships. While relatively cost-effective, the
method has implicit limitations as it requires the use
of fresh tissues (ethanol preserved samples and/or
minimal invasive sampling are not applicable), and the
number of loci and alleles that can be resolved tends
to be relatively small.
Cave crickets belonging to the eastern Mediterranean
genus Dolichopoda (Rhaphidophoridae) have been exten¬
sively studied on micro-, meso-, and macro-geographical
scales (Sbordoni et al, 2000 and references therein).
Allozymes proved informative to discriminate geograph¬
ically distant populations, to recognize processes of
cryptic speciation, to assess levels and patterns of gene
flow within species, and, more generally, to describe
the microevolutionary phenomena of adaptation and
speciation related to colonization of caves.
In the case of subterranean aquatic organisms,
levels and patterns of allozymic divergence in
the exclusively stygobiont isopods of the genus
ENCYCLOPEDIA OF CAVES
302
E
Stenasellus accurately reflect the geological evolution
of the peri-Tyrrhenian area, emphasizing the role
played by the tectonic evolution of the area in shap¬
ing the diversification of the group and allowing the
calibration of a specific allozymic molecular clock
(Ketmaier et al., 2000).
In the early 1970s DNA kinetics became an
increasingly popular venue of research and evolu¬
tionary biologists started looking at methods based
on DNA/DNA hybridization as a promising source
of information alternative to protein polymorphisms.
The rationale behind this approach is based on the
double-stranded nature of DNA; the firmer the in-
vitro bonds between previously single-stranded
DNA fragments of two different species the higher
their genetic affinity because of the relatively low
number of mismatches between the heterospecific
DNA fragments. This technique provides informa¬
tion on the degree of similarity among species and,
hence, on phylogenetic relationships. Critiques have
been raised that, while accurate at deeper phyloge¬
netic levels, DNA/DNA hybridization tends to be
more prone to experimental errors when closely
related species are compared.
Caccone and Powell (1987) presented one of the
very few study cases centered on cave organisms
based on this technical approach. DNA/DNA hybrid¬
ization data were used to reconstruct phylogenetic
relationships within and between the North American
cave crickets genera Euhadenoecus and Hadenoecus.
They found a large degree of genetic divergence
among populations and species, used the yielded phy-
logeny to calibrate a molecular clock, and concluded
that these insects had, at the time when the article was
published, one the fastest rates ever reported for
invertebrates.
Popularity of protein electrophoresis and DNA/
DNA hybridization peaked in the late 1970s/early
1980s; however, the number of studies based on these
approaches progressively waned in the mid to late
1980s and virtually disappeared from top-ranked
journals by the early 2000s. But the amount of molecu¬
lar data gathered on cave organisms has certainly not
decreased in the past two decades; rather the opposite.
Central to this quantum leap in the use of molecular
markers has been development of applications of the
polymerase chain reaction (PCR) for investigating vari¬
ation in DNA on a large scale. The possibility to easily
and cost-effectively amplify a large number of genes
also from old and/or badly preserved samples
resulted in the production of increasingly large data
sets on DNA sequence variation between and within
species. The versatility of PCR and the fact that
sequences produced worldwide have to be deposited
in open access databases ( e.g ., GenBank) have allowed
scientists to pinpoint with extreme accuracy genes
with levels of variation appropriate to the case under
study. Genes of the mitochondrial genome (mtDNA)
have played a prominent role in the past two decades,
as they are relatively easy to access, do not recombine
(and hence coalesce faster), and are (with a few excep¬
tions) maternally inherited. Sequencing of mtDNA has
proved instrumental in the fields of biogeography,
phylogeography, and speciation among closely related
species.
Allegrucci et al. (2009) sequenced multiple mtDNA
genes to reconstruct the phylogeny and phylogeogra¬
phy of the genus Dolichopoda at the scale of the eastern
Mediterranean area. Relationships among species
resulted to reflect essentially a geographical pattern,
with taxa coming from nearby areas being also geneti¬
cally more closely related. The biogeographic scenario
hypothesized to explain patterns of divergence fore¬
saw dispersal as the main force promoting divergence
followed by dispersal when the paleoclimatic condi¬
tions were favorable for surface dispersal.
A combination of vicariance and dispersal was also
invoked by Ketmaier et al. (2003) to explain the current
distribution and mtDNA phylogeography of the iso¬
pods belonging to the genus Stenasellus. The authors
also took advantage of geological data to calibrate rates
of substitutions for the cytochrome oxidase subunit
I mtDNA gene and to compare them to those available
for other invertebrates.
Leys et al. (2003) presented mtDNA evidences sup¬
porting a scenario of multiple independent origin and
allegedly sympatric subterranean speciation in a group
of diving beetles from calcrete aquifers in Western
Australia. These subterranean systems host at least
12—13 sister species pairs or triplets of diving beetles,
which can be readily discriminated from one another
on both genetic and morphological bases. Species
occurring sympatrically have nonoverlapping body
sizes; this suggests a potential mechanism of sympatric
speciation by niche partitioning of species within each
calcrete.
MtDNA data need to be complemented with the
analysis of more conserved nuclear (nuc) genes when
the aim is to disentangle ancient cladogenetic events or
to understand ecomorphological evolutionary processes.
Allegrucci et al. (2010) used a selection of mt- and
nucDNA genes to unravel tempo and mode of evolu¬
tion in a group of cave crickets belonging to the family
Macropathinae (Rhaphidophoridae). This group shows
a clear Gondwanan distribution and hence represents
an ideal system to test whether molecular-based phylo¬
genetic hypotheses mirror the geological events that
led to the breakup of Gondwana. Most of the splits in
the obtained phylogeny fit to a vicariance scenario
based on the ancient tectonic events related to the
ENCYCLOPEDIA OF CAVES
EVOLUTION OF LINEAGES
303
Gondwana fragmentation. However, some relation¬
ships remained controversial and could be explained
either in light of sporadic events of long distance
dispersal or as an artifact due to incomplete taxon
sampling.
North American Sclerobuninae harvestmen ( Opiliones)
include three genera; each of these genera comprises
series of taxa ranging from the surface ones with normal
pigmentation and eye development to those inhabiting
caves and showing different levels of troglomorphism,
including reduction or lack of tegumentary pigments,
elongated legs and palps, and reduced or absent eyes.
Derkarabetian ef al. (2010) coupled the sequencing of
multiple mt- and nucDNA genes with the morphometric
analysis of a suite of characters to gain insights into
morphological homoplasy and rates of morphological
changes in the framework of temporal evolution of
the cave systems inhabited by these harvestmen.
Phylogenetically unrelated taxa evolved troglomorphism
multiple times convergently; in one circumstance a tro-
glomorphic facies was found in specimens from a high-
elevation stony debris habitat, suggesting that troglo¬
morphism can potentially evolve also in non-cave habi¬
tats. A strong positive and linear relationship between
degree of troglomorphism and divergence time was also
found, making it possible to predict taxon age from
morphology.
At the other tip of the scale (z.e., at the intraspecific
level), nuclear microsatellite loci are the markers of
choice to obtain a finer description of the population
genetic structure of a given species.
The increasing demand for low-cost sequencing
has led to the development of technologies that paral¬
lelize the sequencing technology, resulting in the pro¬
duction of thousands of sequences at once. Parallel
(or next-generation) sequencing was launched about
ten years ago and machines started being marketed
at reasonable prices less than five years ago. The
decreasing costs of this new technology and the
humongous amount of data produced in just a frac¬
tion of a single run are promoting a gradual but deci¬
sive transition from the genetic to the genomic level
of investigation. Scanning a large portion of the
genome is becoming increasingly feasible also for
nonmodel organisms (hence including cave species)
and will allow scientists to shed light on processes
that were considered out of reach only a few years
ago. Among others, certainly the most intriguing are
those aimed at understanding the genetic basis of
adaptive and regressive evolution in cave species with
different degrees of troglomorphism. As to our knowl¬
edge a study focusing on subterranean organisms and
using next-generation sequencing has yet to be pub¬
lished but this certainly represents a fascinating and
very promising area of research that will attract the
attention and efforts of cave biologists in the near
future.
See Also the Following Articles
Asellus aquaticus: A Model System for Historical
Biogeography
Niphargus: A Model System for Evolution and Ecology
Bibliography
Allegrucci, G., Rampini, M., Gratton, P., Todisco, V., & Sbordoni, V.
(2009). Testing phylogenetic hypotheses for reconstructing the
evolutionary history of Dolichopoda cave crickets in the eastern
Mediterranean. Journal of Biogeography, 36(9), 1785—1797.
Allegrucci, G., Trewick, S. A., Fortunato, A., Carchini, G., &
Sbordoni, V. (2010). Cave crickets and cave weta (Orthoptera,
Rhaphidophoridae) from the southern end of the world: A molec¬
ular phytogeny test of biogeographical hypotheses. Journal of
Orthoptera Research, 19(1), 121—130.
Caccone, A., & Powell, J. R. (1987). Molecular evolutionary diver¬
gence among North American cave crickets. II. DNA-DNA
hybridization. Evolution, 41, 1215—1238.
Culver, D. C., Kane, T. C., & Fong, D. W. (1995). Adaptation and natural
selection in caves. Cambridge, MA: Harvard University Press.
Derkarabetian, S., Steinmann, D. B., & Hedin, M. (2010). Repeated
and time-correlated morphological convergence in cave-dwelling
harvestmen (Opiliones, Lanatores) from Montane Western North
America. PLos One, 5, el0388. doi:10.1371/journal.pone.0010388.
Dillman, C. B., Bergstrom, D. E., Noltie, D. B., Holtsford, T. P., &
Mayden, R. L. (2011). Regressive progression, progressive regres¬
sion or neither? Phytogeny and evolution of the Percopsiformes
(Teleostei, Paracanthopterygii). Zoologica Scripta, 40(1), 45—60.
Felice, V., Visconti, M. A., & Trajano, E. (2008). Mechanisms of pig¬
mentation toss in subterranean fishes. Neotropical Ichthyology, 6(4),
657-662.
Holsinger, J. R. (2000). Ecological derivation, colonization, and speci-
ation. In H. Wilkens, D. C. Culver, & W. F. Humphreys (Eds.),
Subterreanean Ecosystems. San Diego, Elsevier.
Howarth, F. G. (1993). High-stress subterranean habitats and evolu¬
tionary change in cave-inhabiting arthropods. The American
Naturalist, 142(suppl), 65—77.
Jeffery, W. R. (2005). Adaptive evolution of eye degeneration in the
Mexican blind cavefish. Journal of Heredity, 96(3), 185—196.
Ketmaier, V., Argano, R., & Caccone, A. (2003). Phytogeography
and molecular rates of subterranean aquatic Stenasellid isopods
with a peri-Tyrrhenian distribution. Molecular Ecology, 12(2),
547-555.
Ketmaier, V., Messana, G., Cobolli, M., De Matthaeis, E., &
Argano, R. (2000). Biochemical biogeography and evolutionary
relationships among the six known populations of Stenasellus
racovitzai (Crustacea, Isopoda) from Tuscany, Corsica and
Sardinia. Archiv fur Hydrobiologie, 147, 297—309.
Leys, R., Watts, C., Cooper, S. J. B., & Humphreys, W. F. (2003).
Evolution of subterranean diving beetles (Coleoptera: Dytiscidae:
Hydroporini: Bidessini) in the arid zone of Australia. Evolution,
57, 2819-2834.
Niemiller, M. L., & Poulson, T. L. (2010). Subterranean fishes of
North America: Amblyopsidae. In E. Trajano, M. E. Bichuette, &
B. G. Kapoor (Eds.), Biology of subterranean fishes (pp. 169—280).
Enfield, NH: Science Publishers.
Parzefall, J., & Trajano, E. (2010). Behavioral patterns in subterranean
fishes. In E. Trajano, M. E. Bichuette, & B. G. Kapoor (Eds.),
ENCYCLOPEDIA OF CAVES
304
E
Biology of subterranean fishes (pp. 81 — 114). Enfield, NH: Science
Publishers.
Sbordoni, V., Allegrucci, G., & Cesaroni, D. (2000). Population genetic
structure, speciation and evolutionary rates in cave-dwelling
organisms. In H. Wilkens, D. C. Culver, & W. F. Humphreys (Eds.),
Ecosystems of the world 30: Subterranean ecosystems (pp. 450—483). San
Diego: Elsevier.
Trajano, E. (2007). The challenge of estimating the age of subterranean
lineages: Examples from Brazil. Acta Carsologica, 36, 191—198.
Wessel, A., Erbe., P., & Hoch., H. (2007). Pattern and process:
Evolution of troglomorphy in the cave-planthoppers of Australia
and Hawaii—Preliminary observations (Insecta: Hemiptera:
Fulgoromorpha: Cixiidae). Acta Carsologica, 36, 199—206.
Wilkens, H. (2010). Genes, modules and the evolution of cave fish.
Heredity, 105(5), 413-422.
EXPLORATION OF CAVES—GENERAL
William B. White
The Pennsylvania State University
The skills, training, and equipment necessary to
explore caves vary tremendously with the size and
complexity of the cave. In common, however, are the
environmental realities that caves are dark, usually
wet and muddy, and have temperatures that are close
to the regional average. The terrain is rugged with low
crawls, narrow fissures, pits, puddles, pools, flowing
streams, and piles of loose rocks. Only rarely does a
cave passage have a level floor. The equipment needed
by the explorer includes reliable light sources, clothing
that provides protection from the environment, and
any additional equipment needed for vertical and
water-containing parts of the cave system.
SKILL LEVELS IN CAVE EXPLORATION
Small, Near-Horizontal Caves
In regions of low to moderate relief, caves tend to
form as near-horizontal passages although there may be
many small-scale irregularities in the form of short
climbs, drops, and piles of broken rock. Exploration of
these caves requires a certain amount of equipment and
a certain skill level, but generally nothing that requires
extensive training. Even in small, near-horizontal caves,
there are certain minimum requirements for exploration.
Above all, caves are dark. Beyond the twilight zone
at the entrance, darkness is absolute. Failure of light
sources is the single most common risk in caving.
Maneuvering one's way through even a small cave by
touch alone is exceedingly difficult and for a cave of
even moderate complexity, probably impossible.
Requirement (1): The explorer must understand the
necessity of reliable light sources, backup sources, and
the expected lifetime of light sources.
Cave temperatures are usually constant and near
the mean annual temperature of the region. Caves in
temperate climates are in the range of 8—12°C. Caves
at high latitudes and/or high altitudes are colder,
often in the range of 0—4°C. Caves tend to be wet and
humid. The combination of low temperature and high
humidity accelerates heat loss with risk of hypother¬
mia. Tropical caves may pose the opposite problem.
High temperatures, 20—25°C, combined with high
humidity restricts heat loss with the resultant risk of
overheating and possibly heat exhaustion. Some caves
contain ponded water and/or flowing streams, which
require crawling, wading, or swimming in water that
is also at or close to cave temperature. Requirement (2):
The explorer must dress appropriately both for the
expected environmental conditions and also for the
expected activity. A fast-moving exploration team has
different heat management requirements than a team
engaged in survey or photography.
The cave terrain is rugged. Passages are irregular
often scaling from dimensions of tens of meters to
crawlways at the limit of human penetration. Ceilings
are irregular and protruding ledges are common.
Passage floors are often muddy. Loose rocks (break¬
down) occur as isolated fragments and as piles tens of
meters high. Passages may be discontinuous with ver¬
tical offsets that require climbing up or down. Some
passages are traversed on ledges or by hopping along
canyons that have been incised into the main passage
floor. Requirement (3): The explorer must be prepared
with foot, head, and body protection against the abuse
required for traversing cave passages.
Vertical Caves
Vertical caves are those that have substantial eleva¬
tion differences between the lowest and highest points
in the cave. To be considered a vertical cave, the eleva¬
tion difference should occur over a relatively short
horizontal distance. There exist caves, some volcanic
caves, for example, with elevation differences between
the top and bottom of more than 1000 m but with only
gently sloping passages. The vertical component of
caves may consist of pits, sequences of pits sometimes
interconnected by small passages, and vertical crevices
of various sizes, depths, and shapes. Pits are often
pathways for descending water ranging from trickles
down walls to large waterfalls.
Vertical caves may be entered from the top, from the
bottom, or from various points between. The require¬
ments for exploring vertical caves are more rigorous
than those for small horizontal caves and usually
require special training on the rigging and use of ropes
and on techniques for descending and ascending ropes.
Encyclopedia of Caves.
© 2012 Elsevier Inc. All rights reserved.
EXPLORATION OF CAVES—GENERAL
305
Large Cave Systems
What constitutes a "large" cave as distinguished
from a "small" cave is to some extent in the eye of the
beholder. As used in this article, a large cave is one
that cannot be explored in its entirety in a single day.
Exploration of large caves requires developing techni¬
ques for push caving, for project caving and for expedi¬
tion caving possibly including cave camps. Push
caving is done by a team that simply drives itself to
continue exploration for 20- to 30-hour stretches.
Requirements are mainly superb physical conditioning
and the mental stamina to keep going to exhaustion
and beyond. Project caving has an organizational
structure. Teams return to the same cave over and
over. Some type of organizational structure must be
assembled. Records must be kept. The advances in
exploration and survey are recorded so that each
exploratory push adds to the previously explored cave
in a systematic way. In addition to the standard caving
skills, the requirement for an expedition caver is a will¬
ingness to accept a measure of discipline and control.
The term expedition caving is applied to caves that are
either or both difficult of access and difficult to reach
the limits of exploration. Expedition caving may
require an expedition to reach the cave and may also
require extensive rigging and advance camps within
the cave to advance the exploration.
Underwater Caves
It has, of course, been known for a long time that
caves exist below regional base levels or below sea
level. Until the past 20—30 years, the exploration of
these caves was very limited because they were full of
water and the necessary equipment did not exist.
Although a few heroic explorations were made going
back into the nineteenth century using pressurized
suits and air compressors, it was the invention of
self-contained underwater breathing apparatus
(SCUBA) that allowed explorations to be extended to
underwater caves.
Requirements for underwater exploration are the
most formal and the most extensive. Formal training
on the equipment is needed along with formal certifi¬
cation of cave-diving qualifications. A distinction can
also be made between clear water diving like that
found in many of the large spring caves in Florida
and the Yucatan of Mexico and what may be called
northern sump diving. The water-filled passages of
many temperate and northern climate caves are lined
with silt, such that movement of a diver's body
quickly stirs up the silt and reduces visibility to
near zero.
BASICJiQUIPMENT
Light Sources
The most basic items of equipment are the light
sources. Without light, no exploration is possible and
if all lights of an exploration party are lost, the party
can expect to wait in the dark for rescue. Cavers are
advised to each carry three sources of light, which are
called the primary, secondary, and tertiary sources.
New technology, particularly the development of
white light-emitting diodes (LEDs), has greatly
enlarged the caver's choices of light sources.
For many years the primary light source was the
miner's carbide lamp. Most American cavers used a
helmet-mounted lamp that consisted of a lower cham¬
ber containing lumps of calcium carbide, an upper
chamber containing water, and an adjustable valve
that allowed water to drip into the carbide producing
acetylene gas. The acetylene produced in the lower
chamber was ejected through a nozzle where it was
burned to produce a bright yellow flame. A version
more popular in Europe used a belt-mounted carbide
and water chamber connected by a hose to the helmet-
mounted nozzle and reflector. The chemical reaction is
CaC 2 + 2H 2 O^C 2 H 2 + Ca(OH) 2
Carbide lamps had many advantages. They were
nearly indestructible. A small repair kit with spare
nozzles, filters, gaskets, and other parts allowed
replacement of failed parts in the cave. A waterproof
bottle holding 500 g of carbide would keep the lamp
going for tens of hours. The large flame made a great
deal of light. The lamp could also be used to warm a
meal. The sooty flame was often used to mark survey
stations. A carbide lamp combined with a plastic gar¬
bage bag tent offered some protection against hypo¬
thermia. The downside of carbide lamps was that the
spent carbide is a white, pasty solid consisting mostly
of calcium hydroxide but also containing all of the
impurities that were present in the carbide. Spent car¬
bide must be carried out of the cave because it is
strongly alkaline and toxic to cave organisms. A fur¬
ther complication is that security authorities regard
carbide as an explosive material and there are exten¬
sive restrictions on carrying or shipping the material.
Carbide lamps have been to a large extent replaced
by battery operated light sources. These are available
in a great variety of forms. The miner's Wheat lamp
features a helmet-mounted fixture that holds the bulb
and reflector with power provided by a belt-mounted
rechargeable battery. Other belt-mounted lamps are
used. Other primary light sources are helmet-mounted
lamps with a small battery pack attached to the back
of the helmet. These use disposable alkaline batteries.
ENCYCLOPEDIA OF CAVES
306
E
typically the AA size. They are much lighter than other
lamps but require the caver to carry a sufficient supply
of spare batteries.
Filament-type lamps are rapidly being replaced by
arrays of white light-emitting diodes (LEDs). LEDs are
more rugged than filament bulbs and are much more
efficient, producing light for a longer time on a set of
batteries. Some cavers carry dual light sources on their
helmets—an LED array as their primary source and a
filament lamp that can be focused to provide a sharp
bright beam for probing high ceilings and possible
side passages. The disadvantage to helmet-mounted
lamps is that they provide flat lighting on an irregular,
generally monochrome landscape. A hand-held lamp
can provide better contrast on irregular cave floors.
The secondary source is often a flashlight of some
kind. Small, rugged, waterproof flashlights such as
the aluminum-bodied Maglites are widely used.
Secondary sources are needed to provide light when it
is necessary to refuel or repair the primary source.
However, it should be remembered that the secondary
will become the primary source if the original primary
fails completely. For this reason the secondary source
is often a second lamp similar to the primary source.
The tertiary source is the backup to the backup. In
the early days, candles and a waterproof container of
matches were often used. Escaping from a cave with a
candle is not easily done. The present-day choices are
the small LED devices, which are available in a variety
of forms, some of which are even designed for key
rings. The single LED will provide marginally enough
light to find one's way through the cave and it will
work for 30 hours or more on a single battery.
Hard Hats
The most important item in the caver's outfit is the
hard hat. Caves have low and irregular ceilings; there
are projecting ledges and other objects on which to
bang one's head. Loose rocks get knocked down pits.
Hard hats are essential. Two styles are in common use:
miners' hats with a snout, and rock-climbers' helmets,
which are rounded (Fig. 1). In either style, there is an
outer shell of tough polymer that provides the actual
protection and an inner harness that keeps the shell
from coming into contact with the head. Hard hats
should have chin straps with safety release features.
Hard hats without straps can be easily knocked off
during climbing. However, cavers have been known to
become trapped by their hard hats when the caver slid
down a narrow fissure only to discover that the hard
hat would not fit; thus the necessity for a release
feature on the chin strap. Hard hats should have
mounting brackets for lamps, whether carbide or
FIGURE 1 A comparison of caving helmets. Left: a rock-climbing
helmet fitted with a lamp bracket. Middle: a modern caving helmet
with attached LED headlamp and battery pack on the back. Right: a
miner's helmet with carbide lamp.
electric. Helmet-mounted light sources leave the hands
free for climbing or other activities.
Clothing
Beginning cavers in small caves may wear any
clothing that is rugged and that they don't mind
getting dirty. More advantageous is a coverall type of
outfit that protects the entire body, keeps mud off the
inner clothing, and can be peeled off after exiting
the cave. Specialty coveralls of tough nylon are avail¬
able in a variety of bright colors. The desired clothing
is dictated by the cave, its temperature, and presence
of water and mud.
Underclothing should be selected to preserve heat,
particularly if the cave is wet. Wool has this property.
Cotton does not. Special clothing of polypropylene has
become available for serious cave exploration.
Footwear should provide ankle support and have a
good gripping sole. In the United States, ankle-high
boots with cleated soles are often favored. European
cavers have a preference for high-topped rubber boots
("Wellingtons"). These have the advantage of allowing
the caver to wade calf-deep water with dry feet. Wool
or polypropylene socks are recommended.
Cave Packs
Because of climbing, crawling, and the frequent
necessity for squeezing through tight places, it is
unwise for the caver to carry very much in his or her
pockets. Most cavers carry packs in which to transport
food, water, spare batteries or spare carbide for their
lights, a first aid kit, repair kits for lights, surveying
equipment, and anything else that might be needed
ENCYCLOPEDIA OF CAVES
EXPLORATION OF CAVES—GENERAL
307
underground. Most needed personal gear can be
carried in a small side pack, but if the trip involves ver¬
tical caving, larger packs for ropes, ladders, and other
climbing equipment will be needed. Packs must be
made of tough material that will not tear as the pack is
pushed, dragged, or rolled through the cave. It must
remain tightly closed so that the contents are not lost.
Straps work, whereas zippers tend to become clogged
with mud. Tough waterproof packs are used to trans¬
port camping gear when advanced camps are needed
VERTICAL CAVING
Vertical caves range in depth from a drop of just a
few meters into a cave entrance to deep alpine and
high plateau systems where sequences of pits have
reached depths of 1500 m and more. The techniques
for descending and ascending vertical caves have
evolved and improved greatly during the past several
decades. Specialized equipment is well-established.
The use of it requires training and practice, which is
best accomplished under the guidance of experienced
vertical cavers.
Going Down
The early days of pit exploration involved winches
and lowering lines along with rope and cable ladders
of various types. In early twenty-first century caving,
the only surviving technique from the early days is the
cable ladder. The ladder consists of two strands of
woven cable as the legs of the ladder. The rungs are
short (typically 15 cm) lengths of aluminum tubing,
crimped to the cable a convenient distance apart. Thirty
meters of cable ladder can be rolled to a diameter of
30 cm or less and stuffed in a pack. Sections of ladder
can be strung together for long drops but today are
rarely used for this purpose. Ladders are mainly used
for short drops to avoid the nuisance of rigging climb¬
ing gear and for tight crevices where single-rope techni¬
ques are awkward. Ladder climbs should be belayed
because of the possibility of slipping off the ladder.
Nearly all descents of vertical caves are made with
what are known as single-rope techniques. The ropes
themselves have undergone substantial innovation
with older woven (laid) ropes being replaced by sheaf
ropes. Sheaf ropes have a core of twisted polymer to
provide strength surrounded by a braided sheaf. The
sheaf allows the rope to slide smoothly through des¬
cending devices without the twisting experienced on
the older laid ropes. Caving ropes are strong but are
sensitive to abrasion. They must be rigged where they
do not rub against the rock. This often requires that the
rope be attached to a set of secondary anchors bolted
to the wall along the drop. These are known as rebelays
and require the caver to learn how to pass them, both
descending and ascending. In part, the extensive use of
rebelays has been made possible by the invention of
the battery-powered drill which allows setting bolts in
awkward places without excessive effort.
Descending a pit is a question of connecting oneself
to the rope and sliding down it in a controlled manner,
a procedure known as a rappel. Many rappelling
devices have been devised. One of the most popular is
the rappel rack, a U-shaped steel rod. Across the U are
placed a series of brake bars, usually of aluminum.
These have holes at one end for threading over the U
while the other ends are notched so that the brake bars
can be hinged away from the rack, thus allowing bars
to be added or removed according to the weight of
rope hanging below. The rope is threaded through the
brake bars in such a way that the bars are locked
against the U-rod of the rack and friction between the
rope and the brake bars controls the speed of descent.
The caver wears a harness of nylon webbing that is
attached to the rack by a carabiner.
Going Up
Pits can be ascended by cable ladder, which works
equally well going up or going down. However,
ascending a pit or sequence of pits using single-rope
techniques requires some means for climbing the rope.
Prusik knots are a form of slip knot that will clamp
around the main rope when tension is put on them
and release when the tension is released. A set of three
Prusik knots, two connected to the climber's feet and
the third attached to the chest harness, provides a slow
and tedious method of rope climbing but is useful as
an emergency measure. Most cavers use ascenders of
some type. There are several designs but all are based
on a moving toothed cam that presses against the
main rope when the climber's weight is on it. It
releases and can be moved upward when the climber's
weight is released. A number of ways have been
devised for arranging the ascenders, each with their
adherents. The overall result is that the caver essen¬
tially walks up the rope.
Some caves are explored from the bottom, which
means that explorers need to ascend drops or pits
without benefit of a rope. Standard rock-climbing tech¬
niques are used but are more limited because of poor
lighting, wet and muddy surfaces, and the presence of
weak or rotten rock. Most vertical ascents require dril¬
ling holes in the rock, setting bolts, and attaching
anchors that support the climber while new and higher
bolts are set.
ENCYCLOPEDIA OF CAVES
308
E
LARGE SYSTEMS: EXPEDITION CAVING
AND PROJECT CAVING
It is useful to make a distinction between caving and
cave exploring. Caving is simply traversing the cave,
whether for recreation, scientific observations, photog¬
raphy, or other objectives. Cave exploring is the system¬
atic probing of all passages to obtain a complete
description of the cave. For present-generation
explorers, this implies that the passages will be sur¬
veyed and a map produced. Systematic exploration
and survey is much more time-consuming than a
casual trip through the cave and, for caves of even
modest size, may require multiple trips or even multi¬
ple years of effort; thus, the terms expedition caving or
project caving.
The approach to expedition caving depends on
whether or not the frontier of exploration can be
reached from some entrance in a reasonable period of
time. If so, the expeditions can be based at the surface.
Exploration parties enter the cave, proceed quickly to
the point at which exploration is to begin, explore and
survey, and then return to the surface to eat and sleep.
The alternative is a cave in which reaching the frontier
of exploration requires so much time that there is no
time left for exploration, much less returning to the
surface. Such caves require setting up underground
camps.
With either type of large cave, complete exploration
and survey requires a sustained effort and thus an orga¬
nizational structure. The structure may take the form of
a formal organization or a loosely knit "project." In
either case, some leadership is necessary to coordinate
expeditions, data collection, and data storage. Survey
results need to be deposited in central files and
between-expedition activity is necessary to plot draft
maps and prepare objectives for the next expedition.
Lengths and organization of expeditions also
depend on the accessibility of the cave. Many caves in
the United States and Europe are within easy access of
roads. The style is to conduct frequent weekend or
weeklong expeditions. Expeditions to caves in inacces¬
sible regions or that require extensive travel such as
the deep caves of Mexico or the giant caves of
Malaysia often extend for periods of months. For these
expeditions, arrangement for supplies and their trans¬
port to the expedition site becomes an important part
of expedition management.
Perhaps the most elaborate effort at project caving is
the Cave Research Foundation. Although it has greatly
expanded its original objectives, the Cave Research
Foundation was organized to explore and survey the
complex Mammoth Cave System in Kentucky. It has
been at the task for more than 50 years and the end is
not yet in sight. In the course of this 50-year effort, the
length of the system was expanded from something
like 40 km to almost 600 km. Few places in the
Mammoth Cave System are exceptionally difficult; it's
just that there is so much of it.
WATER AND UNDERWATER CAVING
Big-River Caves
Most caves are wet and many contain pools and
flowing streams. Most of these caves can be traversed
only by getting wet. However, a few caves exist that
contain large, high-gradient rivers not different in kind
from white water rivers on the surface. Exploration of
such caves requires techniques beyond those of basic
caving.
Cave exploration that requires total immersion in
water requires protective clothing beyond the usual
warm underclothes and coveralls. Wet suits are tight-
fitting full-body outfits made of polymer foam with a
protective smooth outer polymer layer. As the name
implies, the wet suit allows water to soak through but
provides an insulating layer to preserve body heat.
Wet suits are widely used in any cave exploration that
requires extensive exposure to water. They are often
worn under coveralls for protection against snags and
abrasion. Less common are dry suits, which are full-
body outfits of water-impermeable polymer. They
keep water out, but they also keep perspiration in. Dry
suits work best for waterfalls and whitewater streams
where the caver may be immersed for only short peri¬
ods. Once out of the stream, the suit can be opened to
allow perspiration and waste heat to escape.
Exploration of big-river caves is one of the most
hazardous types of cave exploration. The distinction
between a stream cave and a big-river cave can be
made on the basis of whether or not a caver can wade
the stream and not be swept away. Exploration of
big-river caves requires much the same techniques as
exploring vertical caves. Anchors must be placed and
ropes played out to prevent the explorers from being
washed away. Waterfalls must be rigged with ropes.
Deep pools may require swimming. Great care must
be taken to avoid being pinned by the force of flowing
water.
Underwater Caves
Many caves end in sumps. Sumps are places where
the ceiling of the passage dips below the surface of
standing water. Some sumps are short and shallow.
A dive of a few meters or a few tens of meters brings
the explorer back up into an air-filled passage.
ENCYCLOPEDIA OF CAVES
EXPLORATION OF CAVES—GENERAL
309
Some sumps are longer and deeper. Some caves are
completely underwater. Cave systems in Florida, the
Yucatan Peninsula, and the Bahamas have been
drowned by the sea-level rise that took place at the
end of the ice ages. Underwater caves can be quite
long. Six caves are known in Florida with lengths in
excess of 5 km. The underwater cave Ox Bel Ha in the
Mexican state of Quintana Roo has a surveyed length
of 70.65 km and a dozen other underwater caves are
known in the Yucatan with lengths exceeding 5 km.
Both sump diving and underwater cave exploration
require complex SCUBA equipment, extensive train¬
ing, and, above all, experience. Formal training and
certification is offered by the Cave Diving Section of
the National Speleological Society. Underwater caving
is far less forgiving of mistakes and equipment failures
than any other type of caving. Rescues are rare. Body
recoveries are all too common.
The horizontal limits of underwater exploration are
dictated by air supply. The distance a diver can travel
before one-third of the air supply is consumed was for
a long time limited by how far the diver could swim.
A new development is the use of underwater propul¬
sion systems (scooters), which allows divers to go
much farther with the same air supply. Most of the air
in SCUBA tanks is wasted; it is breathed once and then
exhaled as a stream of bubbles. Another development,
however, is the invention of the rebreather, a device
that recycles the air by extracting carbon dioxide. This
ability greatly extends the diver's range.
The depth limits of underwater exploration are lim¬
ited by the need for decompression. Pressure increases
with depth, causing an increased solubility of gases,
particularly nitrogen, in the diver's blood. Ascent from
an underwater cave must be done slowly, lest nitrogen
form bubbles in the blood, bringing on a painful condi¬
tion known as the bends. Special gas mixtures have
been devised to partially alleviate the problem.
SAFETY
Whether recreational caving in relatively small,
near-horizontal caves or undertaking expedition cav¬
ing in deep vertical caves, caving parties must take
responsibility for their own safety. Rescue of disabled
persons from caves is an extremely complex and
expensive operation. Most rescue teams are volunteers.
Their services should be reserved for serious accidents
and not wasted on people who simply forgot to take
extra batteries for their flashlights.
For beginning cavers in relatively small, near-horizontal
caves, the rules are simple: (1) Never cave alone. A
party of three works well. If one caver has accident,
another can go for help while the third remains with
the injured caver. (2) Carry the necessary light sources
and know how many hours of service they will
provide. Be out of the cave before the lights go out.
(3) Always notify someone about where you are going
and when you expect to return. If the entire party does
not return, there is someone to call for help. (4) Be
careful. Any injury that incapacitates a caver will
require a massive rescue effort.
Most countries with a large number of caves or a
large number of cavers have some type of formal
cave rescue organization. In the United States, it is
the National Cave Rescue Commission (NCRC). The
NCRC is on call throughout the United States and
maintains close communication with police and other
fire and rescue organizations. It also offers formal
training courses for those who might wish to partici¬
pate in cave rescue activities.
CONSERVATION
Caves are fragile. Time moves slowly underground.
Any damage done to a cave—broken speleothems,
graffiti, litter—will take a very long time for natural
processes to repair. Organisms that live in caves are
also easily disrupted. There is an ethic among cavers
to do their explorations with the minimum possible
impact on the cave and its inhabitants. Anyone who
visits caves, no matter how casually, should make
the maximum effort to preserve the underground
environment.
CLOSING COMMENTS
This article is a summary, not a technique manual.
Those who are contemplating becoming cave explorers
are urged to join experienced explorers and learn from
their experience. Most countries have formal cave
exploring organizations. In the United States the orga¬
nization is the National Speleological Society with
headquarters in Huntsville, Alabama. The bibliogra¬
phy lists some books on techniques and also books
that describe some of the great explorations that have
occupied cavers for many years.
See Also the Following Articles
Exploration of Caves—Vertical Exploration Techniques
Exploration of Caves—Underwater Exploration Techniques
Entranceless Caves, Discovery of
Recreational Caving
ENCYCLOPEDIA OF CAVES
310
E
Bibliography
Brucker, R. W., & Watson, R. A. (1976). The longest cave. New York:
Alfred A. Knopf.
Damon, P. H., Sr. (1991). Caving in America. Huntsville, AL: National
Speleological Society.
Dasher, G. R. (1994). On station. A complete handbook for surveying and
mapping caves. Huntsville, AL: National Speleological Society.
Exley, S. (1994). Caverns measureless to man. St. Louis, MO: Cave Books.
Farr, M. (1991). The darkness beckons. London, England: Diadem Books.
Hanwell, Price, D., & Witcombe, R. (2010). Wookey Hole: 75 years of
cave diving and exploration. Wells, Somerset, U.K: Cave Diving Group.
Judson, D. (1984). Caving practice and equipment. Newton Abbot, U.K:
David & Charles.
Marbach, G., & Tourte, B. (2002). Alpine caving techniques. Allschwil,
Switzerland: Speleo Projects.
NSS Caver Training Committee (1992). Caving basics. Huntsville, AL:
National Speleological Society.
Padgett, A., & Smith, B. (1987). On rope. Huntsville, AL: National
Speleological Society.
Steele, C. W. (1985). Yochib: The river cave. St. Louis, MO: Cave Books.
Steele, C. W. (2009). Huautla: Thirty years in one of the world's deepest
caves. Dayton, OH: Cave Books.
Stone, W., & Am Ende, B. (2002). Beyond the deep. New York: Warner
Books.
EXPLORATION OF CAVES—
UNDERWATER EXPLORATION
TECHNIQUES
Jill Heinerth
Heinerth Productions, Inc.
When some cavers reach a water-filled sump, a sigh of
disappointment rises in their throats—the end of acces¬
sible exploration. Yet to cave divers, this is the portal
to a new frontier of discovery. In the past 70 years, the
underwater exploration of caves has evolved from a
high-risk adventure for self-described gear inventors
to an evolved application of new technology driven by
human endurance.
The evolution has been rapid and extensive and
world record penetration dives of two decades past
are now being conducted as routine excursions in a
cave-diving class. Exploration has moved from wide
open passages into tight, secluded sites, highly techni¬
cal expeditions to remote locations, and sump dives to
previously unimaginable depths. Beyond stretching
the limits of human endeavor, cave-diving exploration
has joined the ranks of serious "technical diving."
Three factors have contributed to increased range
and productivity within cave-diving exploration:
wider availability of training and mentoring, improved
quality and availability of cave-diving equipment, and
advanced technology.
Currently, cave diving educational programs are
available from training agencies on every continent. The
distinction between for-profit and nonprofit training
organizations is quickly disappearing. From Finland to
China to Australia, organizations and social networks
have united people with cave-diving interests and the
desire to seek general and specialized educational
opportunities. For the most part, that explosion of avail¬
ability and interest has been relatively recent. After for¬
ward progress was halted by a water-filled sump in the
Mendip Hills of the United Kingdom in 1935, cavers
organized mentoring programs and equipment for the
exploration in the flooded caves of Somerset. The goal
of the British Cave Diving Group (CDG) was to teach
good cavers how to dive as opposed to training divers
how to explore underwater caves. In North America,
training emerged much later. American cave diving
pioneer Sheck Exley published Basic Cave Diving:
A Blueprint for Survival in 1977, taking a serious look at
accident analysis and creating the foundation for for¬
malized training. Since that time, numerous interna¬
tional organizations have used his basic rules as the
centerpiece for teaching safe cave diving. In the past
decade, entry-level recreational diver certifications may
have leveled off, but participation in technical diving
programs is still growing. In a recent survey conducted
by the SDI/TDI Group (Scuba Diving International™
and Technical Diving International™) almost four out of
ten divers indicated they were "diving tech" or they
were interested in taking a technical diving program of
some sort within 12 months (Lewis, 2011). Cave diving
represents a large segment of the technical diving com¬
munity, which includes deep-wreck divers and anyone
diving beyond the range of direct and immediate ascent
in an emergency situation.
The earliest training programs focused on essential
cave-diving skills. These days, specialized training by
accredited and insured instructors is available in activ¬
ities including side-mount and no-mount exploration,
deep mixed-gas cave diving, diver propulsion vehicles,
photography, surveying, cartography, and the use of
closed-circuit rebreathers in caves. Classes for many of
these pursuits are available in Florida, Mexico, the
Bahamas, France, Italy, Spain, Scandinavia, Australia,
South Africa, Brazil, and Russia, with more destina¬
tions, instructors, and infrastructure growing around
the world.
Equipment development for cave divers was aggres¬
sive in the 1970s, driving many of the safety enhance¬
ments that we take for granted in diving today. Valve
manifolds, octopus second stages, buoyancy devices,
and reels were developed and refined in this decade.
Accurate underwater cave maps matured correspond¬
ingly, headed in America by Sheck Exley, Bob
Friedman, and Frank Martz. Lighting instruments were
also improved and new "Goodman" light-head handles
enabled the use of two hands for reel work and survey.
Additional tanks, called stage bottles, permitted divers to
Encyclopedia of Caves.
© 2012 Elsevier Inc. All rights reserved.
EXPLORATION OF CAVES—UNDERWATER EXPLORATION TECHNIQUES
311
extend their penetrations significantly and the mapped
underwater cave passages in north Florida were
extended more than 30 miles. However, with increased
activity came higher mortality rates. In 1974, fatalities
peaked with 28 individuals perishing in water-filled
caves (Bozanic and Halpern, 1974).
In the early 1980s, cave diving became the subject of
the mainstream media. A television program. Descent
into Darkness, introduced the sport to wider audiences.
During this period, the first NSS-CDS Cave Diving
Manual was published. Toward the end of the 1980s,
cave diving became truly technical in nature. Closed-
circuit rebreathers (CCRs) (Fig. 5), exotic gases, and
diver propulsion vehicles (DPVs) increased the range
of individual divers and groups, including the United
States Deep Caving Team and the Woodville Karst
Plain Project. In 1987, Dr. Bill Stone made the first 24-
hour cave excursion using a CCR (Fig. 4), ushering in
the dawn of a new age of extremely technical dives.
Currently, CCRs, DPVs, and small, high-powered
HID and LED lights are facilitating remarkable explo¬
ration. Advanced dive computers, capable of multigas
and multialgorithm profiling, have made decompres¬
sion diving safer and more efficient. Decompression
algorithms are traded on iPhones and BlackBerrys
instead of being held secretly in the hands of elite
teams. Time-worn, popular caves such as Ginnie
Springs (Devil's Ear and Eye), Peacock Springs State
Park, and other sites are enjoying a new era of explora¬
tion, as young divers penetrate beyond the ends of
lines, often in passages that were previously thought
to be too small for continued exploration. In some
cases, completely new complex base levels and connec¬
tions to other openings are being reported and
surveyed.
One of the tools supporting new cave exploration is
the side-mount diving technique (Fig. 1). Divers have
come to recognize that this versatile configuration is
one of the most comfortable, stable, and safe ways to
enter smaller passages that were previously over¬
looked. Although this style of cave diving is not new,
off-the-shelf equipment solutions from trusted manu¬
facturers, new textbooks, and training opportunities
have expanded the practice of side-mount diving. At
the time of writing, several international training agen¬
cies have specialized programs in side-mount diving,
including the International Association of Nitrox and
Technical Divers (IANTD), National Association for
Cave Diving (NACD), and National Speleological
Society/Cave Diving Section (NSS-CDS). The follow¬
ing agencies provide training and certification in side-
mount open water diving: International Association of
Nitrox and Technical Divers (IANTD) and the
Professional Association of Diving Instructors (PADI).
The first side-mount cave exploration took place
beneath the hills of northern England in the early
1960s. British cave diver Mike Boon explored the
sumps of Hardrawkin Pot in Yorkshire, England, by
slinging the cylinder on his side with a bandolier-style
harness.
FIGURE 1 Brian Kakuk explores the beauty of "The Badlands" of Dan's Cave in Abaco, Bahamas. His efforts are bringing the country
ever closer to preserving this unique region. South Abaco Blue Holes Conservation Area, as a part of the National Park system. Photo by the
author.
ENCYCLOPEDIA OF CAVES
312
E
The impetus for side-mount diving in North
America came from a much darker episode. In the late
1970s, a diver fatally pinned himself in an impossibly
narrow crevice in Royal Spring in north Florida.
Divers Sheck Exley and Wes Skiles were unable to
reach him in their traditional back-mounted cylinders.
Exley, holding a tank at his side, slid into the crevice
beside the dead diver, breaking him loose from the
cave. Skiles, seeing this configuration in action, quickly
realized the potential for exploration of smaller pas¬
sages, and thus, side-mount cave exploration began to
expand through the United States. Other pioneering
explorers including Woody Jasper, Forrest Wilson,
Court Smith, Mark Long, Tom Morris, and Lamar
Hires (Fig. 2) continued to evolve and refine the sys¬
tem, lending to its popularity today.
Perhaps the greatest paradigm shift in underwater
cave exploration is the recent adoption of rebreather
technology. In the late 1990s, Dr. Bill Stone pioneered
the development and marketing of the first fully elec¬
tronic CCRs that were targeted to noncommercial
divers and explorers. At Wakulla Springs, the U.S.
FIGURE 2 Lamar Hires demonstrates the perfect trim afforded
by the side-mount diving system. Photo by the author.
Deep Caving Team, using Stone's Cis-Lunar MK5P
rebreathers, along with his advanced sonar-based
mapper and FatMan scooters, accomplished the first
accurate three-dimensional model of an underwater
cave environment (Fig. 3). Some of the most impres¬
sive deep cave penetrations of the decade were logged
by members of the Woodville Karst Plain Project
(WKPP) using a purpose-built semi-closed circuit
rebreather (SCR). In 2007, a seven-mile traverse by
Jarrod Jablonski and Casey McKinley (Kernagis et al.,
2008) linked two separate openings and had the cave¬
diving community wondering whether they had
reached the boundaries of human endurance. Yet, in
the past decade, supported by the wide commercial
availability of CCRs, a new wave of exploration is
under way. AP Valves (Inspiration and Evolution),
Innerspace Research (Megalodon), KISS, and others
have established a viable worldwide commercial
rebreather market. Several organizations support
rebreather training programs, giving cave divers access
to safe, standardized instruction. Once a rare curiosity,
rebreathers are now common technology at most tech¬
nical diving sites.
Rebreather diving is not a new idea. Rebreathers
were available long before traditional open circuit
scuba equipment. Open-circuit divers use either back-
or side-mounted high-pressure cylinders affixed to
buoyancy control devices (BCD) called icings. A regu¬
lator attached to the cylinder lowers the pressure of
the supply gas so that it can be delivered to the diver
on demand as she inhales. As the diver exhales, the
resulting bubbles are vented directly into the water
column. Since humans only use a very small portion
of the oxygen molecules in an inhaled breath for
metabolism, they exhale vital leftovers and create
FIGURE 3 The author driving the Wakulla Mapper and using the
Cis-Lunar MK5P rebreather during the United States Deep Caving
Team explorations of the late 1990s. Photo by Wes Skiles/United States
Deep Diving Team. Used with permission.
ENCYCLOPEDIA OF CAVES
EXPLORATION OF CAVES—UNDERWATER EXPLORATION TECHNIQUES
313
FIGURE 4 A side-mount diver may easily remove a single tank
to negotiate a restriction and minimize impact on the cave itself.
Photo by the author.
FIGURE 5 The author diving in a Sentinel Expedition rebreather,
which offers not only a gaseous carbon dioxide monitor, but also a
large heads-up display with lights and vibrating buzzer that alerts
the diver to life-threatening scenarios. Photo by the author.
waste. A rebreather, on the other hand, takes advan¬
tage of the exhaled bubbles by recapturing them,
scrubbing them clean of carbon dioxide, and returning
them to the diver with the addition of supplemen¬
tal oxygen, maintaining an appropriate level for
metabolism. The first rebreathers were pure oxygen
rebreathers which could only be used at very shallow
depths, but allowed much longer, bubble-free opera¬
tion for divers.
As interest grew in rebreathers and applications
demanded technology for deeper depths, SCRs were
developed. SCRs create some intermittent bubbles.
They inject a steady stream of oxygen-enriched air
(nitrox) into a set of breathing bags. As the diver's
exhaled breath is recaptured in the system, a few bub¬
bles are vented off and fresh nitrox is injected to make
up the difference. The equipment is carefully cali¬
brated so that the injection stream makes up the
correct volume of oxygen lost through bubbles and
diver metabolism. SCRs act as gas extenders and pro¬
vide longer no-decompression times over standard-air
diving, but have the disadvantage of venting gas in
the form of bubbles. Various advances over the years
enhanced this technology to reduce wastage through
bubbling and increase monitoring capability.
Fully closed circuit rebreathers offer the greatest
advantages over standard scuba. A CCR only vents
gas on ascent. It has the highest efficiency and nets
divers the longest no-decompression times. This tech¬
nology can be up to 200 times more efficient than stan¬
dard open-circuit scuba, dependent on depth, the ratio
of efficiency improving at deeper ranges. CCRs are
now developing into true life support devices that bet¬
ter protect the diver from human error and undetected
dangers such as carbon dioxide buildup. The Sentinel
rebreather from VR Technology brings to the market
the first application of an advanced onboard carbon
dioxide monitoring system. The three-pronged
approach monitors the volume of oxygen that passes
through the solenoid valve estimating the production
of carbon dioxide, tracks the thermal profile of the
reaction zone within the bed of scrubber material,
and monitors gaseous carbon dioxide within the
breathing loop using infrared technology, netting a
reasonable expectation of scrubber life as well as
detecting errors in preparation of the scrubber or fail¬
ures of the system (Gurr, 2010).
Carbon dioxide is thought to be the least under¬
stood and greatest risk to rebreather divers. Difficult to
track and impossible to detect at autopsy, carbon diox¬
ide poisoning may turn out to be the silent killer that
has taken the lives of many dozens of rebreather
divers in the past decade. Professor Simon Mitchell
from the Department of Anaesthesiology at the
University of Auckland and Auckland City Hospital,
leads the industry in researching the issues of carbon
dioxide buildup and retention in divers. His paper.
Hyperbaric Conditions, gives a comprehensive account
of relevant physiology (Doolette and Mitchell, 2011).
Mitchell describes that the act of diving itself alters
respiratory control. He describes that sensitivity to
CO 2 may be reduced by an increased partial pressure
of oxygen, increased partial pressure of nitrogen, and
increased work of breathing—all aspects of rebreather
diving. In particular, he believes that an increased
work of breathing results in varying degrees of CO 2
tolerance which may subsequently lead to the diver
failing to recognize and correct a potentially life-threat¬
ening scenario (Mitchell, 2010).
Cave-diving fatalities have declined in the past 30
years, yet rebreather diving deaths are on the rise. The
industry is at its pioneering stage where instructors
and divers are working out common codes of practice.
ENCYCLOPEDIA OF CAVES
314
E
manufacturers are unregulated in many regions, and
equipment testing is sometimes undertaken with
human divers using gear that has no independent
oversight. In ever more litigious times, it will be
incumbent on the industry to self-police through the
adoption of policies such as CE or equivalent third-
party testing and documentation. Barring this, product
liability lawsuits and outright banning of the use of
such equipment at dive sites seem likely. In the dawn¬
ing days of this exciting technology, it is clear we still
have a long way to go to improve procedures and
equipment safety. (CE EN14143: Essentially any person
or company in Europe that is importing, manufactur¬
ing, selling, or using rebreathers requires a valid CE
marking on the equipment. If it does not have a valid
CE marking, then they are breaking the laws or regula¬
tions in their country, and the enforcement body in
that country can ban the import, obtain court injunc¬
tions to stop manufacture or sale, and seize the goods.
If there are accidents from the use of such equipment,
it may become a criminal liability for the directors of
the company involved. EN14143 is a testing and docu¬
mentation protocol for all PP02 monitors for use with
a rebreather that is imported, manufactured, sold, or
used in work in Europe.)
In addition to the diving apparatus itself, the tools
for cave survey are also undergoing a renaissance.
Traditional underwater cave survey is conducted using
knotted survey line or fiberglass tape and compass.
These traditional tools can only net extremely rudi¬
mentary surveys that are dependent on myriad imagi¬
native additions by true artists of cartography.
Underwater cave maps take years to assemble in even
a basic form and are often fraught with errors and gen¬
eralities. Accurate digital compasses, such as those
provided on Suunto diving computers, add signifi¬
cantly to accuracy, but the future lies in far more accu¬
rate tools.
After proving the value of the Wakulla 3D Mapper
in the late 1990s, Dr. Bill Stone went back to the draw¬
ing table to make the next huge leap in exploration
technology. In 2007, DepthX made a historic dive in La
Pilita cenote in Tamaulipas, Mexico, accurately map¬
ping the sink in three dimensions, autonomously. It
appears that the real future of cave mapping may not
include humans. The untethered device dropped to a
maximum depth of 100 m while compiling 340,000
sonar wall hits during the course of an approximately
1-hour-duration mission. The device's sonar arrays
"scanned" the wall in 360° and even noted the pres¬
ence of a 20-m-diameter tunnel leading off from the
western wall of the chamber at a depth of approxi¬
mately 40 m. As Stone continues to miniaturize this
apparatus, it appears the future of underwater cave
mapping will not only be more accurate but also safer.
Numerous challenges and rewards await the next
generation of cave-diving explorers. Physical and tech¬
nological advances are increasing the depths and pene¬
trations that are humanly possible. The coming era
may feature exciting new technologies from hard suits
to miniature robots to thermal imaging and new
decompression strategies. With education, mentoring,
and experience, there are few limitations that stand
before the imagination of those willing to push the lim¬
its of underwater cave exploration.
Bibliography
Bozanic, J., & Halpern, R. (1974). Cave diving fatalities. A summary.
Workshop presentation for the NSS-CDS Annual Conference,
Marianna, Florida.
Doolette, D. J., & Mitchell, S. (2011). Hyperbaric conditions.
Comparative Physiology, 11—39.
Exley, S. (1986). Basic cave diving: A blueprint for survival (5th ed.).
Huntsville, AL: National Speleological Society.
Exley, S. (1994). Caverns measureless to man. St. Louis, MO: Cave
Books.
Gurr, K. (2010). Technical diving from the bottom up. VR Technology.
Heinerth, J. (2010). The essentials of cave diving. High Springs, FL:
Heinerth Productions, Inc.
Heinerth, J., & Oigarden, W. (2008). Cave diving: Articles and opinions.
High Springs, FL: Heinerth Productions, Inc.
Kakuk, B., & Heinerth, J. (2010). Side mount profiles. High Springs, FL:
Heinerth Productions, Inc.
Kernagis, D., McKinlay, C., & Kincaid, T. (2008). Dive logistics of the
Turner to Wakulla Cave Traverse. In P. Brueggeman & N. W.
Pollock (Eds.), Diving for science, Proceedings of the American
Academy of Underwater Sciences 27th Symposium. Dauphin
Island, AL.
Lewis, S. (2011). The six skills and other discussions: Creative solutions
for technical divers. Techdiver Publishing.
Mitchell, S. (2010). CCL: Big, bad and hard to measure. Presentation at
Eurotek, October 2010, Birmingham, England.
Stone, W. (2007). DEPTHX: Zacaton—Mission 1 begins. <www.stone
aerospace.com/neivs-/news-zacaton-mission4.php/>.
EXPLORATION OF CAVES—VERTICAL
CAVING TECHNIQUES
Mark Minton and Yvonne Droms
U.S. Deep Caving Team
Disclaimer: Vertical caving is a potentially hazardous activity.
The following discussion is not intended to teach proper verti¬
cal technique or to cover all details of vertical caving. Get
proper training before entering any vertical cave.
INTRODUCTION
Although many caves can be explored by simply
crawling, walking, and climbing over rocks and
ledges, a great many require the use of ropes and
Encyclopedia of Caves.
© 2012 Elsevier Inc. All rights reserved.
EXPLORATION OF CAVES—VERTICAL CAVING TECHNIQUES
315
related equipment for ascending and descending verti¬
cal drops. Even caves that are predominantly horizon¬
tal may have occasional pits, while others are almost
entirely vertical with dozens of drops and relatively lit¬
tle horizontal extent. Individual pits can range from a
few meters to hundreds of meters in depth. The
world's deepest cave (Krubera) is over 2000 meters
deep and contains scores of drops.
The simplest vertical cave passages can be negoti¬
ated without equipment by chimneying and other free-
climbing techniques, although this can be dangerous
since there is no safety net if one slips or if handholds
break. For slightly more difficult drops one may get by
with a simple handline or a ladder. However, serious
vertical caves require specialized equipment, training,
and practice to explore safely. Ropes, and the associ¬
ated hardware to use them, are the only practical
method for getting up and down the vast majority of
vertical caves. The current state of the art is called sin¬
gle rope technique, or SRT. We shall confine our dis¬
cussion to the most commonly used variants of SRT,
although there are many different systems with ardent
adherents and specialized applications.
ROPE
Modern ropes for caving are made of nylon or
polyester and use kernmantle construction, which con¬
sists of a core of long, multistranded, twisted fibers sur¬
rounded by a tightly woven sheath. Unlike dynamic
ropes for rock climbing, which are designed to stretch in
order to absorb the shock of a fall, static caving ropes are
designed to have as little stretch as possible. This contri¬
butes to the efficiency of climbing the rope, especially on
long drops where stretch can be significant. For many
years the standard diameter for caving rope was 11 mm,
and this is still widely used in the United States.
However, nowadays many cavers are opting for smaller
diameters: 10-mm rope is common and advanced explo¬
ration may use 9-mm or even 8-mm rope. These smaller-
diameter ropes are lighter and large amounts can be
packed into a relatively small space for transport, but
one must be increasingly careful to avoid abrasion when
rigging them.
RIGGING
In the early days of vertical caving it was common
practice to simply tie the rope to whatever was handy
and throw it down the pit. This was easy to do but
often resulted in dangerous abrasion of the rope as
well as the possibility that loose rocks could be
dislodged from the wall. It also frequently made it
difficult to get over the lip at the top of the drop.
Especially when using small-diameter rope, these con¬
ditions are unacceptable. Rig points should be chosen
such that the rope hangs as freely as possible down
the pit. Since this requires the rope to hang at least a
short distance beyond the edge of the drop, an
approach line is usually added for safety in getting
onto and off of the main rope.
Natural rig points (bedrock bridges, large stalag¬
mites, firmly wedged boulders, etc.) are preferable
from a conservation point of view, but more often than
not there is no suitable natural anchor available. In
that case artificial anchors consisting of rock-climbing
chocks, pitons, or, most often, expansion bolts are
used. With the advent of lightweight battery-powered
hammer drills, placing bolts has become very easy.
One should use care to place bolts properly since there
are usually limited spots available, and no one likes to
see a proliferation of bolts at the top of a pit. When
bolts are in high-traffic areas they should be made of
stainless steel and ideally labeled in a manner that lets
future visitors know the embedment depth, date of
placement, and identity of the rigger. Another alterna¬
tive is the use of glue-in bolts, which are more perma¬
nent and have less tendency to work loose with
repeated use. All rig points should be backed up. Do
not depend on a single bolt, piton, or chock. An addi¬
tional advantage of using multiple rig points is that
one can almost always construct a free-hanging rig by
using Y-hangs (Fig. 1).
As important as a good rig point is a secure knot to
tie off the rope. There are several knots that vertical
FIGURE f Y-hang with rebelay. Sketch by Bob Alderson.
ENCYCLOPEDIA OF CAVES
316
E
cavers should know. These include the bowline,
bowline on a bight, figure eight, double figure eight or
rabbit-ears knot, figure nine (used with thinner rope),
alpine butterfly, and double fisherman's knot. The
double fisherman's knot is used to join two ropes
together and is relatively easy to pass. One should also
be familiar with the water knot for nylon webbing.
There are a great many other knots that may come into
play in various situations, but those mentioned here
should serve in the majority of situations.
Even when the rope is placed properly at the top of
a drop, there may be ledges, speleothems, waterfalls,
and so on, that must be avoided farther down. In this
case the rope may be secured at intermediate spots to
eliminate rub points or change its orientation. Such
intermediate rig points are called rebelays and redirects
or deviations. A rebelay is a hard tie-off, where the rope
is secured to an anchor similar to one used at the top of
a drop (Fig. 1). This essentially divides the drop into
two (or more) independent pitches, which, depending
on circumstances, may be climbed simultaneously by
two or more cavers. A redirect or deviation is not an
actual anchor, but simply a rerouting of the rope
through a carabiner tethered on a piece of cord or nylon
webbing such that it pulls the main rope off of its origi¬
nal path and clear of obstacles. Both of these techniques
require additional skill and training to negotiate safely.
Although rarely necessary with careful rigging,
occasional rub points are unavoidable or impractical to
eliminate. In such cases rope pads may be used. Any
relatively abrasion-proof, flexible material may be
used to protect the rope from rubbing against the rock.
This includes pieces of carpet, heavy cloth, wide nylon
webbing, slit rubber hose, or, temporarily, even cave
packs or items of clothing. Rope pads are easiest to
pass if they are tied off independently and placed such
that the rope will simply run across them. When this is
not possible an alternative is to tether the pad directly
to the rope. This is a more secure arrangement in that
the rope cannot slip off of the pad, but it is more diffi¬
cult to pass.
SIT HARNESS
For all vertical work in caves, one should wear a
properly fitting sit harness designed specifically for
this purpose (Fig. 2A). The harness holds the body in a
comfortable and secure upright position on rope even
when there are no walls or floor within reach. A typi¬
cal sit harness is made of wide nylon webbing and
consists of a waist belt and two leg loops. It is closed
by a lockable metal carabiner or quick link, often called
a maillon, which provides a central attachment point to
which various pieces of hardware may be connected.
Frequently there are auxiliary gear loops for stowing
vertical gear when it is not in use and for carrying
other pieces of equipment as well. A harness designed
for caving typically has a lower central attachment
FIGURE 2 The frog system for ascending rope. (A) Frontal view showing harness. (B) Frontal view, climbing. (C) Side view, climbing.
Sketches by Bob Alderson.
ENCYCLOPEDIA OF CAVES
EXPLORATION OF CAVES—VERTICAL CAVING TECHNIQUES
317
point than one for rock climbing, as this optimizes
ascending efficiency.
While on rope it is more efficient to suspend a cave
pack from the central link of the sit harness with a
tether than to wear it on one's back. This prevents the
weight of the pack from pulling the caver away from
the rope. It also supports the weight of the pack when
one is resting. The tether should be just long enough
not to interfere with climbing, but not so long as to
allow the pack to get caught or entangled beyond the
caver's reach.
DESCENDING
While cable ladders may occasionally be used to nego¬
tiate vertical drops in caves, by far the most common
method for going down pits is rappelling (also called abseil¬
ing) on a single rope. Many devices have been designed
that apply friction to the rope in a variable and controlla¬
ble fashion. Of these, the only ones in common use for
caving today are the rack, bobbin, and figure eight. An
older technique now rarely used is body rappelling,
where the rope is wrapped around the caver's body to
provide the friction necessary to control descent. Body
rappelling is less comfortable and more dangerous than
using a mechanical descender because the caver is not
physically attached to the rope. For safety in approaching
a drop, see the discussion of cow's tails below.
The figure eight is the simplest mechanical
descender and consists of a stiff metal frame roughly in
the shape of the numeral 8. The rope is threaded
through the frame to provide friction. Disadvantages to
the eight are that it gives only moderate control and
must be removed from the harness in order to be
attached to the rope, which could allow it to be
dropped. It also applies a twist to the rope which many
cavers do not like. The figure eight is only suitable for
relatively short, nontechnical drops.
Rappel racks consist of a steel frame, basically in the
shape of the letter U, which is attached to the central
link of the sit harness. A quick link as opposed to a
locking carabiner should always be used to attach a
rack to the sit harness because a carabiner is far more
prone to accidental unlocking. Perpendicular to the
long axis of the U are mounted a number of steel or
aluminum brake bars, which may either be fixed or
pivot to open and close onto the other side of the rack.
The rope is threaded under and over these bars to pro¬
vide friction. An advantage of the rack is that friction
can be varied based on both the number of bars in use
and their spacing. Bars may be added or removed in
the course of a rappel if necessary. A rack can be
locked off by wrapping the rope around it so that the
descent stops and the caver may use both hands for
other tasks. Some cavers do not like aluminum brake
bars because they wear more quickly and leave a gray
residue on the rope.
A bobbin consists of two metal plates which pivot
open to reveal two metal spools or bobbins around
which the rope is threaded to provide friction. Unlike
a rack, a bobbin should always be attached to the sit
harness link with a locking carabiner rather than a
quick link, because the latter, due to its smaller diame¬
ter, can sometimes inadvertently pry open the safety
latch. A separate carabiner on the main link is used to
help control friction. Auto-lock variants of the bobbin
also have a safety lever which must be held in a spe¬
cific position in order to proceed. If the lever is
released, and also in some cases if it is squeezed,
descent will be halted, for example, in the event that
the rappeller becomes unconscious. It is more difficult
to change the amount of friction in the course of a rap¬
pel with a bobbin than it is with a rack, although if
one knows a rope will be slow the bobbin can be
rigged differently at the beginning of the rappel to pro¬
vide less friction. A bobbin can also be locked off by
wrapping the rope around it. Bobbins work better on
10-mm or smaller diameter ropes.
ASCENDING
Going up a rope requires considerably more effort
than going down, since one is fighting gravity. In addi¬
tion, cavers more often ascend ropes at the end of a
caving trip when energy reserves are already lower.
Keeping this in mind is important so that fatigue does
not lead to a dangerous situation. All mechanical
ascending devices employ some sort of one-way cam
that will slide up the rope freely and then lock and not
slide back down when weighted. Proper arrangement
and fitting of ascending devices and good, well-prac¬
ticed technique are essential to efficient rope-climbing.
A great number of different systems have been devel¬
oped for ascending ropes, but we will cover only the
most universal, the frog systein. Frog is the world stan¬
dard, used almost exclusively everywhere except in
parts of the United States. One reason for its popular¬
ity is ease of crossing rebelays. Frog is less efficient on
long free drops where the rope walker system may be
preferred, but for general vertical caving it is ideal.
A basic frog system consists of a chest ascender, an
upper ascender, and cow's tails (Fig. 2A). The chest
ascender, which actually rides below the sternum, is
designed to lie flat against the caver's belly and is
attached directly to the sit harness at its central link. It
is held up by a chest harness, whose only purpose is
to hold the chest ascender in a vertical position; it does
not support the climber. The chest harness may thus
ENCYCLOPEDIA OF CAVES
318
E
be made of very light webbing or cord since it does
not bear any weight, although it should be adjustable for
comfort when not in use.
The upper ascender is fitted with a long piece of
cord or nylon webbing with one or two foot loops at
the end to stand in. For safety it is attached by a short¬
er piece of webbing to the central link of the sit har¬
ness. Some cavers use the longer cow's tail (see below)
for this, but that makes for a less versatile system over¬
all. The lengths of these attachments are critical to effi¬
cient climbing and will vary with the height and
technique of the climber. In use, the upper ascender is
placed on the rope above the chest ascender and typi¬
cally has a handle for easy gripping, although lighter
and more compact versions without a handle are pre¬
ferred by some cavers.
To climb a rope, the caver sits so that his/her
weight is held by the chest ascender. The feet are
tucked under the buttocks and the upper ascender is
simultaneously slid up the rope as high as possible.
The caver then stands up in the foot loop(s). While
doing this the chest ascender will automatically slide
upward. One then repeats this sit-stand process.
Although simple in principle, it takes coordination and
practice to climb efficiently (Fig. 2B,C).
Cow's tails are an essential safety element in any
modern vertical system. They are used to attach a
caver to a safety line or directly to an anchor point
while still allowing some freedom of movement in
order to perform the various maneuvers needed to
gain access to rig points and to cross rebelays. Cow's
tails are best constructed from knotted lengths of
dynamic rope; those made from sewn webbing have
been shown to produce significantly higher shock
loads in the event of a fall. Typical cow's tails are in
the form of a V with one side longer than the other for
versatility. A nonlocking carabiner is attached to each
arm of the V using a knot that will hold it stiffly in
place. A knot at the bottom of the V is attached to the
central sit-harness link. In addition to their use in verti¬
cal work, cow's tails make handy tethers for other
items that might be dropped or lost.
For longer climbs, especially if one is carrying a
heavy load, a small additional ascender may be attached
directly to one foot. The Petzl Pantin is made espe¬
cially for this purpose. Flaving an ascender directly on
the foot helps hold the climber's body in a more
upright position. It also provides an additional foot¬
hold for getting past rebelays and pitch heads, espe¬
cially if the passage is constricted.
One additional type of "ascender" that is rarely used
but can come in handy in certain situations is the Prusik
knot. This is a simple knot tied with a loop of smaller-
diameter rope or cord that will slide up the main rope
when unweighted but will tighten and hold fast when
weighted. A set of knots is very light and compact and
could be used to climb a short drop in a remote location
where it might not be worthwhile to carry normal climb¬
ing gear. In an emergency, Prusik knots can also be
fashioned by cutting off the bottom of the main rope.
CHANGING OVER
For various reasons one may need to change direc¬
tion of travel while on rope. It is possible to change
from descending to ascending or vice versa, although
the former is easier than the latter. When descending it
is a relatively simple matter to attach one's ascenders
and transfer weight to them. The descending device
can then be removed and the caver can climb up the
rope. When ascending it is possible to change direction
by simply moving one's ascenders in reverse. By
briefly holding open the cams and alternately sliding
the ascenders downward, one can walk down the
rope, but this is a slow and inefficient process not
suitable for long distances. It is more complicated to
successfully transfer one's weight from ascenders to a
descending device and then be able to remove all of
the ascenders. These techniques should be practiced
on the surface where provisions have been made to
lower the climber in the event of a problem.
A similar situation occurs when one must pass a
knot in the rope. This may occur either because the
original rope was not long enough and an additional
piece was tied on, or because the rope was damaged
and the bad section was isolated by a knot. In either
event, a vertical caver must be prepared to pass an
unexpected obstacle on rope while going either up or
down.
TRAVERSES AND TYROLEANS
Occasionally one needs to cross a void horizontally
rather than vertically. This can occur when a passage
continues on the opposite side of a pit and it is more
convenient to go straight across rather than down and
back up, or when it is desirable to avoid deep water. If
there is a small ledge to provide footing, a rope may
be rigged a short distance above the ledge and tied off
periodically so that there is never much slack between
anchor points. One may then cross the ledge safely by
clipping one or both cow's tails onto the rope and
walking across, using the rope for support. If there is
some elevation gain, an ascender may also be used to
provide an easier grip. At each anchor, one cow's tail
at a time is moved past the knot so that the caver
remains securely attached at all times. Traverses may
also be done when there is no ledge to walk on.
ENCYCLOPEDIA OF CAVES
EXPLORATION OF CAVES—VERTICAL CAVING TECHNIQUES
319
although with considerably more difficulty. In such
cases a separate rope may be rigged for the feet to pro¬
vide support and balance.
If the space one needs to cross is not near a wall,
then a rope may be rigged directly across the void.
This is called a Tyrolean or Tyrolean traverse (Fig. 3). In
this case the rope should be as tight as possible to min¬
imize sag. Such rigging places extra force on the
anchor points and on the rope and should be done
carefully with this in mind. Many times a Tyrolean
will be slightly inclined, in which case it is easy to
slide down in one direction but is significantly more
difficult going up the other way. Even an essentially
horizontal Tyrolean will inevitably have some sag, so
at some point one must travel uphill. To cross a
Tyrolean, the caver clips in a cow's tail and upper
ascender (without using the foot loops) as well as a
carabiner at the waist. The caver hangs underneath the
rope and may throw one or both legs over the rope for
support. Using the ascender, one pulls along the rope
to the other side. On a long or steep Tyrolean, a chest
and/or foot ascender may be used as well. If the
Tyrolean is highly inclined, a second rope may be
rigged more loosely beneath the first. On descent, a
rappel device may be used on the lower rope while a
cow's tail is used for support on the upper one.
An alternative to a Tyrolean or even some wall
traverses is to rig a rope loosely across the void in a
large U shape. The caver rappels down one side of the
rope until the other side is within reach. Ascenders are
then attached to the rope going up, and the rack is
removed as soon as it is unweighted. One then
climbs in the usual fashion. Even though this requires
more rope and more distance traveled, it is sometimes
easier and more efficient than doing a traverse or a
Tyrolean.
VERTICAL FROM THE BOTTOM UP
The majority of vertical caves are rigged from
the top down. In other words, one enters the cave and
descends various drops to get to the bottom. This is
FIGURE 3 The Tyrolean traverse. Sketch by Bob Alderson.
certainly the easiest approach, because pitches may be
rigged with the aid of gravity. However, occasionally
one encounters a vertical drop from the bottom with
no ready access to the top. In this case more involved
procedures are required to get up the drop. A few
very deep caves have been explored almost exclusively
from the bottom up. The most famous is Austria's
Lamprechtsofen, which was explored to a depth
(height) of over 1000 meters before an upper entrance
was connected.
Many techniques have been used to climb vertical
drops from the bottom, some quite ingenious. The sim¬
plest drops may be scaled by free climbing or chim¬
neying. Sometimes one caver can boost another up
into a high lead. When the climb is relatively short, a
ladder can be brought in to gain access. A related tech¬
nique involves a scaling pole, where a segmented
metal pole is carried into the cave and assembled at
the climb. A rope or cable ladder is suspended from
the top of the pole and climbed to gain access to the
upper passage. If there are appropriately located pro¬
jections or formations, a rope may be used to lasso an
anchor or a grappling hook can be used. In a few cases
helium balloons have been used to drape a line over
an anchor so that a rope could be pulled up. These
techniques are rather risky because one cannot exam¬
ine the anchor to see if it is strong enough or if the
rope is securely in place.
The most reliable method of reaching a high lead
from the bottom is bolt climbing. This used to be a
laborious process using hand-driven expansion bolts,
but with modern, lightweight hammer drills it has
become almost routine, so much so that now any
dome can be considered a viable lead. To bolt up a
drop, one sets an expansion bolt in the wall from
which is suspended a short section of webbing ladder
called an etrier. An adjustable daisy chain is used to
pull the climber's waist up to the bolt and keep him/
her upright while climbing the etrier. A quickdraw is
also attached to the bolt, through which runs a
dynamic belay rope, much the same as in rock
climbing. A separate person controls the belay, which
will catch the climber in the event that a bolt fails.
By repeating this process with another etrier, daisy
chain, and quickdraw the climber advances up the
wall, alternately moving the etriers and daisy chains
upward. Domes over 100 meters high have been
climbed by this method, and much passage has
been discovered that would never have been seen
otherwise.
To retreat from a climb that does not go or is no lon¬
ger needed, one may either downclimb the drop in
reverse, retrieving protective gear along the way, or
rig a pull-down, which usually involves leaving some
sort of anchor behind. The former is more efficient in
ENCYCLOPEDIA OF CAVES
320
E
terms of hardware, but is more time-consuming. In a
pull-down, the main rope is doubled through the top
anchor and the caver rappels on one side with the
other end tied off or blocked, or on both of the ropes
simultaneously. Once on the bottom of the drop, the
caver pulls one end of the doubled rope through the
anchor until the rope falls to the ground. Some sort of
hardware at the top of the drop is generally sacrificed
in this case.
Bibliography
Judson, D. (1995). Caving practice and equipment. British Cave
Research Association and Birmingham, Alabama: Menasha Ridge
Press.
Marbach, G., & Tourte, B. (2002). Alpine caving techniques. Allschwil,
Switzerland: Speleo Projects.
Smith, B., & Padgett, A. (1996). On rope. Huntsville, Alabama:
National Speleological Society.
Warild, A. (2007). Vertical. Newtown, Australia: Alan Warild.
ENCYCLOPEDIA OF CAVES
FOLKLORE, MYTH, AND
LEGEND, CAVES IN
Paul Jay Steward
Cave Research Foundation
Since the beginning of time man has fantasized about
caves and the mysteries they hold. They are a doorway
to the Underworld and have played a vital role in the
narrative of cultures throughout the globe. Many of the
stories associated with caves conjure up images of a dark
and mysterious place inhabited by gods and demons
and associated with everything from resurrection, fertil¬
ity, worship, sacrifices, and the beginning of life, to a
resting place for the dead. Others envision the Earth as
hollow, teeming with lost worlds and civilizations.
The most popular of these stories are the ones
handed down from ancient mythology. The Greeks
believed the Underworld to be the kingdom of the dead
ruled by Hades, a greedy god whose only concern was
increasing the number of souls in his dark world. In
Northern Ireland there is a cave called Saint Patrick's
Purgatory. It is used as a place of worship and is
believed to be the entrance to Hell.
Although it was a one-way trip for most, getting to
the Underworld was not an easy chore. Between the
world of the living and the dead, five rivers must be
crossed: the Acheron (river of woe), the Cocytus (river
of lamentation), the Phlegethon (river of fire), the Styx
(river of hate, which surrounds the Underworld with
nine loops), and the Lethe (river of forgetfulness).
Hermes, the god of travelers, was born in a cave on
Mount Cyllene. He was quite familiar with the trail and
would lead the souls of the dead through the darkness.
Erebus was the god of darkness in the Underworld.
The last river to cross was the Acheron. Charon would
lead you across Acheron in a ferry, provided you were
buried with a coin placed under your tongue for pay¬
ment. Those who could not pay were trapped between
the two worlds forever.
Across the river loomed the gates to the Underworld.
Cerberus, a three-headed dog with a dragon tail,
guarded the gates and would allow the souls to enter
but never leave. Once through the gates, the souls
appeared before a tribunal consisting of Hades, Aeacus,
Minos, and Rhadamanthys. Souls who had committed
crimes against the gods were sent to Tartarus, while the
souls of the just were sent to Elysium Fields.
Many of the caves located on the island of Crete are
associated with mythology. Zeus, the king of the gods,
and brother to Hades, was born in a cave located on
Mount Aegeum. Today it is called the Cave of
Psychro. Amnisos Cave is also located on Crete and is
believed to be the sanctuary of Artemis, the goddess of
fertility. Aeolos, the god of wind and air, kept his
winds locked in a cave and only let them out as
instructed by the gods of Olympus. Somnus, the god
of sleep, resided in a cave. The twins, Romulus and
Remus, were found in a basket along the banks of a
river and brought to the Cave of Lupercal where they
were suckled and raised by a she-wolf. Many of the
gods from the oceans lived in caves and grottoes. Even
the goddess of music loved to play her instruments in
a cave.
The Underworld was not just reserved for gods and
souls of the dead. Cunning little elves and dwarfs
spent much of their time living and hiding in caves.
Harpies were filthy, winged monsters, with the faces
of women and the bodies of vultures, that lived in
caves. Hydra, the many-headed serpent, and Cacus, a
fire-breathing troll, lived in caves. Polyphemus, a
Cyclops, lived in a cave near Mount Aetna and sur¬
vived on meals of human flesh. The serpent Python
dwelled in a cave. Typhon, the scourge of mankind,
lived in Kilikian Cave, and the famed Medusa lived in
the secret Cave of the Gorgons.
Sibyl's Cave is located in the ancient town of
Cumae, near Naples, and was the home of Sibyl, a
famous female prophet. Many of her prophecies were
Encyclopedia of Caves.
321
© 2012 Elsevier Inc. All rights reserved.
322
F
written on leaves, which she placed at the mouth of
her cave for people to find. These "Sibylline Leaves"
were bound into books and consulted by the Romans
in times of crisis. It is also through this cave that Sibyl
was said to have led Aeneas to the Underworld.
Makua Cave, also known as Kaneana Cave, is
located on the western shore of the Hawaiian Island of
Oahu. The cave is named after the Hawaiian god
Kane, the god of creation. It is believed that the cave is
the womb from which mankind emerged and spread
throughout the Waianae Coast. In ancient times
Kahunas were forbidden to enter the cave for they
believed it was the home of Nanaue, the shark man of
Kaneana. It was said this half-man, half-shark beast
would drag his victims to the back of the cave and
devour them.
Located on the high slopes of Mount Amarnath, deep
in the Himalayan Mountains, lies Amarnath Cave. This
is one of the holiest sites in the Hindu religion. The cave
contains several ice stalagmites that grow from a natural
spring. The largest of these is believed to be a representa¬
tion of the Hindu god Shiva. The smaller ones represent
the Hindu gods Ganesha, Parvati, and Bhairava.
Every year during the months of July and August,
tens of thousands of pilgrims bear the harsh climate and
rugged terrain to trek to the cave to worship these phal¬
lus-shaped formations. According to the legend, in the
cave Shiva disclosed the secrets of creation and immor¬
tality to his companion, Parvati, and those who make the
difficult journey to the cave are promised salvation. Also
in the cave with Parvati was a pair of mating doves that
overheard the conversation and learned the secrets. The
doves made the cave their eternal home and have been
reborn every year since. Many pilgrims report seeing a
lone pair of doves during their trek.
To the Maya, caves and cenotes were sacred places
used for rituals and sacrifice. They called their lower
world Xibalba, or Place of Phantoms. Many of their
temples were built over cave openings, and much of
their water came from the caves. The Maya believed
they were born from the Earth goddess and originated
out of seven caves.
Caves are mentioned throughout the bible. They are
used for shelters, as tombs to bury the dead, as places
to hide from one's enemies, and a place for quiet
worship.
Lot and his two daughters left the city of Zoar and
found shelter in a mountain cave. Abraham purchased
Machpelah Cave, also called Cave of the Patriarchs,
from Ephron and used it as a family tomb. This cave is
one of the holiest sites known in the Jewish religion.
The Patriarchs Abraham, Isaac, and Jacob, along with
the Matriarchs Sarah, Rebecca, and Leah, are buried in
the cave. It is also believed the cave is a passageway to
the Garden of Eden.
Seeing their armies defeated by Joshua, the five
kings of the Amorites fled to Makke'dah Cave. After
being captured and killed, their bodies were returned
to the cave and the cave sealed with large stones.
David found refuge in Adullam Cave after fleeing
from Saul. David wrote many of the Psalms in this
cave. Also in this immense cavern, David trained his
outlaw army of 400 men. In Engedi Cave, David found
Saul asleep and cut off a part of his robe. Obadiah hid
Elijah and 100 prophets in caves on Mount Carmel.
Caves were frequently used as stables that also con¬
tained rooms for people to sleep. Many believe that
Jesus was born in one of these caves. After dying on
the cross, Jesus was brought to a cave and buried.
Several historical books written by Egyptians several
hundred years before the birth of Christ contain stories
about the lives of Adam and Eve. It is believed that after
being banned from the Garden of Eden, God sent
Adam and Eve to live in a strange broad land, covered
in sand, strewn with stones, and void of all vegetation.
Here, God commanded them to live in the Cave of
Treasures. The cave became a family shrine and many
generations would live, die, and be buried in there.
Another cave associated with Adam and Eve is
Lilith Cave. According to Jewish folklore, Adam's first
wife was Lilith, not Eve, as commonly accepted.
Unfortunately, it was not a match made in heaven.
Adam and Lilith fought and argued constantly. Any
attempts by Adam to mate with Lilith were met with
rejection. Lilith refused to lie under Adam in the stan¬
dard missionary position. She felt they were equal and
therefore wanted to make love in equal positions.
Adam wanted no part of this arrangement. In a rage,
Lilith left the Garden and took up residence in a cave
along the shores of the Red Sea in an area fraught with
demons. In the cave, she made love to many demons
and populated the world with thousands of demon
children. She became known as the Mother of Demons,
and is said to still reside in her cave.
Ti-Tsang Wang is the Chinese god of mercy. He
wanders through the caverns of Hell searching for
souls to save and help escape from the Underworld.
Amaterasu is the Japanese Sun goddess. One day
her mischievous brother, Susanoo, destroyed her beau¬
tiful garden. Amaterasu was so mad she hid herself in
a cave and covered the entrance with a large rock. This
made the world dark and cold. For many weeks she
hid in the cave. One day she heard laughter, music,
and singing coming from outside her cave. Amaterasu
was curious as to how all the other gods could sound
so happy with the world so dark and cold. They told
Amaterasu another god more beautiful than she had
come and brought light to the world. As she peered
out from the cave she saw her reflection in a mirror.
Quickly, she was pulled out from the cave and told
ENCYCLOPEDIA OF CAVES
FOOD SOURCES
323
that the light was her own reflection. Amaterasu had
never seen herself before and was amazed at her own
beauty. Afterward she returned to her throne in the
heavens to warm and light the Earth.
No talk of caves would be complete without a few
ghost stories. The quiet, isolation, and absolute dark¬
ness found underground create the perfect setting for
imaginations to run wild. Many of the caves across the
world have horrible tales of ghosts tormenting unsus¬
pecting visitors. One of the oldest and most document¬
ed cases in the annals of American haunted places is
the Bell Witch Cave.
The cave is along the south side of the Red River, in
the rural farming town of Adams, Tennessee. Above
the cave is a known burial ground for the Choctaw
Indians. While many believe it is Indian ghosts that
haunt the cave, others believe it is a troubled spirit
known as the Bell Witch.
Numerous stories abound of unusual happenings
in the cave. Researchers have discovered high energy
levels and evidence of paranormal activity in and
around the cave. Visitors have reported strange
sounds, eerie feelings, and even sightings of appari¬
tions. In the cave, camera equipment is known to
jam, stop working, and rewind for no apparent
reason. Photographs taken in the cave sometimes
contain white or orange glowing orbs floating about.
Orbs are a common occurrence in locations associ¬
ated with paranormal activity. Others have reported
seeing a dark-haired woman drifting throughout the
cave. The current owners of Bell Witch Cave allow
tours from May 1 to October 31. They, too, tell of
strange sounds and happenings in the cave. It is not
unusual for a tour to be cut short by a disturbed
guest wanting to return to the surface and vowing
never to return.
FOOD SOURCES
Thomas L. Poulson
Jupiter, Florida
USE OF THE SCIENTIFIC METHOD,
CAVEATS, AND TRADEOFFS
This article uses the scientific method to hypothesize
that organisms in caves are food-limited and the
adaptations of different species to test predictions
about the degree of their limitation by food supply.
There are a few cave types where organisms are not
very food-limited and these are exceptions that prove
the rule. The species in these food-rich caves show few
or no adaptations to food scarcity. At one extreme
are caves with abundant food from organic pollution,
sulfur-based chemoautotrophy, natural deposits of
fossil fuel, or extensive bat guano. At an intermediate
scale are areas with nonsoluble ridges above limestone
valleys where streams sink at the contact and can
carry large amounts of decomposing plant material
into caves. At the other extreme are caves in glaciated
areas where food supply is variable, but there are few
or no cave-adapted species. If they had been present,
most became extinct when glaciers covered the caves.
Trade-offs for supply and kinds of food in caves are
examined. These include quantity versus quality,
quantity versus risk of injury, and quality versus
unpredictability. In addition, predictability of food
influences whether a species can be a specialist or
generalist (another trade-off), and in terrestrial habitats
the usability of a food type depends on the moisture of
the substrate.
LIMITED FOOD SUPPLY IN MOST CAVES
Observations and First Principles
The absence of light in caves leads to the hypothesis
of food limitation; no light means no photosynthesis.
During growing seasons the above ground world is
overwhelmingly green; the amount of all animal life
combined is only about 20% of live and dead plant
material plus dead plant parts. This is because animals
that eat plants or dead plant parts are only about 15%
efficient, and animals that eat plant eaters are on
average only 20% efficient. This inefficient transfer of
energy along a food chain is due to costs of doing
business. Costs can include hunting, pursuit, capture,
killing, eating, digestion, assimilation, making specific
proteins and fats and carbohydrates, and reproduction.
Organ system costs are circulation, respiration, excre¬
tion, neural processing, and metabolism. Then there
are losses of feces and urine. Because the sum of costs
goes up with steps in the food chain, it should be no
surprise that the mass of organisms must be progres¬
sively reduced from plants to herbivores to first-order
carnivores. In addition, no second- or third-order pre¬
dators live exclusively in caves, especially not endo¬
thermic birds and mammals that have very high costs
due to heat production and high metabolic rates.
Actually the food limitation in caves is even greater
than the above plant-based food chain argument
suggests, because food chains in caves mainly start
with losses from the above ground food chains. These
losses include death, feces, and urine with further
losses to the fungi bacteria that decompose waste
Encyclopedia of Caves.
© 2012 Elsevier Inc. All rights reserved.
324
F
products. Even though fungi and bacteria have rela¬
tively low costs of doing business, there is still further
energy lost with each species along a food chain of
decomposers.
The food limitation in caves should be still greater
because the amount, and especially sizes, of organic
matter that gets into caves is progressively reduced
with depth below the surface. This is because most
caves do not have entrances of greater than 1 cm
dimension, and those that have bigger entrances have
very few of them relative to cave length. Why is this
true? Even without the usual soil filter over the lime¬
stone, the soil moving downward with water tends to
plug most of the openings (look at a road cut in lime¬
stone). The very small vertical joint cracks in the lime¬
stone layers and horizontal bedding plane cracks do
not change much in size with depth, except in the rare
greater than 1-cm pathways that may enlarge and
become a cave (Palmer, 2007).
The restriction on entry of organic matter is even
greater because the pathways of greater than 1 cm are
convoluted. It is more like a series of pinball machines
than a vertical series of sieves of decreasing pore size.
Thus we should expect mostly millimeter-sized partic¬
ulate organic matter and mostly dissolved organic
matter to be brought into caves by percolating water.
The small amounts of tiny particulate and dis¬
solved organic matter that percolate into most caves
are likely to be low in food quality and very slowly
used by microorganisms that are the main base of
cave food chains. The logic is that the easily decom¬
posed organic matter is eaten, lost by leaching, and
used by microorganisms as it moves from the surface
through the soil and through decreasing sizes and
numbers of opening in the limestone. Soil detriti-
vores, such as earthworms, and decomposers, such as
fungi, "eat" the particulate organic matter. Bacteria
use the easily metabolized sugars and other simple
organic molecules leached out of particulate organic
matter by water. Thus most dissolved organic matter
to reach caves without entrances are large and com¬
plex molecules, such as fulvic and humic acids, that
are metabolized very slowly. The colonies of
Actinomycete bacteria, which use complex organic
molecules and give caves their characteristic musty smell,
grow less than 1 mm over decades. As is often true for
slow-growing microorganisms, Actinomycetes are known
to produce antibiotics such as streptomycin, which
protect them from being eaten.
Even the rare particulate organic matter that is
washed or that percolates into caves, with the usual
centimeter to decimeter scale entrances, is so thor¬
oughly leached that it is of low food quality and only
slowly used. In Mammoth Cave streams, leached and
black particulate organic matter does not disappear
over decades between rare 25- to 100-year floods that
wash in new organic matter (Poulson, 1992). However,
fresh leaf and wood litter entering caves via sinking
streams decomposes just as fast as in a surface desert
stream. Even in caves with very small entrances, fresh
sawdust or newly broken pieces of lumber are decom¬
posed so rapidly by fungi and bacteria that they use
up all of the dissolved oxygen in the water and so can
kill even specialized cave organisms with low meta¬
bolic rates.
The relations of coarse particulate, fine particulate,
dissolved organic matter, and biofilms to input by
sinking streams and ceiling drips has been extensively
studied by K. Simon and colleagues (chapter 2 in
Culver and Pipan, 2009).
Species Adaptations Predicted from Hypotheses
of Limited Food
If the hypotheses about extreme food limitation in
caves are correct then one prediction is that cave spe¬
cies with the highest costs of doing business should be
both relatively rare and should have many adaptations
to cope with a shortage of food. Table 1 lays out pre¬
dictions of decreasing adaptations to scarce food with
decreasing costs of doing business, decreasing size,
and decreasing complexity of organization of organ
systems and nervous systems from fish and salaman¬
ders to protozoa and bacteria. For each cost, bacteria
are given a score of 1 and estimates of the increasing
costs are given. The least increases are for cellular
functions, which should be the same, plus anabolism,
where the number of proteins, fats, and carbohydrates
increase from bacteria to fish and salamanders. The
greatest increases are for locomotion and foraging
where the costs increase dramatically with increasing
size and complexity. Summing estimated costs gives a
range from 4 to 68 for predictions of increased speciali¬
zations to cope with limited food supply.
Table 2 shows that the predicted specializations from
Table 1 are matched by observed increased specializa¬
tions to food limitation from bacteria to fish and sala¬
manders in caves. Among species that have evolved
from troglophiles to troglobites, from snails and flat-
worms to fish and salamanders, all appear white and
eyeless, but they are increasingly different from their
troglophilic relatives and surface ancestors in their
external morphology. Their body builds are less robust
with lower weight per length and lower metabolic
cost per length, their legs and sensory appendages are
longer, and their macroscopically visible sense organs
on body and appendages are larger, fancier, more
numerous, and more dense. Their internal anatomy,
where we can assess it macroscopically, also indicates
ENCYCLOPEDIA OF CAVES
FOOD SOURCES
325
TABLE 1 Energetic Costs 1 of “Doing Business”
Organisms
Locomotion
Foraging
"Capture" Ingestion
Digestion
Cell and Circulation
Costs
Respiration
Excretion
Sum of
Costs
Fish and salamanders
30
20
3
15
68
Arthropod predators
20
20
3
10
53
Arthropod detritivores
15
15
3
8
41
Snails and flatworms
6
5
2
4
17
Protozoa
3
2
1
2
8
Bacteria
1
1
1
1
4
1 Costs are estimated relative to bacteria, which are given a score of 1 in each category. Organisms are listed in order of decreasing size and decreasing complexity of organization.
The sum of costs is a predictor of the degree of food limitation. The adaptations listed in Table 2 support these predictions.
Based on these costs, it is hypothesized that fish and salamanders should be most limited by scarce food in caves and bacteria should be least limited.
TABLE 2 Troglomorphic specializations of kinds of organisms to food scarcity. These in caves support predictions from Table 1, based on
complexity of organization and costs of “doing business." Specialization are adaptations in terms of efficiency, metabolic economy, and
increased chance of successful reproduction. The number of pluses indicates the level of adaptation
Specialization
Fish and Arthropod Arthropod Snails and Protozoa and
Salamanders Predators Detritivores Flatworms Bacteria
Foraging
Search pattern
Reduced cost of transport
Sensory appendage length
Sensory organ density
Sensory organ kinds
Brain center l 1
Brain centers 2/3 2
Assimilation and growth
Metabolic
Routine metabolic rate
Starvation resistance
Reproduction
Egg size
Parental care
Multiple breeding attempts
Sum of adaptation
Sum of costs (Table 1)
+ + +
+ +
+
0
+ +
+ +
0
0
+ +
+ +
+
0
+ +
+ +
+ ?
0
+ +
+ +
+ ?
0NA
+ +
+ +
+ ?
0NA
+ ?
+ ?
0?
0NA
+ +
+
0?
0
+ +
+ +
+
0
+ + +
+ +
++
0?
+ + + +
+ + +
++?
0NA
0
0
0
0
+ +
+
+ ?
0
27
22
11
0
53
41
17
0
1 The first level of projection of sensory information, e.g., vagal (mouth taste), facial (skin taste), and somato sensory (lateral line touch).
Processing and translation of sensory information from brain center 1.
greater ability to find food from arthropods. Refer to
Mohr and Poulson (1966) or to Poulson and White
(1969) for drawings of sense organs, such as smell
and movement and touch and taste, and primary and
secondary processing parts of the brain that process
sensory information in troglophiles versus troglobites of
Amblyopsid cavefish and of crayfish. Troglobite forag¬
ing behavior is also more efficient at finding and utiliz¬
ing scarce food that is patchily distributed in space.
Where we have data, troglobites that are hungry move
long distances between turns and so increase their
chances of encountering an area with food. Once they
ENCYCLOPEDIA OF CAVES
326
F
find a food item their rate of movement slows, their
rate of turning increases, and they remain in the area
even after they eat a food item because there are likely
to be more food items in the same area.
Once troglobites encounter a food item, their effi¬
ciency at ingesting it and using it also increases from
flatworms to fish. Bacteria simply absorb molecules
across a cell membrane, and flatworms may catch live
prey using their mucus and ingest it by extruding
a pharynx. Among arthropods, some troglobitic spi¬
ders use a series of webs like a trap line to increase
chances of capturing some prey (Poulson, 1992). It is
suggested that they will have especially potent poisons
to subdue the largest possible prey. Among fishes,
Amblyopsids use many motion sensors on a very large
head to detect prey, and have dense touch receptors to
contact prey, huge mouths to suck in especially large
prey items, and extremely efficient digestion with very
fast growth in the rare cases where large prey items
are captured and eaten.
Aquatic and Terrestrial Differences
in Food Supply
Because water dissolves simple compounds faster
as it flows, there are major differences in the variability
and predictability of food, in time and space, in terres¬
trial and aquatic habitats in caves. This is true for food
input by abiotic agents such as flowing water, falling
into a pit, or blowing into a large entrance. It is also
true for input by biotic agents such as commuting bats
or cave crickets. The main difference is in whether the
food is diluted and dispersed widely in aquatic habi¬
tats or remains concentrated in a local area in terres¬
trial habitats. In terms of relative food limitation this
means that aquatic cave species will be more limited
and more specialized for energy economy than terres¬
trial species.
A corollary prediction is that aquatic detritivore
species should be slower growing and live longer than
terrestrial species. Indeed the data show that the differ¬
ences in longevity are incredibly large with the maxi¬
mum life span of terrestrial detritivore troglobites
(millipedes) on the order of a decade and the maxi¬
mum life span of aquatic species (crayfish) on the
order of a century. The selection for long life is that
renewal of the detritus at the base of the aquatic food
chain can be rare in time, on the order of decades (e.g.,
statistically 10-, 20-, or even 50-year floods). With the
same food base of plant detritus, terrestrial species
also get rare pulses of food, but it is left by floods on
stream banks and ceilings, so is not progressively
leached. Thus it will remain approximately the same
relatively high food quality for many years. This
means that the most food-limited terrestrial species,
such as millipedes and bristletails, can grow and com¬
plete their life cycles in less than a decade.
Exceptions That Prove the Rule of Food
Limitation in Caves
There are past and present situations that are both
much less food-limited and much more food-limited
than most of the limestone caves in today's temperate
zones of the Earth. These are considered at decreasing
scales in time and space. The corollary, from earlier
logic, is that this should affect the degree of species
specialization to food limitation.
At a long time scale of glacial cycles there were
probably times when less food washed into caves than
at present, especially for aquatic species that may have
been isolated in caves for tens of millions of years.
During both cold—wet periods with glacial advances
and during dry—warm periods with glacial retreats,
there would have been less food input into caves.
Because the more specialized troglobites are presumed
to date from preglacial times, present adaptations to
food scarcity are a complex result of fluctuating food
supplies, with times of greatest food supply "crunches"
being most important.
At an intermediate scale across space and over time
are contrasts between caves in flat-bedded limestones
in temperate and tropical regions, caves in flat-bedded
and steeply folded limestones in temperate regions,
and flat-bedded limestone caves and lava tube caves.
It is argued that wet tropical caves should be less
food-limited than temperate zone caves based on less
seasonality, greater rainfall, year-round occupancy by
bats in a higher percentage of caves, and faster rates of
cave development (with more entrances) in the wet
tropics. This is a strong hypothesis, but a Brazilian sci¬
entist (Eleonora Trajano) has argued that the tropics
are not uniform. For example, glacial cycles have had
great effects in the tropics, and there are both wetter
and more arid zones present within the tropics. It is
likely that data on degree of adaptation to food limita¬
tion within a taxonomic group are not yet adequate to
support or reject those predictions based on less food
limitation in the wet tropics than in temperate zone
caves.
The caves in steeply folded and thin strata of the
Appalachian Valley and Ridge Province of the United
States should be less food-limited than caves in the
flat-bedded and thick limestones of the Interior Low
Plateau Province. This prediction is supported mainly
by aquatic species with a greater number of species,
greater density of individuals, and higher proportion
of troglophile to troglobite species in caves of the
ENCYCLOPEDIA OF CAVES
FOOD SOURCES
327
Appalachian Valley and Ridge Province (Culver and
Pipan, 2009). The logic for this hypothesis is that the
geology results in more large cave entrances and more
regular and greater input of particulate organic matter
to valley and ridge caves where the ridge tops are
insoluble rocks, such as sandstone and shale, and the
valley sides and bottoms are soluble limestone with
caves. Water flowing off the insoluble ridge tops dis¬
solves entrances where it encounters the limestone, so
large amounts of unleached plant organic matter are
washed into the caves almost every year. The trade-offs
for cave species are that the food supply is reliable (+)
and plentiful (+), but that the risk of being dislodged
from rock refuges in streams and killed by abrasion or
being washed into pools with predators is high (—).
Under these circumstances short-lived troglophiles,
which can take fast advantage of good times and
recover quickly after bad times, coexist with more
energy-efficient troglobites that have longer life cycles
(especially isopod and amphipod species).
The contrast of limestone caves versus lava tube
caves is even more striking and even more strongly
supports the overarching hypothesis of the selective
influence of low food supplies in caves (Poulson and
Lavoie, 2000). Over geological time the changes in
food input and limitation are reversed in limestone
caves and in lava tubes. As erosion continues over
thousands to hundreds of thousands of years, lime¬
stone caves become more open to food input and so
less food-limited. But unlike lava tubes, we cannot
readily date their time of initial formation. The long
times prevent us from assessing how adaptations to
food scarcity change. In contrast we can exactly date
the time of lava tube formation and follow what hap¬
pens in lava tubes of different ages. Lava tubes are
most open to food input soon after sudden formation
when they are colonized both by troglophiles and tro¬
globites and population densities are at the highest.
Over only tens to hundreds of years both wind-borne
and water-borne sediments fill the larger voids above
the lava tubes and abiotic agents introduce less food.
Population densities de-cline and only those species
adapted to food scarcity persist.
On a short time scale of one to tens of years, organic
matter in the form of green "plants" (including blue-
green bacteria) and lint in terrestrial parts of commer¬
cialized caves can greatly increase. None of the
human-caused input of organic matter such as lint,
which can be widespread, or algae and moss, which is
localized around lights, seems to be eaten by special¬
ized cave species or above ground native species but
alien millipedes occasionally invade and graze algae
and mosses around lights even when dense congrega¬
tions of camel crickets are nearby and do not come to
the moss and algae. Why are neither outside species.
such as mice, or any cave species not attracted to the
plant growth or to lint-covered surfaces? One hypothe¬
sis is that these kinds of organic matter are of poor
food quality. Certainly blue-green bacteria and mosses
are heavily defended because few or no herbivores eat
them above ground, and much of lint is synthetic
fibers that are complex and either not decomposed or
slowly decomposed. A weaker hypothesis is that
extensive modern human visitation of caves is only
about 100 years old, so no species have had time to
evolve to utilize either lint or green photosynthesizers
associated with extensive human visitation. Still
another, not mutually exclusive hypothesis, is that
long-lived troglobites avoid areas because their preda¬
tors are also attracted. Long-lived terrestrial troglobites
must spread their risk of reproductive failure over
many years and so anything that might lower the
chances of successful reproduction, such as increased
predation risk, is avoided.
Pollution of aquatic habitats by decomposable
organic matter can occur across a large spatial scale,
everything downstream of the input point, and its
effects of decreasing the diversity of cave species occur
on a short time scale, often within weeks or months.
This is the same pattern seen outside of caves where a
huge abundance of only a few species of short-lived
and low-oxygen-tolerant species replaces a much lower
abundance and much higher diversity of species that
are not tolerant of low oxygen. Whether above- or
below- ground, the dominant food base with organic
pollution is colonial sewage bacteria and the only ani¬
mal may be Tubificid worms. Oxygen is low because
the biological oxygen demand is high when bacteria
and fungi rapidly decompose large quantities of
organic matter. This may not explain why energy-
efficient troglobites do not still occur, because they
have very low metabolic rates and so should be able to
live where oxygen is low though probably not at the
extremely low oxygen concentrations where sewage
bacteria survive.
Other relevant observations are of the path of com¬
munity recovery when input of excess decomposable
organic matter, such as human or animal wastes, is
stopped. First the sewage bacteria and worms are
replaced by somewhat longer lived and more energy-
efficient troglophiles such as crayfish and isopods.
Then, after sewage bacteria and worms disappear, the
area is recolonized by long-lived and energy-efficient
troglobites, the troglophiles become rare, and the origi¬
nal diversity of species is restored. The temporary
dominance of troglophiles suggests that either the
demographic pressure of faster reproducing species
or direct interference competition by the troglophiles
(low-metabolic-rate troglobites should win with indi¬
rect resource competition based on energy efficiency)
ENCYCLOPEDIA OF CAVES
328
may explain the dominance of troglophiles when oxy¬
gen is not limiting. Still, another hypothesis is that
pollution by decomposable organic matter is often
associated with pollution by toxins and that, compared
to troglophiles, troglobites are especially compromised
by bioaccumulation of toxins over their much longer
life spans and biomagnifcation of toxins if they are far¬
ther along food chains. Whatever the explanation, the
changes with organic pollution underscore the usual
food limitation in caves and the needs for adaptation
to scarce food by troglobites.
FOOD SOURCES AND FOOD TYPES,
ESPECIALLY IN TERRESTRIAL HABITATS
This section deals with the differences between
input by abiotic agents and biotic agents and the more
striking results of different kinds of biotic agents of
input in terrestrial than in aquatic habitats. Considered
first are aquatic versus terrestrial contrasts for the
same food sources and types, then terrestrial extremes
of food types, and the section ends with a detailed
look at different fecal types of biotic input in terrestrial
habitats.
Aquatic versus Terrestrial Contrasts for
Extremes of Food Quality
As Table 3 shows, hypothesized differences in vari¬
ability and predictability of food on land versus in
water depend on whether there is abiotic input by
water or biotic input by trogloxenes. There are oppo¬
site and common trade-offs for land and water for
quantity versus quality of abiotic and biotic inputs. On
the minus side for plant detritus, a combination of sur¬
face decomposers and leaching by water usually
removes all but the compounds most recalcitrant to
digestion, so bacterial and fungal decomposers are also
much reduced. On the plus side with 10- to 100-year
floods, especially in a few places near cave entrances,
renewal of unleached detritus may "feed" the cave
communities for another 10 to 100 years. Even leached
wood, twigs, and leaves decompose slowly in caves
because they are mostly lignins and cellulose that only
some bacteria and fungi can digest, and very slowly at
that. Once the detrital pieces get small enough to be
eaten by worms, amphipods, and isopods, the rate of
decomposition is enhanced. Cycles of bacteria-fungi-
protozoa biofilm buildup, ingestion of pieces and
digestion of the biofilm by animals, egestion of now
smaller particles, and recolonization of pieces by bio¬
film organisms all decrease the size of the pieces and
so the surface area to volume increases and decompo¬
sition slowly accelerates.
For biotic input by trogloxenes the quantity of food
is less (—) but the quality of food is much greater (+).
In addition, biotic input is much more predictable (+)
both in time, every year and every month, and in
space, usually near entrances and in the same spots.
However, these advantages are much less when inputs
are over water, as seen in the second column of
Table 3.
In water, leached feces of trogloxenes are not much
higher quality than leached plant debris, and both
TABLE 3 Trade-Offs (+ and —) of Food Variability and Predictability in Caves: How Quantity and Quality Vary in Space and Time
Trade-Off Category
Aquatic
Terrestrial
Abiotic Input 1
Occurrence
+ Gets everywhere
— Localized
Quantity
— Low densities
+ Locally high
Quality (leaching)
— Continued leaching
+ No further leaching
Renewal predictability
= / — Not every year
— Rare
Successional communities
— None
+ Some
Biotic Input 2
Regional occurrence
± Not everywhere
+ Localized
Occurrence in a cave
± Not everywhere
+ Localized
Quantity
— Dilution
+ Stays in place
Quality
— Only leached feces
+ Feces and others
Renewal predictability in a cave
± Moderate
+ Yearly
1 Abiotic input is by flowing water, wind, and gravity.
2 Biotic input is by trogloxenes such ns hats, wood rats, and cave crickets. Inputs are feces, dead bodies, and eggs.
ENCYCLOPEDIA OF CAVES
FOOD SOURCES
329
TABLE 4 Extreme Food Types in Terrestrial Caves: How Food Traits Influence the Ecology and Life History of Species That Eat Each
Type and How This Influences Species Diversity
Twigs and Leaves
Feces 1
Dead Bodies
Cricket Eggs
Food Traits
Caloric density
Low to very low
Moderate
High
Very high
Digestibility 2
Low to very low
Moderate
High
Very high
Occurrence
Among caves
High
Moderate
Low
Very low
Within a cave
High
Variable
Low
High
Amount per area 3
Very high
Variable
Low
High
Renewal rate
Very low
High
Low
High
Monopolized by single species?
No
Variable
If small
Yes
Species Hypotheses
Specialization?
No
Variable
No
Yes
Life span
Long
Variable
Short
Short
Species diversity
High
Variable
Low
Very low
Ms shown in Table 5, there are many kinds of feces, so many of the categories are variable.
2 Digestibility is largely determined by the N:C ratio.
3 Amount per area is at locations where the food type is present, i.e., nonzero values.
become diluted and dispersed over wide areas. The
only advantage of this reduction in quality is that a lit¬
tle food occurs everywhere and is easy to find.
Consumers of this low-quality food must be very effi¬
cient digesters and are thus slow-growing and are
probably long-lived. They also must be opportunistic
omnivores. Species that have predatory ancestors, espe¬
cially fish and salamanders, will get some benefit from
eating detritus with a biofilm if their preferred food of
copepods, isopods, and amphipods is especially scarce.
Species that have herbivorous or detritivorous ances¬
tors, such as crayfish and isopods, will greatly benefit
from eating occasional live prey. For all of these rea¬
sons there is little opportunity for specialization of rela¬
tively few species in aquatic cave habitats.
Terrestrial Extremes of Food Types: Leaves,
Feces, Dead Bodies, and Cricket Eggs
Many more species are on land because they have
opportunities for specialization and almost no minus
trade-offs (Table 3, last column). With less leaching,
food quality is higher. Food quality varies from low
with partly leached litter, to feces, to dead bodies, to
very high with eggs. With no dilution, input stays
where it is deposited, and this makes the food input
more predictable in time and space. The consequences
of these on-land advantages are explored in the next
section, but suffice it to summarize here that a high
diversity of terrestrial cave species occurs for four rea¬
sons (Table 4): (1) There are many food types with a
wide range of quality and quantity; (2) within many
food types there is a heterogeneity of digestibility for
different components and so there can be successional
replacement of species during decomposition; (3)
predictability in time and/or space at, within, and
among cave scales allows some species to specialize
(with narrower niches more species can coexist);
and (4) there is different local usability of the same
food types because of interactions among microhabitat
and moisture.
Table 4 contrasts aspects of food type and aspects of
the species that are expected to use each type. Quality
of food includes both caloric density and digestibility
as indexed by carbon to nitrogen ratio. Predictability of
food type includes both frequency of occurrence within
and among caves and amounts where it occurs. These
food type trait differences lead to hypotheses about the
likelihood that species will specialize to use each type.
At one extreme are variably leached leaves and twigs
that are low in quality but high in amount; for this
type consumers with little specialization and long life
spans are expected. At the other extreme are cricket
eggs that are high in quality but low in amount; for
this type consumers with great specialization and short
life spans. Corollaries of these primary predictions are
secondary predictions about the diversity of species
likely to use each food type. The highest diversity of
ENCYCLOPEDIA OF CAVES
330
F
consumer species is predicted for litter because it is
most heterogeneous, can undergo very slow succes-
sional decomposition by different species, and has a
low renewal rate. Its lowest quality components
can last for many years, often decades, and cannot
be monopolized by one or a few species. The lowest
diversity of species is predicted for cricket eggs
because they are homogeneous, have no successional
use, have a yearly renewal rate, and can be monopo¬
lized by one species. This is a more specific case of the
generalizations discussed above for cave species in
general. Thus if food is of low quality and low renewal
rate, in our example litter we expect species to be
generalized with the plus and minus tradeoffs of a
jack-of-all-trades but a master of none. We expect the
opposite trade-offs for a specialist that is favored by
food of high quality and regular renewal, such as our
example of cricket eggs. These predictions are best
supported by extensive data on the most specialized
cave species known, namely Carabid beetles that
specialize on eating cave cricket eggs (Griffith and
Poulson, 1993).
A cricket egg is a real prize that can be quickly
eaten by one beetle of about the same size as the egg,
the ultimate in monopolization of and energy return
from a single food item. A beetle spends only 10% of
its time digging up eggs but 90% of its energy digging,
so it is not surprising that a beetle defends its ongoing
investment of digging given the high caloric reward of
an egg. In fact the deeper the hole the more aggres¬
sively a beetle defends the hole. Once a beetle digs up
an egg it runs with it until it is in a place safe—under
a rock, up the wall, or on the ceiling—from other bee¬
tles that will try to steal the egg because this is much
less expensive than digging up its own egg.
Given the high payoff it is not surprising that beetle
predators of cricket eggs have sophisticated searching
behaviors. First, they slow their rate of walking and
increase their rate of turning when on the silt or sand
substrates where cave crickets lay the overwhelming
majority of their eggs. Second, beetles dig in spots
where crickets have raked silt or sand into tiny mounds
over the egg-laying site and may avoid the holes that a
cricket makes to test the suitability of a site by inserting
its ovipositor. The beetles may recognize an egg mound
not only by texture but also because the raking of the
cricket ovipositor removes beetle scents left as they
forage for eggs. Third, the beetle digs only to a depth
equal to that of the cricket's ovipositor, because if there
is no egg at that depth there will be none deeper.
An even stronger support for these predictions of
specialization to eat cave cricket eggs is the convergent
evolution of this habit and in body size among at least
five different species of Carabid beetles that eat eggs
of at least three different species of cave crickets. In
each case the cricket egg is about 5 mg and in most
cases closely matched to the weight of the beetle. In
some cases the ancestral species was smaller and
in some cases larger than about 7 mg. Also, each beetle
species chooses a substrate habitat of silt or sand closely
matched in moisture content to that best for cricket egg
development. In each case the beetle's life history is
closely matched to the seasonal cycle of cricket egg lay¬
ing. As predicted in each case, the egg can be monopo¬
lized by one individual beetle.
As with cricket eggs, another food item that is high
in quality and can potentially be monopolized is dead
bodies. Corpses that are too large to be carried away
by scavengers are usually at the bottom of deep pits
where surface species can blunder in and fall to their
death. This is rare within and among caves, and we
have no data to suggest that any cave species even has
this food item as part of its regular diet. On the other
hand, under maternity colonies of bats that number in
the thousands ( e.g ., gray bats) to millions (free-tailed
bats), baby bats regularly fall from the ceiling and are
quickly eaten by hordes of small scavengers, especially
Dermestid beetles and camel crickets. Tiny corpses that
can be carried away are rarely seen in caves because
they are carried away by scavengers (see Mohr and
Poulson, 1966). In fact, if we incapacitate or kill a small
beetle or cricket and watch, we usually see that either
beetles or crickets of the same species as the food item
almost immediately find it and carry it away. Thus in
this case the dying or dead species does not support
any more species to the cave community; it merely
enlarges the feeding niche of existing cave species that
we predicted should be opportunistic feeders.
At the opposite extreme are food types that cannot be
monopolized and have a high diversity of consumers,
especially leaf and twig litter of very low quality and
large piece size. Recall that in Table 4 it was predicted
that the low quality and low predictability of renewal
with very long times for decomposition, often decades,
would prevent monopolization or specialization and
might support a relatively high diversity of cave species.
Also, relatively undecomposed litter is carried into caves
only by infrequent floods (e.g., 20- to 50-year events)
and avoids continued leaching only if deposited at the
extreme high watermark on ceilings or high on stream
banks. Only there is it available as a moderately high
quality food source every year for terrestrial cave spe¬
cies. In such situations many generalist species make up
the cave community. Among eaters of fungi and bacte¬
ria, either grazed on or digested off of as a biofilm, are
primary consumers such as springtails, bristletails, and
millipedes. Predator generalists that eat the primary
consumers include daddy longlegs, spiders, pseudoscor¬
pions, Carabid beetles, and various mites. Springtails
live for several years and millipedes may live for a
ENCYCLOPEDIA OF CAVES
FOOD SOURCES
decade. Primary and secondary consumer predators
such as Carabid beetles live for several years and
Linyphiid spiders for more than a decade.
So we see that the predictions for the extremes
of terrestrial food types (Tables 3 and 4) are
upheld, but the real test of the generality of these pre¬
dictions is to look at an intermediate food type that
has a wide range of food qualities and quantities, that
is, feces.
A PLETHORA OF FECAL TYPES
In terrestrial habitats, of all the biotic inputs of food
to caves by trogloxenes, their feces, and the usability of
their feces, are most different and offer the greatest
number of feeding niches for cave animals. It can be
said that you are what you eat and your feces reflect
even more what you eat. Consider, as a human
331
omnivore, your own feces when you eat hot spicy food
compared to salad compared to meat and potatoes (the
cave community would be incredibly richer if humans
lived in caves). In addition, the usability of any one
fecal type can vary with its predictability in time and
space and with the moisture of the substrate where it
is deposited. Is it, as in Goldilocks, too wet, too dry, or
just right? Thus if, like Doctor Doolittle, we could
question cave species about food they would tell us
that cave species have an incredibly rich vocabulary
for feces. Their vocabulary for feces is every bit as rich
as the vocabulary of Inuit Eskimos for frozen water,
which includes words for different forms and amounts
of snow and ice. As Eskimo survival is related to snow
and ice, cave animals would tell us how different
forms and amounts of feces are critical for niche sepa¬
ration of cave species that feed on the feces of raccoon,
pack rat, bat, cave cricket, and cave sand beetle
(Table 5).
TABLE 5 In Terrestrial Habitats in Caves, There Can Be a Diversity of Fecal Types and a Diversity of Species Eating Feces
Raccoon
Pack Rat
Bat (insectivorous)
Cave Cricket
Cave Beetle
Fecal Traits
Quality
Caloric density
High
Moderate
Low
Very low
High
% easily digestible
50
30
20
60
70
Low digestibility
Hair, seeds
Hair, wood
Chitin
Little
None
Diet heterogeneity
High
Moderate
Low
Low
Very low
Quantity Where it Occurs
Turd size mg.
13,700
60
20
5
0.1
Turds per day
10
20
50
5
5
Spatial dispersion
Random
Latrine
Roost to flight paths
Roost to walk-by
Under rocks
Predictability
Among caves
20%
30%
10%
90%
70%
Area within cave
10%
5%
30%
40%
15%
Seasonality
Moderate
Little
Great
Little
Little
Drying risk
Winter
Usually
Little
Rarely
None
Leaching risk
None
None
None
Moderate
None
Predictions/Tests
Species specialization
To fecal type
No
A little
Yes
Yes
Yes
To microhabitat
No
Rocks, dirt
No
Rocks, dirt
Sand, dirt
Species succession
Yes, a lot
Yes, with depth
Yes, with depth
A little
None
Species diversity
+ +
++ +
++ + +
+ + +
+ +
Note: For each fecal type, there should be increasing specialization of feeding with increasing predictability and quality of the feces but decreasing diversity of
species. These data, predictions, and tests are from the Interior Low Plateau of east central United States and from Texas and New Mexico.
ENCYCLOPEDIA OF CAVES
332
F
In Table 5, quality, quantity, and predictability of
feces are considered, and how they lead to hypotheses
and predictions about whether cave species specialize
on one fecal type and whether distinct communities of
species are associated with each fecal type. For quality,
the most important component is ease of digestibility
and the heterogeneity of the fecal producer's diet.
For quantity important variables are where feces
occur, turd size, numbers of turds per day, and spatial
dispersion ( e.g ., random or clumped). Predictability
includes spatial and temporal occurrence within and
between caves as well as abiotic constraints on
whether the feces are dried or leached, both making
them less useable. Quality, quantity, and predictability
of feces together lead to opposite predictions for
whether a species will specialize on a single fecal type
and whether a diverse community of species will be
associated with a single fecal type. The best predictors
of species specialization are high quality, homogeneity
of quality, and high predictability in time and space,
for example, fresh bat feces at the top of a pile of
guano under maternity roosts. The best predictors
of species diversity are heterogeneity of quality, which
leads to successional replacement of species, and large
quantities, which leads to different species using dif¬
ferent densities of feces, for example, from the top of
a bat guano pile to areas beyond the periphery of a
guano pile under a maternity roost. Because maternity
colonies of bats are rare, the overall specialization of
species to using bat feces is relatively low. What of
other fecal types? We will now give a summary for
each type. For details about communities associated
with each fecal type, especially bats with completely
different diets (fruit and blood in the tropics in addi¬
tion to insects in the temperate zone), see Poulson and
Lavoie (2000).
Raccoon feces should not support specialist fee¬
ders but its heterogeneity in quality and low predict¬
ability should allow use by a wide variety of species
at the community scale. Raccoons rarely wander into
caves and are only rarely found far from entrances.
On this basis we should predict only generalist
opportunistic feeders on feces. In fact, in early stages
of decomposition in summer there can be domination
of the most easily digestible components by dense
squirming masses of maggots of Sphaerocerid flies
that usually have a wide feeding niche and are only
locally abundant. With fresh raccoon feces, the large
size of individual turds and a tangling of turds in
piles allows maggots to do best due to their collective
facilitative behavior with an abundance of external
digestive enzymes and increased rate of develop¬
ment allowed by generation of body heat. As
successional decomposition proceeds, with fewer
and fewer digestible components available, larvae of
scavenger and fungal-eating species become the most
common. At the end of decomposition, when the
only remaining components are nearly undigestible
{e.g., raccoon hair or seeds) there is an even larger diver¬
sity but a very low density of very small generalist fee¬
ders. These can use tiny and scattered amounts of
remaining digestible components and include spring-
tails, mites, and scavenger beetle adults (Leiodids).
In winter the community using raccoon feces has
few, if any, animal species but all major size classes of
fungi. Low temperatures and dry conditions just inside
cave entrances preclude use by most animals that in
summer eat fungal hyphae or disrupted hyphal net¬
works in the process of feeding. The result is that fungi
cannot accumulate enough energy to produce increas¬
ingly large fruiting bodies from Phycomycetes to
Ascomycetes to the mushrooms of Basidiomycetes.
In contrast to the random deposition of raccoon
feces in place and time, pack rats have fecal latrines
that can be in the same spot for decades, centuries,
and even millennia. Consequently we might predict
that many cave species could specialize on this fecal
resource; however, they do not. Let us consider why
not. The problem is that pack rats occur in few caves,
regionally and locally, and where they do occur their
latrines are usually in very dry entrances where, at
best, only fungi can do well as users of feces. Because
pack rats urinate elsewhere, for example, as one way
of scent marking trails in and out of the cave, there is
no mitigation of dry conditions by urination as can
occur with bats that almost always urinate over their
fecal piles in caves.
In cave entrances that have moist microclimates,
pack rat fecal latrines can have a rich community of
fecal users. The piling of pellets helps to create a
stable microclimate which buffers variation in the cave
microclimate and collectively creates a resource with
low surface area to volume. The largest latrines have a
spatial zonation, from pile top where fresh feces are
deposited daily to pile edges where both fresh and old
feces are scattered as a mosaic. A large species of fun¬
gus gnat grazes fungi on the freshest feces and a smal¬
ler species of fungus gnat eats a combination of fungi
and bits of feces deeper in the pile and toward the
periphery. Still deeper in the pile the feces are only
powder consisting of mostly hard-to-decompose lignin
and cellulose. Use of partially intact feces at the
periphery depends on whether the substrate is rock,
mud, or sand. On rock there is a smelly slurry of solu¬
ble nutrients with domination by saprophytic mites
that can tolerate the low oxygen. On mud there can be
large numbers of two species of springtail and a few
predatory pseudoscorpions. On sand most nutrients
are leached and there are more species with no domi¬
nance. Thus the overall diversity of species using piles
ENCYCLOPEDIA OF CAVES
FOOD SOURCES
333
of rat pellets depends on depth in the center of the pile
and the degree of leaching into the substrate at the
periphery of the pile.
As with rat feces, sand beetle (e.g. Neaphaenops) feces
are localized and their use depends on the substrate; the
difference is that sand beetle feces are much more
predictable because there are so many areas where cave
crickets lay eggs eaten by sand beetles. Thus we might
predict some specialization in use of their feces. As
explained previously, a 5-mg cricket egg is a huge meal
for a 7-mg sand beetle, and the successful beetle usually
retreats under a rock to prevent another beetle from
stealing the egg. Under the rock the beetle ingests the
egg and urinates and defecates as it digests the egg over
a 1- to 2-week period. The community of consumers and
their predators that use the feces depends on whether
the rock is on a sand or silt—mud substrate. On silt the
main consumers are scavenger beetles and springtails,
and both mites and pseudoscorpions are predators. On
sand the main consumers are bristletails, springtails, and
mites and spiders are the main predators.
The feces of insectivorous bats can be like rat feces
in occurring in piles, unlike rat feces in having no sub¬
strate constraints, and like raccoon feces in having rare
places in which piles occur. In the rare caves with heat
traps and adequate moisture where free-tailed or gray
bat maternity colonies exist, the piles are so large that
they create their own microclimate and the substrate is
not an important constraint on use by cave animals.
The microclimate requirements are so exacting that
maternity colonies occur in fewer caves than rare
hibernating roosts or transient colonies. The maternity
colonies have stayed in the same caves for decades
and probably centuries, and so it has been possible for
some cave animals to specialize on this locally super¬
abundant food resource. When mother bats are nurs¬
ing in a maternity colony the input of feces and urine
is prodigious, and the surface of the guano pile can be
a seething mass of varying combinations of mites,
springtails, and camel crickets. These feed both
directly on feces and urea from urine and also indi¬
rectly on the bacteria and fungi that metabolize feces
and urine. As detailed in Poulson and Lavoie (2000),
the animals using feces in huge piles of bat guano vary
with depth in the pile and with decreasing density
of feces from the center of the pile to the periphery.
This zone blends into a fly-by scattering of fecal pellets
analogous to the walk-by scattering of fecal splotches
of cave crickets away from their entrance roosts.
Cave cricket and bat feces share a huge spatial vari¬
ation in local density from areas under roosts to walk-
and fly-by scatterings of turds on the way to and from
near-entrance roosts. In both cases there is a gradient
to higher diversity and more adaptation to food scar¬
city from the periphery of roosts to the much less
dense feces and lower renewal rates in fly- and walk-
by areas. The difference from bats is that cave crickets
occur in almost every cave within their geographic
range so a more predictable community is associated
with the cricket's scattering of feces.
Unlike the huge piles of bat feces under maternity
roosts, the veneers of feces under cave cricket roosts are
another example of the Goldilocks principle for sub¬
strate moisture; it should be not too wet or too dry to be
used by the highest diversity of cave animals. If the sub¬
strate is steeply sloped with lots of drip water from spe-
leothems, then leaching removes what little nutrients
are in the feces. If the substrate is too dry, for example,
due to air movement in cold and dry winter microcli¬
mates, then only fungi can use the feces. The most
diverse communities with highest densities of each
species are found on flat rocky areas with the locally
deepest veneers of fresh feces, such as under protected
ceiling microhabitats of mini domes. In these areas both
a common snail generalist and a rare snail specialist are
found as well as other specialists including a millipede
and two mites as consumers and a Pselaphid beetle and
web worm as predators. On both mud and rock sub¬
strates Oligochaete worms may preempt the feces, and
on sand the feces can be used only if rates of fecal input
are high enough to create crusts that buffer the local
moisture microclimate (Poulson, 1992).
Cave Crickets as Key Industry Species
Cave crickets are key species by virtue of their
high frequency of occurrence, high densities where
they occur, and high impact per individual as the
largest terrestrial cave invertebrate. Within their geo¬
graphic range they are found in virtually every cave
with high densities in roosts just inside cave entrances.
They are the basis for three separate cave communi¬
ties. At entrances they go outside to feed about every
10 days and return to their roosts where they digest
crop contents that may be 100—200% of their body
weight. As they digest they defecate and one commu¬
nity depends on this fecal input. Adult crickets walk
from entrance roosts to reproductive areas deep in the
caves and defecate along the way. The walk-by feces
they deposit, at very low densities with infrequent
renewal, are the most important and reliable energy
source for the most energy-efficient terrestrial species
in the cave. The eggs they lay in sand and silt areas are
eaten by Carabid sand beetles, and beetle feces are the
basis for the third cave community that depends on
cave crickets. Cave crickets are not only key species in
several of the karst areas of the United States but also
in parts of Europe and perhaps in Asia and South
Africa.
ENCYCLOPEDIA OF CAVES
334
F
See Also the Following Articles
Adaptation to Lozv Food
Guano Communities
Chemoautotrophy
Bibliography
Culver, D. C., & Pipan, T. (2009). The biology of caves and other subter¬
ranean habitats. New York: Oxford University Press.
Griffith, D. M., & Poulson, T. L. (1993). Mechanisms and conse¬
quences of intra-specific competition in a Carabid cave beetle.
Ecology, 74(5), 1373-1383.
Mohr, C. E., & Poulson, T. L. (1966). The life of the cave. Living World
of Nature Series. New York: McGraw-Hill.
Palmer, A. N. (2007). Cave geology. Dayton, OH: Cave Books.
Poulson, T. L. (1992). The Mammoth Cave ecosystem. In A. Camacho
(Ed.), The natural history of biospeleology (pp. 569—611). Madrid,
Spain: Museo Nacional de Ciencias Naturales.
Poulson, T. L., & Lavoie, K. L. (2000). The trophic basis of subsurface
ecosystems. In H. Wilkens, D. Culver, & W. Humphreys (Eds.),
Ecosystems of the world 30: Subterranean ecosystems (pp. 231—250).
Amsterdam: Elsevier.
Poulson, T. L., & White, W. B. (1969). The cave environment. Science,
165(3897), 971-981.
FRIARS HOLE SYSTEM, WEST
VIRGINIA
Stephen R.H. Worthington and Douglas M. M edville'
Worthington Groundwater
Virginia Speleological Survey
The Friars Hole System, in Greenbrier and Pocahontas
counties. West Virginia, is the longest cave, in terms of
surveyed passage (73.3 km, or 45.5 miles), in both the
Valley and Ridge and Appalachian Plateaus provinces in
the eastern United States. The cave's geological setting, in
moderately faulted and gently dipping Mississippian
limestones, is in the midpoint of a range of Appalachian
cave settings, from those found in flat bedded unfaulted
carbonates to the west, in the Cumberland Plateau, for
example, to those in the more highly faulted and folded
limestones in the Valley and Ridge Province to the east;
Burnsville Cove in Virginia, for example.
As such, the Friars Hole System is an excellent
example of a large cave, the development of which has
been influenced by minor structures such as low angle
thrust faults and the availability of competing flow
paths to two regional base level springs 20 km apart.
As a result, the cave has a complex hydrological his¬
tory and a plan view that reflects both abandoned flow
paths and currently active internal drainages within
the cave system.
SETTING OF THE CAVE
Location
The Friars Hole System is located close to the bound¬
ary between the Valley and Ridge and Plateau provinces
of the Appalachian Mountains. Friars Hole itself is a
150-m-deep dry valley between the summits of Droop
Mountain and Jacox Knob, which are 850 m asl. Friars
Hole extends for 13 km from the sinks of Bruffey Creek
and Hills Creek in a southerly and southwesterly direc¬
tion to Spring Creek (Fig. 1A). Sandstones, shales, and
minor limestones of the Mississippian Mauch Chunk
Series underlie these summits and the valley sides, but
there are a series of inliers of Union Limestone along the
valley bottom (Fig. 2). Streams flow onto the limestone
sink soon after reaching it, and the entrances to Friars
Hole System are found there. A number of these sinking
streams feed active stream passages within the cave,
which underlies a 6-km length of Friars Hole. Much of
the cave lies under the nonkarstic flanks of the valley,
with passages lying up to 185 m below the surface.
Hydrology and Hydrogeology
Hills Creek is the largest sinking stream along Friars
Hole and its hydrology is complicated. At low flow it
sinks completely in its gravel bed and is seen in suc¬
cession in Cutlip Cave, Clyde Cochrane Cave, and
Friars Hole System, where it forms Rocky River, the
largest stream in the cave (Fig. 2). At high flow that
part of Hills Creek that does not sink in its bed con¬
tinues on the surface for 500 m beyond the main sink,
and enters Hills-Bruffey Cave, which also takes the
high flow discharge from Bruffey Creek. Hills-Bruffey
Cave drains to the east to Upper Hughes Cave, Lower
Hughes Cave, Martha's Cave, General Averell Pit, and
finally Locust Creek Cave, which is the cave at Locust
Spring on the east side of Droop Mountain. In very
high flow conditions the water backs up at the
entrances of Hills-Bruffey Cave and overflows on the
surface to its final sinkpoint at Cutlip Cave, from
which it flows south through Clyde Cochrane Cave
and then through the Friars Hole System via Rocky
River and rises at JJ Spring on Spring Creek, 19 km to
the southwest of the Cutlip Cave entrance (Fig. IB).
The next sinking stream to the south of Hills Creek is
Rush Run and, like Hills Creek and Bruffey Creek, has
part of its discharge flowing to Locust Spring and part
to JJ Spring.
The streams that sink directly into the Friars Hole
System all drain to the southwest to JJ Spring. This
spring drains some or all of the flow from an area of
265 km 2 , and the streams draining via the known cave
Encyclopedia of Caves.
© 2012 Elsevier Inc. All rights reserved.
FRIARS HOLE SYSTEM, WEST VIRGINIA
335
Caves:
CC Clyde Cochrane Cave
Cutlip Cave
Fox Cave
Friars Hole Cave
General Averell Pit
Hills Bruffey Cave
Hughes Creek Cave
Locust Creek Cave
Martha's Cave
Robbins Run Hole
The Boarhole
The Portal
Zotter's Folly
West
Virginia
<>
-- area of
main map
FIGURE 1 Topography and hydrology of the Friars Hole area. Friars Hole is the dry valley between the sinks to Hills Creek and Spring
Creek (A). Mapped caves in the drainage area of Friars Hole. Arrows represent traced flow paths of cave streams (B).
Cave passages:
CN Crowsnest Room
CO Crossover Passage
CR China Room
Downlets
DW Droughtway
FT Friars Trunk
Lew's Last Climb
Lost Passage
MC Monster Cavern
PM Parker Mountain Passage
RR Rocky River
I I Union Limestone outcrop
LL
LP
FIGURE 2 Cave passages of Friars Hole
System, showing the relationship to surface out¬
crops of limestone.
ENCYCLOPEDIA OF CAVES
336
F
in the Friars Hole System account for only a third of
this discharge (Fig. 1). JJ Spring is located on Spring
Creek, which flows across the Union and Pickaway
limestones for several kilometers. At low flow both
Spring Creek and its tributaries sink in their beds on
reaching the Union Limestone, and the bed of Spring
Creek is dry for 13 kilometers. At high flow the sinks
are overwhelmed and there is flow along the length of
Spring Creek (dashed line in Fig. 1A). There must be an
extensive cave system underlying Spring Creek and its
tributaries. One fragment of this is seen in Robbins Run
Hole, most of which lies 25 m below the bed of Robbins
Run. A second, flooded, fragment has been explored at
Circulating Cenote, one of the springs for the Friars
Hole System. Several other caves have been explored
between the Friars Hole System and JJ Spring, including
Fox Cave, Zotter's Folly, the Portal, and the Boarhole,
but these are all located on the flanks of Spring Creek
rather than directly under the creek (Fig. IB).
Although JJ Spring is the perennial spring for Friars
Hole, there are also three nearby overflow springs, the
Cannon Hole, the Circulating Cenote, and Dale's
Spring. The Cannon Hole is in fact an estavelle since at
low flow it absorbs the surface flow of Spring Creek.
The complexity seen at the springs of the Friars Hole
System is common in karst aquifers, where distributar¬
ies are common where conduit systems discharge to
the surface. Furthermore, the pattern of one perennial
spring plus several intermittent overflow springs is
also common. It is likely that the drainage to JJ Spring
and the associated springs follows a single large con¬
duit for most of the distance between Friars Hole and
the springs, as shown in Figure IB.
Geology
Almost all of the Friars Hole System is formed in
the 48-m-thick Union Limestone and the underlying
30-m-thick Pickaway Limestone, which both belong to
the Mississippian Greenbrier Group. The Union
Limestone is primarily a pure limestone, with 91% of
the section at Friars Hole being composed of oosparites
and biosparites. However, there are four calcilutites,
and these range in thickness from 50 cm to 3 m and
have insoluble fractions that range from 47% to 63%.
The Pickaway Limestone is much less pure than the
Union Limestone, and consists of a regular alternation
of calcilutites and calcarenites of up to 3 m in thickness.
The underlying Taggard Formation is 7 m thick, and is
composed predominantly of shale. It is only seen in the
cave in the passages underlying the Crowsnest Room
(Fig. 2), and JJ Spring is perched on the Taggard Shale.
The strata dip is to the northwest at 2° within the
Friars Hole System and the dip is similar along the
flow path to the springs. However, to the east of the
cave underneath Droop Mountain there is a monocline
which probably has thrust faulting associated with it.
Thrust faulting is also seen at many locations within
the cave (Fig. 3A). Many thrust faults in the cave have
a strike of around N25°E and dip to the west, though
there is much variability. The major joint set in the
cave strikes at N60°E to N70°E. Although a majority of
passages in the Friars Hole System are developed
along joint sets and solutionally enlarged bedding
plane partings that are parallel or subparallel to the
strike, several substantial passage segments follow
joints and bedding partings down-dip to the northwest
(Fig. 2).
HISTORY OF EXPLORATION
Although the Snedegars entrance to the Friars Hole
System was used for saltpeter mining during the U.S.
Civil War, contemporary exploration of the cave began in
the early 1950s with the descent of the 30-m Crookshank
Pit, followed by its connection to Snedegars Cave in June
1964. The entrance to Friars Hole Cave, 4 km to the south¬
west of Snedegar Cave, was dug open in August 1964
and led to over 6 km of passage in a separate cave.
In March 1976, following excavation of a doline con¬
taining a sinking stream 1800 m to the northeast of the
Friars Hole entrance, 7 km of passage were surveyed
in Rubber Chicken Cave. Within two months, connec¬
tions between it and the nearby Friars Hole and
Snedegar-Crookshank caves were established, result¬
ing in a single cave (the Friars Hole System) with a
combined length of 24 km of surveyed passage.
In the summer of 1976, a major extension was found
in Canadian Hole, a short, multipitch cave 600 m to
the north of the Snedegar Saltpetre entrance. Over the
next year, over 8 km of passage were surveyed, culmi¬
nating in November 1977 with a connection to the
Friars Hole complex to the south. A final connection to
another nearby cave (Toothpick Cave) was made in
November 1978, resulting in over 42 km of surveyed
passages in the entire complex.
Continued work in the Friars Hole System since
then has increased the total surveyed length of the
cave to 72.7 km by the summer of 2002. A significant
extension to the cave took place in 1995—2000 where
an upper passage (Fig. 2, PM) entering the Friars Hole
trunk was followed for 2.2 km to the west and north¬
west, ending beneath the Robbins Run valley 600 m to
the west of the Friars Hole valley and establishing
Robbins Run as a paleo infeeder to the Friars Hole
System. Additional surveys since 2002 have increased
the surveyed length of the system to 73.3 km.
ENCYCLOPEDIA OF CAVES
FRIARS HOLE SYSTEM, WEST VIRGINIA
337
FIGURE 3 Cave passages of Friars Hole System, showing passages following thrust faults (A), active inlets and cave streams (B), strati¬
graphic control of main drains (C), and a possible sequence of development of the cave (D).
HYDROLOGY OF THE CAVE
Seven of the entrances of the Friars Hole System
(Friars Hole Cave, Rubber Chicken Cave, Toothpick
Cave, Crookshank Cave, Canadian Hole, and the
Staircase and North Entrances to Snedegars Cave) are
active stream sinks (Figs. 2 and 3B). The remaining
three entrances. Radio Pit and the Saltpetre and Maze
entrances to Snedegars Cave are abandoned stream
sinks. In addition to the active stream sink entrances to
the cave there are more than 100 other inlet passages
where streams currently enter the cave. These range
from small flows to large sinking streams.
In the cave, many of the inlet passages can be fol¬
lowed upstream to sediment blockages quite close to
the surface, and thus these locations are clustered
underneath the surface inliers of the limestone (Figs. 2
and 3B). These passages are typically tall vadose can¬
yons that range from less than 50 cm up to several
meters in width. They usually descend joints through
the Union Limestone, but also often flow down-dip
for short distances when they reach one of the impure
limestone beds. Most inlet passages descend through
the Union Limestone within a few tens of meters,
though a few are longer. The longest passage in
the Union Limestone is Lew's Last Climb, which was
initiated along a low-angle fault, descending over a
distance of 600 m before reaching the Pickaway
Limestone; subsequently a tall, narrow canyon pas¬
sage has downcut below the fault (Figs. 2 and 3A).
There is some integration of stream passages in the
Union Limestone, but most takes place in the
Pickaway Limestone, where the stream passages have
much lower gradients. Within the known cave the
major streams integrate into three major drainages
(Fig. 3B). The eastern drainage has the largest cave
stream. Rocky River, and drains some or all of the flow
from a catchment area of 45 km 2 . Rocky River is a large
strike-oriented cave stream that can be followed in the
cave for more than 2 km to a boulder blockage (Fig. 3B,
point A). There are several long tributaries to Rocky
River, such as the one sinking at the Canadian Hole
ENCYCLOPEDIA OF CAVES
338
F
entrance. This descends through the Union Limestone
with three major drops (11 m, 15 m, and 9 m), and then
flows down-dip in the upper Pickaway Limestone until
it joins the strike-oriented Rocky River.
The central drainage in the cave includes the
streams sinking at the Toothpick, Snedegar Staircase,
Crookshank, and Rubber Chicken entrances to the
cave. The longest stream is the Toothpick stream,
which can be followed almost continuously for 5.5 km,
ending at a sump, which is the lowest point in the
cave (Fig. 3B, point B). The path of the Toothpick
stream is sinuous. After dropping rapidly through the
Union Limestone by two 10-m drops, most of the pas¬
sage is in the basal Union Limestone or uppermost
Pickaway Limestone, and much probably formed at
the intersection of major bedding planes and thrust
faults (Figs. 3 A and 3C).
The western drainage in the cave is aligned along
the strike and closely follows the base of the Friars
Hole valley. Consequently, only short tributaries join
the main stream. The cave stream is last seen in an
aqueous passage called Water World (Fig. 3B, point C).
Along most of its path the stream follows Friars
Trunk, which is one of the largest passages in the cave,
averaging 15 m in width and 8 m in height.
The three separate cave streams seen in the cave
join together before reaching JJ spring (Fig. IB). The
gradient from the lowest point in the cave to the
spring is only 0.002, and it seems likely that most of
this strike-aligned pathway is along a perennially
flooded conduit below the level of Spring Creek.
PALEOHYDROLOGY
Surface Paleohydrology
The Friars Hole System has a long and complex
history, and the evolution of the cave is intimately
related to the evolution of the surface overlying it. The
progressive capture of the surface creek along Friars
Hole has left the valley in a largely relict situation, allow¬
ing the evolution of the surface drainage to be inferred.
Hills Creek and its tributaries originally flowed
along the length of the Friars Hole valley and formed
a tributary to Spring Creek. As the creek eroded down¬
ward it eventually reached the Union Limestone and
was captured underground. With continued surface
erosion the sink point for Hills Creek has retreated
upstream a distance of more than 4 km, and is now
40 m lower than the original sink point. This has
formed a long inlier of Union Limestone along the
eastern part of Friars Hole, which terminates at
the original sink point, which fed the Downlets in the
Friars Hole System (Fig. 2). The elevation of the top of
the Union Limestone at the Downlets is 780 m asl,
which is the highest elevation of this contact along
Friars Hole, and this is the reason that the original sink
point was at this location.
Flow from the original sink point of Hills Creek was
to the southwest through the Friars Hole System, but
Locust Spring is a much closer spring than JJ Spring,
and offers a much steeper hydraulic gradient. As the
sink point of Hills Creek retreated to the north its loca¬
tion became closer to Locust Spring (Fig. IB), and east¬
flowing cave streams such as the one in Locust Creek
Cave must have captured at least some of the flow. At
present much of the flood discharge from Hills Creek
drains into Hills-Bruffey Cave and then east under
Droop Mountain to Locust Spring, but the low-flow
sink point of Hills Creek 500 m to the west of the flood
sink is gradually capturing the flow back toward the
Friars Hole System and to JJ Spring on Spring Creek.
The underground capture of Hills Creek at the
Downlets bisected the drainage along Friars Hole.
Upstream, the 54 km 2 Hills Creek watershed fed this
large sinking stream. Downstream, the catchment area
of 20 km 2 along the remainder of Friars Hole fed a
much smaller creek, which eventually eroded down to
the limestone contact, and was captured underground.
The process of fragmentation of the surface drainage
by capture underground has continued to the present
day, resulting in Friars Hole now being a dry valley.
However, small creeks continue to flow off the clastic
rocks on the flanks of Friars Hole, but now sink soon
after reaching the Union Limestone, which outcrops
along the axis of Friars Hole (Figs. 1 and 2).
Cave Paleohydrology
The Friars Hole System has a long and complex
history, which is not yet fully understood. Despite the
survey of almost 75 km of passages, the exploration of
many has been stopped by sediment fill, boulder
blockages, or water-filled passages. Furthermore, the
currently explored cave only spans one-third of the dis¬
tance between the major sink at Hills Creek and the
springs for the cave (Fig. 1). Clearly, there are many
tens of kilometers of unexplored cave, including both
currently active stream passages and relict passages.
Despite the only partial knowledge of the under¬
ground drainage in the area, a number of factors are
helpful in deducing the paleohydrology of the cave.
We assume that the fossil drainage was guided by the
same principles that determine present flow patterns.
Determining factors include stratigraphy, structure,
and the water table elevation. Modern cave streams
descend rapidly through the pure Union Limestone,
and have much lower gradients as they follow impure
ENCYCLOPEDIA OF CAVES
FRIARS HOLE SYSTEM, WEST VIRGINIA
339
horizons subhorizontally in the basal Union or upper
Pickaway. Low angle thrust faults are common in the
cave (Fig. 3A), with at least 5 km of major passages
being initiated along the intersection of a thrust fault
with a prominent bedding plane.
The water table elevation is important in determining
flow direction, with generally down-dip flow in the
vadose zone and strike-oriented flow in the phreatic
zone. Detailed analysis of cave streams has shown that
some were initiated solely on bedding planes and fol¬
lowed a down-dip course, and others were initiated
along joints or along the intersection of a bedding plane
and a joint or fault. The currently known cave is all
above the present water table, but the low gradient
between the lowest point in the cave and the springs
indicates that most of the intervening conduit, which
closely follows the stratal strike, is below the water table.
The largest passages in the cave tend to follow the
strike of the strata. These main drains are inferred to be
the principal passages at each stage in the history of the
Friars Flole System, and are presumed to have connected
the major sink into the cave (Hills Creek) with the spring
or springs. A number of these main drains can be identi¬
fied in the cave, and are usually found in the lower
Union Limestone or upper Pickaway Limestone
(Fig. 3C). These main drains range in elevation over
80 m, with the higher ones being found in an up-dip
location in the southeast part of the cave, and lower ones
parallel and down-dip to the northwest. This pattern is
most clearly seen in the easternmost part of the cave,
where four major, parallel main drains can be identified,
and are numbered 1—4 in Figure 3D. The first three of
these are abandoned, and the fourth is Rocky River, the
present underground course of Hills Creek.
Solutional scallops on passage walls are commonly
used to infer flow direction and discharge, but are
rarely preserved in the major passages of the Friars
Hole System, where exfoliation has been common in
the impure limestones of many passage walls.
However, scallops are well preserved in the critical
Sheepdip area (Fig. 3D, SD), where there are a series of
round, phreatic tubes in the Union Limestone. These
are unique in the cave because all other passages in
the cave in the Union Limestone are vadose canyons.
Scallop measurements in the Sheepdip area showed
that the paleoflow was to the east, rather than to the
southwest, which is the current flow direction. The
paleo water table at this time can be identified from
the transition from vadose to phreatic passages, and
was at 702 m, and this eastward flow descended to at
least 22 m below the water table.
Scuba diving in flooded cave passages in the area sur¬
rounding Friars Hole has shown that most sumps
descend just a few meters below the water table. The
interpretation of paleoflow shown in Figure 3D assumes
that most conduits were initiated below the water
table and within a few meters of it. Undulating profiles
along some of the main drains confirm that some sec¬
tions must have been below the water table. Conversely,
tall canyons have developed along some main drains,
indicating these carried vadose streams at one time.
The initial sink point of Hills Creek was at the
Downlets, and the underground course of that creek was
a large strike-oriented conduit (Fig. 3D, main drain 1).
The subsequent main drains are further down-dip, have
an average vertical spacing of 17 m, and are associated
with later sinkpoints of Hills Creek. The parallel courses
of the main drains in the eastern part of the cave make
interpretation there straightforward. However, the inter¬
pretation is only tentative for the remainder of the cave.
For instance, flow to the south at 1A and IB is less well
supported than at 1C (Fig. 3D), and any of these three
paths imply a major unexplored passage to the south,
underneath the clastic caprock, and leading to a spring
on Greenbrier River, where the Union Limestone would
have been exposed earlier than on Spring Creek. The
relationships shown for main drains 2 and 3 in
Figure 3D are even less certain, but are consistent with
passage sizes, orientations, and elevations. What is cer¬
tain is that there are unexplored continuations to many
passages, though many are undoubtedly blocked by sed¬
iment or are water-filled.
Age of the Cave
An indication of the age of the cave can be inferred
from speleothem samples taken from it. The ages of 40
speleothems from the cave have been determined by
uranium series dating, and 10 of these gave ages
beyond the 350,000-year limit of the dating method. A
sample from the Crossover Passage, a stage 2 passage
(Figs. 2 and 3D), was paleomagnetically reversed and
thus is more than 700,000 years old. A sample from
the Lost Passage, a stage 3 passage at the cave's south¬
west end (Figs. 2 and 3D), was more than 350,000 years
old and was magnetically normal, but had a U / U
ratio in secular equilibrium, a process that takes more
than 1 million years. This sample probably dates from
the Olduvai subchron between 1.72 and 1.88 million
years ago. If this is true, and assuming a steady rate of
base-level lowering, then the oldest passages in the
cave are 4.1 million years old.
Bibliography
Dougherty, P. H., Jameson, R. A., Worthington, S. R. H., Huppert,
G. N., Wheeler, B. & Hess, J. W. (1998). Karst regions of the
eastern United States, with special emphasis on the Friars Hole
Cave System, West Virginia. In D. Yuan & Z. Liu (Eds.), Global
karst correlation (pp. 137—155). Beijing, China: Science Press.
ENCYCLOPEDIA OF CAVES
340
F
Medville, D. M. (1981). Geography of the Friars Hole Cave System,
U.S.A. In B. F. Beck (Ed.), Proceedings of the 8th International
Congress of Speleology, Bowling Green, Kentucky (pp. 412—413).
Huntsville, AL: National Speleological Society.
Medville, D. M., & Medville, H. E. (1991). Structural controls on
drainage beneath Droop Mountain, West Virginia. In E. H.
Kastning & K. M. Kastning (Eds.), Appalachian karst (pp. 11 — 18).
Huntsville, AL: National Speleological Society.
Sasowsky, I. D., White, W. B., & Medville, D. M. (1989). The remark¬
ably constant longitudinal profile of Toothpick Stream, Friars
Hole Cave System, West Virginia, U.S.A. In Proceedings of the 10th
International Congress of Speleology (pp. 284—286). Budapest:
Hungarian Speleological Society.
Storrick, G. D. (1992). The caves and karst hydrology of Southern
Pocahontas County and Upper Spring Creek Valley. Morgantown,
WV: West Virginia Speleological Survey.
ENCYCLOPEDIA OF CAVES
QAMMARUS MINUS: A MODEL
SYSTEM FOR THE STUDY OF
ADAPTATION TO THE CAVE
ENVIRONMENT
Daniel W. Fong
American University
ADAPTATION TO THE CAVE
ENVIRONMENT
A diverse assemblage of animal groups, ranging from
vertebrates such as fish and salamanders, to crusta¬
ceans such as crayfish and scuds, to insects such as
beetles and springtails, to even clams and sponges,
contain species that are adapted to subterranean habi¬
tats such as caves and are never found on surface
habitats. Most of these species are strongly convergent
in a suite of morphological, physiological, and behav¬
ioral characteristics, termed troglomorphy , that are
clearly related to the dominant environmental factor in
caves: the absence of light (Culver and Pipan, 2009).
Examples of morphological features include reduced
to complete loss of eyes and body pigment, elaborated
extra-optic sensory structures such as longer antennae
and bigger or higher density of taste buds, elongated
walking appendages, and more slender body forms
compared to related surface species. The elaborated
and elongated structures are generally assumed to be
adaptations to total darkness, resulting from direc¬
tional selection for enhanced chemosensory and tactile
performance as well as for increased efficiency of
movement, although this assumption is seldom tested.
The reduction to loss of eyes and body pigment is part
of a general phenomenon of evolutionary loss of fea¬
tures common to most, if not all, phyletic lineages; the
loss of hearing in moths on islands without bats is but
one of many examples (Fong et al., 1995). Therefore the
evolution of elaborated and of even reduced features is
not unique to cave organisms. The study of the evolu¬
tionary mechanisms leading to troglomorphy in cave
species is interesting, however, because both elabo¬
rated and reduced features are exhibited by the same
organism, have evolved independently and repeatedly
across many higher taxonomic groups, and are clearly
related to the consequences of a single overriding envi¬
ronmental constraint: total darkness. It is possible that
different species have evolved troglomorphy through
different mechanisms. But part of the allure of the
study of troglomorphy is that it seems probable that
one universal mechanism is responsible for the fact
that such a diverse syndrome of superficially similar
characteristics is exhibited by a wide array of distantly
related taxa.
GENERAL HYPOTHESES ON THE
MECHANISM OF ADAPTATION TO
THE CAVE ENVIRONMENT
The search for the mechanism behind the evolution of
troglomorphy has focused mainly on the reduction to
loss of eyes and body pigment, historically referred to as
regressive evolution. The lack of emphasis on the elabo¬
rated features results from the assumption that these are
obviously adaptations to darkness, or simply more
examples of evolution by means of natural selection. The
mechanism of regressive evolution is more difficult to
elucidate, however, not because it is difficult to explain
but because many of the proposed explanations are diffi¬
cult to frame into testable hypotheses (see Fong et ah,
1995). Hypotheses on the mechanism of regressive evo¬
lution in cave species fall into three types: nonadaptive,
directly adaptive, and indirectly adaptive.
The nonadaptive hypothesis invokes as its mecha¬
nism neutral mutations and genetic drift, and attri¬
butes no role to natural selection. Simply put, features
such as eyes and body pigment serve no function in
darkness; consequently, mutations affecting structural
Encyclopedia of Caves.
341
© 2012 Elsevier Inc. All rights reserved.
342
G
or regulatory genes underlying these traits are invisi¬
ble to natural selection, or neutral, and would accumu¬
late over time. The consequence of accumulating such
mutations will eventually lead to the breakdown of the
feature. An adaptation is by definition a feature that is
an evolved response to natural selection, and strong
convergence among distantly related taxa such as the
dorsal fin in sharks and dolphins is usually a textbook
example of evidence of selection, or similar responses
to similar selective pressures. Under the neutral muta¬
tion scenario, interestingly, strong convergence for
regressive evolution among diverse taxa in caves is not
considered as similar responses to the common selec¬
tive constraint of total darkness. The critical assump¬
tion of this nonadaptive hypothesis is independence of
the genetic bases governing the development and
expression of the sets of elaborated and reduced fea¬
tures: the elaborated features are adaptations, but the
reduced features are not.
The directly adaptive hypothesis suggests that selec¬
tion favors individuals with smaller eyes or less pig¬
ment per se in darkness. The selective advantage of
having smaller eyes or less pigment is not always
clear, but is usually couched in terms of the energy or
material saved by spending less of it on a useless
structure. The nature of this energy or material, how¬
ever, is nebulous, and in any case, even if such energy
or material savings can be obtained, the selective
advantage can only be realized if the unspent energy
or material can be spent on features that increase fit¬
ness in darkness, such as extra-optic sensory struc¬
tures; therefore the directly adaptive hypothesis is
mechanistically indistinguishable from the indirectly
adaptive hypothesis (Fong, 1989).
The indirectly adaptive hypothesis assumes an evo¬
lutionary tradeoff, or a negative genetic relationship,
among components between the sets of elaborated and
reduced features through antagonistic pleiotropy.
Under this scheme, the evolutionary reduction in fea¬
tures is a correlated response to selection for elabo¬
rated features, that regressive evolution is a direct
consequence of selection for enhanced features in
darkness. Under this hypothesis, components of the
sets of reduced and elaborated features are genetically
dependent, evolving as one integrated unit.
CRITERIA FOR THE STUDY OF
ADAPTATION TO THE CAVE
ENVIRONMENT
A model system for the study of adaptation to the
cave environment must satisfy several criteria (Culver
and Pipan, 2009). The first is that phylogenetic informa¬
tion must exist to indicate the direction of evolution.
In other words, there is independent evidence that the
troglomorphic features are derived rather than ances¬
tral character states. Simply relying on a correlation
between the different morphologies of different species
in different habitats can be problematical. For example,
large eye in surface species has long been assumed the
ancestral state and small eye in cave species the derived
state among fish lineages of the family Amblyopsidae
in North America. A recently constructed phylogeny
based on nucleotide sequence variation, however, indi¬
cates that large eye in the genus Chologaster is likely a
derived condition within the Typhlicththys clade, with
absence of eyes as the ancestral condition (Dillman
et al., 2010). This criterion is extremely important
because it rules out a majority of species as model sys¬
tems because most troglomorphic species in caves are
highly endemic with no close surface relatives for com¬
parison (Barr and Holsinger, 1985), and because a rela¬
tively small number of phylogenies including closely
related cave and surface taxa currently exist. Second,
there should be data providing genetic and selective
perspectives on population structure. Data on genetic
population structure informs on the extent of genetic
isolation between surface and cave counterparts as well
as within and among cave populations. The structure of
the selective environment reflects the extent of hetero¬
geneity between the surface and cave habitats, and
equally important, heterogeneity within the cave habi¬
tat. Third, there should be evidence that both reduced
and elaborated characters show heritable variation in
the cave and the related surface taxa. If selection has
driven the presumed adaptations to fixation in cave
populations, it is difficult to demonstrate active selec¬
tion on these characters. This criterion excludes most
highly troglomorphic species because all individuals
lack any manifestation of eyes and of body pigment:
there is simply no variation. In any case, selection must
act first on the ancestral character states on initial colo¬
nization of the cave environment, and it is the features
found in the ancestral surface taxa that show the pre¬
sumptive ancestral character states. Ideally, these three
criteria are satisfied by a single species with troglo¬
morphic and nontroglomorphic populations occupying
different habitats and with data supporting troglomor-
phy as the derived condition. Finally, there should be
evidence of multiple, independent evolution of troglo-
morphy in different cave populations. The concordance
or discordance of patterns observed among indepen¬
dently derived populations of the same species is
extremely informative to our ability to generalize con¬
clusions across a large number of species belonging to a
diverse assemblage of taxonomic groups. The amphi-
pod Gammarus minus is an excellent model organism to
study the evolution of troglomorphy in light of these
criteria.
ENCYCLOPEDIA OF CAVES
GAMMARUS MINUS: A MODEL SYSTEM FOR THE STUDY OF ADAPTATION TO THE CAVE ENVIRONMENT
343
GENERAL ECOLOGY OF
QAMMARUS MIN US
The freshwater amphipod crustacean Gammarus
minus was first described by the naturalist Thomas Say
in 1818 from specimens obtained from a spring run in
southern Pennsylvania. The species has an extensive
geographic range, occupying a narrow band along the
Appalachian Mountains from southern Pennsylvania to
northwestern Alabama and the northern Mississippi,
plus much of the east central United States including
most of Tennessee and Kentucky and parts of southern
Ohio, southern Indiana, and western Illinois, as well as
across the Mississippi River west to central Missouri,
northern Arkansas, and northeastern Oklahoma (Fig. 1).
Within its range G. minus is a common inhabitant of car¬
bonate springs (Glazier et ah, 1992) where groundwater
resurge onto the surface, especially in karst areas. It is
also a common inhabitant of cave streams throughout
its range, but troglomorphic populations occur only in
some caves in two widely separated, limited geographic
areas (Fig. 1). G. minus also occurs in the streams of
some karst windows, and some karst window popula¬
tions in these two areas may exhibit morphologies that
are intermediate between spring and troglomorphic
cave populations. Figure 2 shows an idealized subterra¬
nean drainage basin and the hydrological connection
among the spring, cave stream, and karst window
habitats.
Populations of G. minus in all habitats probably feed
similarly on organic detritus but derive their nutrients
mainly from the associated microbial fauna and flora.
FIGURE 1 Distribution of Gammarus minus. Solid areas indicated
by arrows are regions with troglomorphic populations: Greenbrier
Valley (GV) in West Virginia and Wards Cove (WC) in Virginia.
Modified from Culver et ah, 1995.
Troglomorphic cave populations may also derive
much of their energy from grazing on the biofilm coat¬
ing the substratum of cave streams. Compared to
spring populations, they are also more likely to prey
on other organisms such as small worms and smaller
crustaceans such as isopods or other amphipods, and
show a stronger tendency toward cannibalism, espe¬
cially on smaller or injured individuals. Density of
spring populations may exceed 200 m -2 , but is usually
an order of magnitude lower for troglomorphic popu¬
lations in cave streams. The densities of nontroglo-
morphic populations in cave streams and in karst
windows are extremely variable but may reach simi¬
larly high values as spring populations. Similar to
most amphipods in the family Gammaridae, G. minus
is sexually dimorphic in size: the body length of
FIGURE 2 Aerial view of an idealized subterranean drainage
basin. Headwater cave streams originate at swallets (indicated by
arrows) where surface water sinks at the contact of nonsoluble rock
and carbonate rock (stippled area) or from percolating water drip¬
ping into cave passages. Cave streams are occasionally exposed at
karst windows where parts of the cave ceiling have collapsed to the
surface. Water flows through a closed conduit before exiting perma¬
nently at a resurgence, or spring, and continues as a spring run
before emptying into a surface stream. Populations of Gammarus
minus may occupy all three habitats within a basin: resurgence, cave
stream, and karst window. From Culver et ah, 1995.
ENCYCLOPEDIA OF CAVES
344
G
mature females is about 67% of mature males. During
reproduction, males and females form precopulatory
pairs in amplexus. Although amplexing pairs are
observed throughout the year, their numbers tend to
peak between December and March in spring popula¬
tions, but do not show a consistent temporal pattern in
troglomorphic cave populations, with some showing a
peak in the winter months and some not. Fertilization
is external. Females deposit eggs into a ventral marsu-
pium formed by overlapping cuticular plates, where
they are fertilized when a male deposits a sperm packet.
There is no larval stage; fertilized eggs develop directly
within the marsupium into miniature amphipods, about
1 mm in length. The newly hatched amphipods are
released from the marsupium into the environment
when the female molts. The life cycle is about one year
in spring populations and probably in nontroglomorphic
cave populations as well as karst window populations,
but may be longer in troglomorphic cave populations.
Specimens collected from any of the habitats can live for
two to three additional years in the laboratory.
VARIATION IN QAMMARUS MINUS
The spring habitats of Gammarus minus are charac¬
terized by hard, alkaline water of pH 6 or higher, con¬
ductivity greater than 100 ps cm 1 , and temperatures at
or near the average annual ambient temperature for
the latitude (10—12°C in the states of West Virginia
and Virginia). The occurrence of G. minus in cold, car¬
bonate springs is ubiquitous. Springs with the appro¬
priate physical and chemical properties but without
G. minus are extremely rare, at least in West Virginia
and Virginia. Within its spring habitat, G. minus is usu¬
ally the dominant macroinvertebrate species in terms
of density and probably biomass. Sexually mature
individuals of spring populations have large com¬
pound eyes with about 40 ommatidia, a first pair of
antennae at about 45—50% of body length, and brown¬
ish body pigmentation (Fig. 3). Body size is extremely
variable among spring populations, with body length
of mature males ranging from 6 to 11 mm. The varia¬
tion in body size among springs may be inversely cor¬
related with the intensity of size-selective predation by
small fish predators such as sculpins ( Cothis sp.).
Gammarus minus is also commonly observed in cave
streams throughout its range. The cave streams within
a subterranean drainage basin resurge onto the surface
at a spring, and the morphology of most cave pop¬
ulations is identical to or only slightly different from
that of hydrologically connected spring populations.
However, populations of G. minus occurring in some
caves in only two small geographic areas, one in the
Greenbrier Valley of southeastern West Virginia and
one in the Wards Cove area of southwestern Virginia
(Fig. 1), are morphologically differentiated from spring
populations. These morphologically distinct cave popu¬
lations of G. minus exhibit the classic troglomorphic
syndrome common to species highly adapted to the sub¬
terranean environment (Jones et al., 1993). They differ
from spring populations and nontroglomorphic cave
populations in having consistently large body size with
mature males attaining 10—11 mm in length, pale to blu¬
ish instead of brownish body coloration, longer first
antennae that are at least 65% of body length, and greatly
reduced compound eyes with only a few to no discern-
able ommatidia (Fig. 3; Holsinger and Culver, 1970).
Gammarus minus also occurs in the streams of some
karst windows. A karst window forms when a portion
of the ceiling of a cave passage collapses through to
the surface; thus the stream in a karst window was
once a subterranean cave stream. Some karst window
FIGURE 3 (A) A live, adult specimen of Gammarus minus from a
spring population, 7-mm body length. (B) Specimen from a troglo¬
morphic cave population, 10-mm body length. Notice the difference
in eye size, antennae length, and body coloration. Photo by the author.
ENCYCLOPEDIA OF CAVES
GAMMARUS MINUS: A MODEL SYSTEM FOR THE STUDY OF ADAPTATION TO THE CAVE ENVIRONMENT
345
populations of G. minus are located in subterranean
drainage basins with a troglomorphic population in a
cave stream and a nontroglomorphic spring popula¬
tion at the resurgence. Among five such karst window
populations in five different subterranean drainage
basins, one is morphologically identical to the troglo¬
morphic cave population, one is identical to the spring
population, and three are intermediate between the
cave and spring populations (Culver et al., 1995), sug¬
gesting the interesting possibility that some of the
karst window populations are evolving from the tro¬
glomorphic phenotype back toward the nontroglo¬
morphic phenotype.
POPULATION STRUCTURE OF
QAMMARUS MINUS
Although water from springs empties into base level
surface streams, often after flowing for only short dis¬
tances along spring runs, the physical and chemical
properties of the spring water differ greatly in magni¬
tude and show much less daily and annual variation
than those of the surface stream water. In addition, the
density of fish predators, such as trout, darters, and
sculpins, is much higher in the streams than in the
springs. Thus migration of G. minus individuals among
even geographically proximate springs situated along
small surface streams may be restricted, but may be
possible during winter when the temperature of the
surface streams drops to values matching those of
springs. Gene flow among geographically distant
spring populations separated by large streams, such as
rivers of the fourth order or higher, appears highly
improbable. Clearly, spring populations of G. minus
exhibit classic isolation by distance, with distance mea¬
sured as hydrological distance rather than Euclidian
distance. Among G. minus populations situated along
small streams and a large river in Greenbrier Valley,
West Virginia, significant proportions of the sequence
variation in the mitochondrial gene COI (35%) and in
the nuclear gene ITS-1 (17.2%) are explained by hydro-
logical distance (Carlini et al., 2009).
Both of the geographic areas that contain troglo¬
morphic populations of G. minus have extensive cave
development and large cave systems exceeding 20 km
in passage length. Culver et al. (1995) hypothesized
that such large cave systems are necessary for the
genetic isolation of cave populations from spring
populations to allow for the evolution of the troglo¬
morphic traits. Genetic isolation of these G. minus
populations probably arises from the fragmented
nature of the habitats. The cave streams in these sub¬
terranean drainage basins are separated from their
spring resurgences by phreatic sediments and porous
rock when the cave stream passes below the water
table (Ford and Williams, 2007). The resultant closed
conduit channels are barriers to gene flow for organ¬
isms such as G. minus. The physical barrier for
G. minus between the cave stream and the resurgence
can be rather abrupt, even in the absence of such a
closed conduit. For example, the cave stream in
McClung-Zenith Cave in Monroe County, West
Virginia, flows underneath a large breakdown pile of a
collapsed portion of the former cave ceiling at the cave
entrance for about 25 m and emerges on the other side
as a spring. This spring harbors a typical surface popu¬
lation of G. minus at a density of over 100 m -2 , yet
G. minus has not been collected in the cave stream just
a short distance away. Consequently, troglomorphic
cave populations are physically and genetically iso¬
lated from hydrologically connected spring popula¬
tions. Because different subterranean drainage basins
are hydrologically connected only through their
respective resurgences and their base level streams,
troglomorphic cave populations are also effectively
genetically isolated from each other. This is supported
by estimates of divergence among populations as indi¬
cated by Fst values obtained from sequence variation
in the COI gene (Carlini et al., 2009). For example,
water in Persinger Cave resurges at Davis Spring.
Divergence between the Persinger Cave and Davis
Spring populations is 0.78, similar in magnitude to
divergence between the Persinger Cave population
and cave populations located in other subterranean
drainage basins each associated with a different
spring: the Hole Cave at 0.83, Organ Cave at 0.69 and
0.86, and Fallen Rock Cave at 0.88.
ORIGIN OF TROGLOMORPHIC CAVE
POPULATIONS OF QAMMARUS MINUS
There is strong evidence indicating cave populations
of G. minus originated via colonization at some time in
the past by individuals dwelling at the spring where
the cave water resurges, and that cave populations in
different drainage basins are each independently
derived from a different spring population. On aver¬
age, cave populations show significantly lower within-
population sequence variation in the COI and ITS-1
genes than spring populations, and that despite the
substantial morphological difference between cave and
surface populations of G. minus, cave populations are
on average more closely related to hydrologically prox¬
imate surface populations than to cave populations in
other drainages (Kane et al., 1992; Carlini et al., 2009).
Cave populations in general show significantly smaller
ENCYCLOPEDIA OF CAVES
346
Population
FIGURE 4 The effective number of codons (ENC; Wright, 1990) in COI sequences from 15 populations of Gammarus minus. High ENC indi¬
cates small effective population size and low ENC indicates large effective population size. The mean (horizontal line), standard deviation
(box), and range (vertical bar) of ENC are indicated for each population (except for OCM where all individuals were fixed for one sequence).
Populations are grouped by drainage basin and habitat type (spring populations are designated by the letter S, cave populations by the letter
C, and the single karst window population by the letters KW). The TKW and VS populations were the only sites sampled within their respec¬
tive drainage basins, and so are not grouped with any other population. Reproduced from Carlini et al., 2 009.
effective population sizes as measured by levels of
codon bias compared to surface populations in the
same drainage (Fig. 4; Carlini et ah, 2009), indicating
they may have suffered genetic bottlenecking during
initial colonization of the cave habitat, or more probably
from periodic episodes of genetic bottlenecking events
since initial colonization such as lengthy dry spells, or
both. This is because, within a drainage basin, effects of
environmental perturbation of the physical and chemi¬
cal parameters such as discharge or temperature would
be more severe at upstream locations and be dampened
farther downstream with increasing volume of flow,
and the most downstream portion of a subterranean
basin is the resurgence at the spring. The evidence
clearly shows that the troglomorphic features are the
derived character states in G. minus. Interestingly, this
also suggests the possibility that the nontroglomorphic
features of karst window populations of G. minus are
secondarily derived character states from the ancestral
troglomorphic state in each basin. An examination of
the structure of the vestigial compound eyes shows that
individuals in four different cave populations have
consistently lost different parts of the compound eye
(Fig. 5; Culver et ah, 1995). Therefore, troglomorphy has
evolved independently in different populations of
G. minus occupying different subterranean drainages.
There is some evidence indicating genetic structur¬
ing of G. minus individuals within a subterranean
basin. Within Organ Cave, sequence variation in the
COI gene was compared between specimens from a
headwater stream at the top of the drainage (OCM)
and from the large river near the lowest accessible
point of the drainage (OCB). Although the OCB indi¬
viduals showed only a small amount of variation, all
OCM individuals were fixed for a single sequence
(Carlini et ah, 2009). The lack of genetic variation at
OCM may be explained by genetic drift. Small headwa¬
ter streams at the top of a subterranean drainage are
more likely than larger, lower level streams to periodi¬
cally dry up during prolong droughts, leading to extinc¬
tions of localized populations of G. minus followed by
recolonization by a small subset of the main population
from lower level streams when the headwater streams
resume their flow.
ENCYCLOPEDIA OF CAVES
GAMMARUS MINUS: A MODEL SYSTEM FOR THE STUDY OF ADAPTATION TO THE CAVE ENVIRONMENT
347
(B)
FIGURE 5 The compound eye structures of two cave populations
of Gammarus minus. (A) Vestigial compound eye from an Organ
Cave specimen. (B) Vestigial compound eye from a Benedict Cave
specimen. The inset in each panel shows the cross-section of the
entire head through the eyes. Vestigial eyes from other cave popula¬
tions are not shown, but different cave populations have lost differ¬
ent parts of the compound eye. Photo by the author.
QAMMARUS MINUS AS
A MODEL SYSTEM
In summary, Gammarus minus is an excellent model
system to study adaptation to the cave environment.
It is a single species with troglomorphic cave popula¬
tions and nontroglomorphic surface populations. Cave
populations, in addition to surface populations,
show heritable, individual variation in the reduced as
well as the elaborated structures (Fong, 1989).
Hydrological and genetic evidence (see Culver et al.,
1995) clearly indicate that troglomorphic populations
invaded cave streams from surface spring habitats,
thus showing the troglomorphic traits are the derived
condition. The selective environments of both surface
and cave populations are well characterized, and data
exist on the genetic population structure, both among
and within populations and habitats. Evidence from
allozyme variation and morphological variation (see
Culver et ah, 1995) as well as from DNA sequence vari¬
ation (Carlini et ah, 2009) strongly indicates that multi¬
ple cave populations have evolved troglomorphy
independently.
See Also the Following Articles
Asellus aquaticus: A Model System for Historical
Biogeography
Niphargus: A Model System for Evolution and Ecology
Astyanax mexicanus: A Model System for Evolution
and Adaptation
Natural Selection
Morphological Adaptations
Bibliography
Barr, T. C., Jr., & Holsinger, J. R. (1985). Speciation in cave faunas.
Annual Review of Ecology and Systematics, 16, 313—337.
Carlini, D. B., Manning, J., Sullivan, P. C., & Fong, D. W. (2009).
Molecular genetic variation and population structure in cave and
surface populations of the freshwater amphipod Gammarus minus.
Molecular Ecology, 18(9), 1932—1945.
Culver, D. C., Kane, T. C., & Fong, D. W. (1995). Adaptation and natu¬
ral selection in caves: The evolution of Gammarus minus.
Cambridge, MA: Flarvard University Press.
Culver, D. C., & Pipan, T. (2009). The biology of caves and other subter¬
ranean habitats. Oxford, UK: Oxford University Press.
Dillman, C. B., Bergstrom, D. E., Noltie, D. B., Holtsford, T. P., &
Mayden, R. L. (2010). Regressive progression, progressive regres¬
sion or neither? Phylogeny and evolution of the Percopsiformes
(Teleostei, Paracanthopterygii). Zoologica Scripta, 40(1), 45—60.
Fong, D. W. (1989). Morphological evolution of the amphipod
Gammarus minus in caves: Quantitative genetic analysis. American
Midland Naturalist, 121, 361—378.
Fong, D. W., Culver, D. C., & Kane, T. C. (1995). Vestigialization and
loss of nonfunctional characters. Annual Review of Ecology and
Systematics, 26(1), 249—268.
Ford, D., & Williams, P. (2007). Karst hydrogeology and geomorphology.
New York: John Wiley and Sons.
Glazier, D. S., Horne, M. T., & Lehman, M. E. (1992). Abundance,
body composition and reproductive output of Gammarus minus
(Amphipoda, Gammaridae), in ten cold springs differing in pH
and ionic content. Freshwater Biology, 28(2), 149—163.
Holsinger, J. R., & Culver, D. C. (1970). Morphological variation in
Gammarus minus Say (Amphipoda, Gammaridae), with emphasis
on subterranean forms. Postilla(146), 1—24.
Jones, R. T., Culver, D. C., & Kane, T. C. (1992). Are parallel
morphologies of cave organisms the result of similar selection
pressures? Evolution, 46, 353—365.
Kane, T. C., Culver, D. C., & Jones, R. T. (1992). Genetic structure of
morphologically differentiated populations of the amphipod
Gammarus minus. Evolution, 46, 272—278.
Wright, F. (1990). The "effective number of codons" used in a gene.
Gene, 87(1), 23-29.
ENCYCLOPEDIA OF CAVES
348
G
GEOPHYSICS OF LOCATING
KARST AND CAVES
Barbara Anne am Ende
Deep Caves Consulting South Riding, Virginia
It can be quite challenging to find caves that have
entrances, as the openings may be small, obscure, or
plugged. Detecting caves with no openings to the sur¬
face at all is that much more difficult. In addition to
explorers, developers and landowners have a pro¬
found need to understand the subsurface in karst ter¬
rains for public safety concerns to avoid or mitigate
collapse issues (Fig. 1).
On occasion, karst cavities in bedrock are acciden¬
tally encountered during well drilling. However, most
techniques are based on methods that rely on some
measurable physical property that can be sensed
remotely. It is the variations in the properties of a
void, whether it is air-, water-, or sediment-filled, such
as reflections of energy or resistivity to current flow,
that suggest the presence of a void.
Principally, the techniques used are cave radio loca¬
tion, electrical resistivity techniques, microgravity, and
ground-penetrating radar. Less useful, but possible
under limited conditions, are seismic studies, thermal
variations, and interferometry. Some people may con¬
sider dowsing to be a viable technique, but it has no
scientific validity.
For most of these techniques, there are tradeoffs.
Each setting may be interpreted better using one
method versus another. Then there are choices to be
FIGURE 1 Various subsurface targets to be detected and distin¬
guished by geophysical methods.
made, for instance, electrode configurations, frequen¬
cies, spacing of detectors, and so on. Generally, low
frequencies penetrate deeper, but do not give the
resolution of high-frequency energy systems. Careful
study of these tradeoffs is necessary before one goes
to the expense and effort of doing a geophysical
study.
MICROGRAVITY
Gravity is the attraction of one body of mass to
another. With respect to the Earth, gravitational attrac¬
tion decreases with increasing distance from the center
of the Earth. All else aside, the pull of gravity is less on
a mountaintop than in the neighboring valley. The geoid
is the surface that would represent sea level, should it
be represented across the globe. The force of gravity is
always perpendicular to the geoid surface. Bouguer
gravity anomalies are the variations in the gravity field
after corrections are made for latitude, elevation, and
terrain. These are the result in variation in density of
the rock. If a cave is present beneath the Earth's surface,
the gravity field would be less than the surrounding
area, as karst features filled with air, sediment, or water
are less dense than bedrock (Fig. 2). Considering the
radius of the Earth is more than 6000 km, the effect of
a cave a few meters tall would be very small, and
therefore such measurements fall within the realm of
microgravity studies. Indeed, the "average" gravitational
measurement is about 980 gal, whereas the anomaly
measured over passages in Maxwelton Cave, West
Virginia, ranged between 20 and 30 pgal for passages
~3 m in diameter and 10—30 m deep.
A challenge with microgravity studies is that a large
cavity at a deep depth may be impossible to distinguish
from a small void at a shallow depth. Nevertheless,
microgravity studies have successfully predicted cave
locations and studies have been conducted in karst
terrain throughout the world.
ELECTRICAL RESISTIVITY
IMAGING (ERI)
The basic technique for ERI is to insert a line of metal
electrodes into the ground. Constant current is injected
into two, and the voltage difference is measured
sequentially between the remaining pairs of electrodes.
This application of current is repeated through the rest
of the pairs of electrodes and an apparent resistivity is
calculated for each measurement. These data are then
inverted to create a resistivity model of the subsurface
generally using an iterative smoothness-constrained
least-squares technique (Fig. 3). Software, such as the
Encyclopedia of Caves.
© 2012 Elsevier Inc. All rights reserved.
GEOPHYSICS OF LOCATING KARST AND CAVES
349
Well 5 drilled Firmi y Roar „ wai!
Surface elevation Sewer
. along Traverse A-A’ .. 10m
■ Bedrock elevation determined ’***•.. j
by well 5 \ /
\ / \ /
Estimated bedrock elevation based VTy‘ (J]
. '* onmlcrogravttymeasurements jj/ j,!
' ‘ ' ■ - ■ - ■ - ■ - 1 - ■ - ■ ■ ■ - 1 ^-1 -1—-1 - 1 ■ 1 - 1 • 1 ill • 1 ■ - J —1. 1-1-.
- | 1 1 - 1 - 1 - 1 ~ -* ~ 1 - 1 - 1
Distance along traverse line
FIGURE 2 A microgravity traverse beneath a building showing the anomalies due to an underlying cavity. Traverse conducted by the
Center for Cave and Karst Studies, Western Kentucky University, Bowling Green, KY.
0 12 24 36 48 60 72
Traverse Distance (m)
FIGURE 3 Resistivity profile over State Trooper Cave, Bowling
Green, Kentucky. Traverse constructed by the Center for Cave and
Karst Studies, Western Kentucky University, Bowling Green, KY.
freeware program RES2DINV, can be used for this pro¬
cessing. Caveats about possible errors in the interpreta¬
tion are discussed in McLean and Luke (2006). In their
extensive RSI imaging at Ft. Stanton Cave, New
Mexico, transects were conducted with measurements
to ~ 50 m. Several passages were predicted and later
discovered.
High resistivity values are considered bedrock. A
water-filled void would have low resistivity, as the
natural water would conduct the current relatively eas¬
ily and salt water would conduct the current even bet¬
ter. An air-filled cavity would have a very high
resistivity. However, it should be noted that alternate
explanations can be interpreted from a line of mea¬
surements. It may be impossible to distinguish
between a single large void, or several smaller, closely
spaced cavities.
Prikryl et al. (2009) performed experiments using
ERI in karst terrain and found that after testing 5 elec¬
trode array configurations, dipole sizes, and numbers
of electrodes, the pole—dipole configuration increased
the depth that was imaged versus dipole—dipole.
The Wenner—Schlumberger configuration penetrates
deeper (>30 m) relative to dipole—dipole, but with
deeper penetration there is loss of signal strength and
resolution. Further, they found interpretations of the
data are less ambiguous with 3D versus 2D surveys,
although 3D surveys take longer and are therefore
more costly.
ERI has been used at a great number of sites through¬
out the world to determine where cave passages
might lie. The technique is subject to interpretation, but
generally considered as reliable as any other method
available for predicting the unknown.
AUDIO-MAGNETOTELLURIC
SOUNDINGS (AMT)
AMT is a technique related to EMI; instead of intro¬
ducing a current, natural-source, multifrequency elec¬
tromagnetic signals from lightning or atmospheric
disturbances produce the current. Karst features must
be at least 250 m across to be detected using AMT
(Weary and Pierce, 2009). Despite this limitation, there
are several characteristics in AMT's favor. The system
works to a depth of about 1 km. The equipment is
ENCYCLOPEDIA OF CAVES
350
G
portable; 2 ground electrodes and 2 magnetic coils
cover an area of 50 m X 50 m, which contrasts favor¬
ably with traditional ERI studies which require elec¬
trode deployment in line lengths 4—5 times traditional
ERI surveys. Two people can do 6—8 soundings per
day.
GROUND-PENETRATING RADAR (GPR)
are configured much like a lawn mower with four
wheels and an upright handle, but heavier. Other units
use flexible antenna several meters long which can be
dragged behind the operator. Generally a larger
antenna allows deeper penetration. Most of these units
can only be used on moderately level and unforested
terrain. The presence of clay in the subsurface can
degrade the ability to use GPR successfully.
Radio detection and ranging (RADAR) is the technique
for sending radio frequency energy out, measuring
the time for the return, and interpreting reflections as
objects of interest. Originally, it was developed for locat¬
ing objects, such as aircraft or ships, above ground. GPR
is the application of the technique to the subsurface.
Electromagnetic waves, generally in the VHF fre¬
quencies (30—300 MHz), are used for GPR. A transmit¬
ter sends a pulse of energy into the ground. The waves
are reflected at contacts between materials of different
physical properties. Typically this would occur at bed¬
ding planes, but in the case of karst, it could be the
wall of a cavity, whether it is air-, water-, or sediment-
filled, as each situation would result in a reflection of
the waves. An antenna receives the reflections and
records the elapsed time from transmission.
If a velocity is assumed for the bedrock/soil/sedi-
ment, the two-way travel time of the waves can be
converted to depth (Fig. 4). Errors can result from
velocity changes associated with variations in moisture
content and rock type. Calibrating the system with
known features or similar rock types can improve
accuracy.
The devices used for conducting surveys are varied.
Some are pulled by autos, others can be dragged
behind a person in a harness. Commonly, GPR units
Water
table --
Sinkhole
Cave ?
E
JZ
Q.
r 0)
6 Q
FIGURE 4 Ground-penetrating radar images of cutter and pinna¬
cle topography at a site in central Florida.
SEISMIC MODELING TECHNIQUES
Petroleum companies use seismic studies of bedrock
to image the subsurface. Conceptually, the technique is
very similar to GPR, but elastic waves are produced,
for example, through explosions or Vibroseis trucks.
Geophones are implanted in the ground along arrays
to measure the time of returns of the waves as they are
reflected off discontinuity surfaces such as bedding
planes or potentially the walls of karst features.
Typically, low frequencies are used and the subsurface
is imaged at depths from 2 to more than 10 km.
Because any feature less than the wavelength of the
seismic signal is unresolvable, this technique holds lim¬
ited use for finding caves. Paleokarst horizons may be
identified as chaotic zones due to their irregular nature
crossing bedding planes. Geotechnical engineers use
modified techniques with different wavelengths and
have the potential for detecting smaller caves at depths
less than a few hundred meters.
INTERFEROMETRY
Synthetic aperture radar (SAR) is a technique of imag¬
ing where a microwave beam (0.3—30 cm) is used from
an overhead platform. Relative motion between the
antenna and ground-based target is processed such
that it produces finer spatial resolution than basic
radar.
Interferometric SAR (InSAR) records Earth surface
deformation by recording the difference between two
or more InSAR images (Fig. 5). By using the phase of
the returned waveform, this technique can measure
centimeter-scale (or possibly millimeter-scale) changes
in land surface height. Currently, only the European
Space Agency has InSAR-capable satellites, though air¬
borne sensors also exist.
THERMAL VARIATION
Thermal infrared imaging has been used to recog¬
nize cave entrances by the thermal contrast between
the more uniform cave temperature and the ground
ENCYCLOPEDIA OF CAVES
GEOPHYSICS OF LOCATING KARST AND CAVES
351
FIGURE 5 Interferometry relies on two passes of an overhead
SAR sensor and the difference in phase to calculate centimeter-scale
changes in land surface elevation.
surface with much wider diurnal and seasonal varia¬
tions (am Ende, 2009). For this technique to be used
for caves without entrances, a cave would have to
extend upward to a very shallow depth, probably less
than a meter, for the cave air temperature to create an
anomaly on the ground surface.
This technique was used by Gunn et al. (2004) in the
United Kingdom where very old coal mines with their
entrances covered by fill were environmental hazards
due to potential collapse, noxious gas leaks, and acid
mine drainage. Airborne thermal imaging, both day¬
time and predawn, resulted in the identification of
known shafts and revealed sites that might represent
covered mines.
CAVE RADIOLOCATION
If a cave can be entered at a remote location and tra¬
versed underground to a distant site, the location can
be determined even if there is no entrance in the vicin¬
ity. For that reason, cave radio will be considered in
this discussion of geophysics, although an entrance
does exist somewhere. The primary purpose for cave
radio is to find the exact spot on the surface of the
Earth that overlies the location underground where a
transmitter has been placed (Gibson, 2010). This can be
used to correct errors with in-cave surveying by using
GPS or a better survey on the surface. Also cave radio
locations have been used to locate where a well should
be drilled to reach water or to drill or dig a new
entrance to a cave.
FIGURE 6 When deployed, the transmitter acts like a bar magnet
and creates a magnetic field (dashed lines). Ground zero lies on the
surface directly above the transmitter. The depth to the transmitter,
d, may be calculated using distance from ground zero, x, and the
angle of the field, a, at a distance, x, from GZ.
A transmitter is taken into the cave where a battery
produces an alternating current in a coil of wire
at a frequency between about 100 Hz and 10 kHz.
Traditionally, the coil is ranged in a large loop, about
0.7 m in diameter. This setup produces an alternating
magnetic field with lines of force much like those
around a bar magnet (Fig. 6). The coil must be pre¬
cisely leveled so that the vertical field is, in fact,
directly above the transmitter. In air-filled caves, this is
not generally a problem. Underwater, however, the
cave floor sediment can be soft and it is difficult to
deploy the transmitter horizontally. To solve this chal¬
lenge, Adrian Richards and Ken Smith developed a
"pinger" that is a 60 cm X 5 cm PVC tube in which a
specially designed, narrow coil is placed. Underwater,
the sealed tube, with some air inside, floats and natu¬
rally places itself in a vertical orientation, thus produc¬
ing a magnetic field in the proper orientation.
While the current is flowing and the magnetic field
is present, a person must search over the ground sur¬
face with a receiving antenna to locate "ground zero"
exactly above the site where the transmitter is located.
An amplifier converts the amplitude of the magnetic
field into sound, and the person on the surface wears
headphones while carrying the receiver. The signal can
be heard within about 50 m of ground zero and with
experience (and open ground), the site can be located
within about five minutes. Ground zero is the spot
where the receiver reads null in all directions. Depth
to transmitter can be calculated mathematically when
ENCYCLOPEDIA OF CAVES
352
G
a 45° field location is found at a measured distance
from ground zero.
Examples of using cave radio include Frank Reid's
location of a point inside Honey Creek Cave, Texas. A
shaft was drilled, successfully intersecting the cave
passage at a depth of 44 m in a direct hit. Brian Pease
pushed the limits of cave radio technology by doing
location at Culverson Creek, West Virginia. The cave
passage was approximately 200 m deep and it was
deeper than Pease could accurately calculate. Later,
Pease provided the technology to conduct 44 underwa¬
ter cave radio locations during the Wakulla 2 expedi¬
tion, most of the passages lying at a depth of 90 m.
TRANSIENT ELECTROMAGNETICS (TEM)
Using the transient electromagnetics technique, typi¬
cally, a square-shaped transmitter loop (which may be
many meters across) is placed on the ground and a
steady current applied to the loop produces a primary
magnetic field. When pulsed off, the decaying second¬
ary magnetic field is measured by sensor coils on
the ground surface. The voltage curve induced in the
receiver coil is called the transient. Resistivity in the
bedrock controls the diffusion and dissipation of cur¬
rent. Generally, a larger loop size is used for deeper
targets and the useful depths for this technique are
one to three loop diameters. Longer decay times come
from deeper structures within the Earth. Traverses are
generated by moving the system of transmitter and
sensors along a series of positions.
Interpreting the data collected through TEM is com¬
plicated. The recorded transient must be normalized
for many parameters such as current intensity, loop
size, received gain, and comparison with modeled,
uniform resistivity. The transient data may be inter¬
preted through curve matching or computer inversion.
TEM is best suited for mapping conductive targets,
such as saltwater-saturated sediments.
DOWSING
Dowsing (also known as divining or witching) is a
technique principally used for finding underground
water, particularly underground rivers. The technique
is performed by a person walking across the ground
holding a forked stick or two thin metal rods bent into
an L shape. When the person walks over underground
water, the forked stick bends downward or the metal
rods cross. No mechanism for this process is known
and its veracity is highly disputed. The scientific com¬
munity has no evidence the technique is repeatable
and testable (Enright, 1995) and therefore dismisses
dowsing as without merit.
Water is not the only matter that is considered pos¬
sible to find through dowsing. Historically, minerals
were dowsed for and recently explosives in war zones
are considered fair game. Caves may be full of water
and thus prime suspects for detecting. On the other
hand, with no clear mechanism and a handful of anec¬
dotal successes, even air-filled caves are considered by
some as subject to detection through dowsing.
Believers in dowsing have anecdotal evidence from
numerous successes. Detractors point out that the suc¬
cesses are no more significant than from random
chance or from extraneous information. Despite the
use of dowsing for centuries, we are no closer to
understanding the technique and it remains within the
realm of believers.
CONCLUSION
Cave radiolocation has been used successfully to
establish surface locations above caves many times.
For unknown cave and karst features, a variety of tech¬
niques have been used. Knowing when to use any par¬
ticular method may require a series of tradeoffs
depending on local circumstances, including what
equipment is available. Electrical resistivity studies
are, perhaps, the most commonly used technique to
locate karst features. In some studies, multiple tech¬
niques are used, which can be useful as one technique
may validate or refine the results from another.
Bibliography
am Ende, B. (2009). Discrimination between caves, overhangs, and
large vugs using long-wave infrared imaging. In W. B. White
(Ed.), Proceedings of the 15th international congress of speleology
(pp. 571-579). Kerrville, TX.
Enright, J. T. (1995). Water dowsing: The Scheunen experiments.
Naturwissenschaften, 82, 360—369.
Gibson, D. (2010). Cave radiolocation. Raleigh, NC: Lulu Enterprises.
Gunn, D., Ager, G., Marsh, S., Raines, M., Water, C., McManus, K.,
et al. (2004). The development of thermal imaging techniques
to detect mineshafts. British Geological Survey Internal Report.
IR/01/24.
McLean, J., & Luke, B. (2006). Electrical resistivity surveys of
karst features near Fort Stanton, Lincoln County, New Mexico.
In L. Land, V. W. Leuth, W. Raatz, P. Boston & D. L. Love (Eds.),
Caves and karst of southeastern New Mexico (pp. 227—232). Socorro,
NM: New Mexico Geological Society (Guidebook, 57th Field
Conference).
Prikryl, J., McGinnis, R., & Green, R. (2009). Characterization of karst
solutional features using high-resolution electrical resistivity sur¬
veys. In W. B. White (Ed.), Proceedings of the 15th international con¬
gress of speleology (pp. 597—602). Kerrville, TX.
Weary, D. J., & Pierce, H. A. (2009). Geophysical prospecting for a
major spring conduit in the Ozark karst system (Missouri, USA)
using audio-magnetotelluric (AMT) soundings. In W. B. White
(Ed.), Proceedings of the 15th international congress of speleology
(pp. 604—610). Kerrville, TX.
ENCYCLOPEDIA OF CAVES
GLACIER CAVES
353
GLACIER CAVES
Jason D. Qulley and Andrew Q. Fountain'
University of Texas Institute of Geophysics,
^Portland State University
Glaciers are some of the most rapidly evolving and
dynamic karst aquifers on the planet. Glacier caves
have frequently been classified as pseudokarst by
many geomorphologists because glacier caves form
by melting instead of dissolution. Considering that the
vast majority of glacier caves are by-products of self¬
organizing hydrological systems, they are, at least
hydrologically speaking, karst aquifers (Gulley et al.,
2009). Over the past 40 years, processes of hydrological
self-organization and cave formation in glaciers have
become the subject of considerable scientific interest
because of the discovery that hydrological processes,
including cave formation, can modulate the speed of
glaciers and ice sheets (including Greenland and
Antarctica). Understanding these complicated links
between glacier hydrology and glacier motion are
important for understanding and predicting the rate of
global sea level rise. In addition to ice melt, the mechan¬
ical transfer of ice as icebergs can also be influenced by
hydrological processes and make important contribu¬
tions to sea level rise.
INTRODUCTION TO GLACIERS
By definition, a glacier is a perennial accumulation
of snow or ice that moves. Glaciers are sedimentary
formations of seasonal snow that accumulates on the
upper elevations of a glacier and buries the previous
season's accumulation. As the annual layers become
buried in subsequent years, they become compacted,
increase in density, and eventually metamorphose
into glacier ice (Benn and Evans, 2010). In most alpine
environments within the United States, Europe, and
South America, ice is relatively close to its melting
temperature. Under these conditions, ice is mechani¬
cally soft, easily deformable, and readily flows down¬
hill. Therefore, snow that accumulates at high
elevations is transferred downhill as ice, to lower,
warmer, elevations where it is removed by melting.
Rates of ice motion and deformation are controlled by
glacier thickness and slope, and can be greatly
enhanced by meltwater that is either produced at or
delivered to glacier beds. In subpolar and temperate
regions of the globe, ice at the melting temperature
allows a film of water to exist at the bottom of a
glacier. Additionally, surface meltwater can penetrate
to the glacier bottom where it may pressurize the gla¬
cier bed. In both instances, water decreases friction
between the ice and its substrate, allowing the glacier to
move faster by sliding than without water. The volume
of water at glacier beds is an important factor in deter¬
mining the degree of lubrication and the amount of slid¬
ing. In polar regions, where temperatures are extremely
cold and little or no melting occurs, glacier motion is
very slow and caused entirely by ice deformation.
Those unfamiliar with glaciers frequently assume
that glaciers are very cold, freezing any water on con¬
tact. In fact, many glaciers in the world are considered
"warm," or "temperate," meaning the ice is at the melt¬
ing temperature. For glaciers in the western United
States and Europe, for example, it is common to see
streams flowing across the glacier surface before plung¬
ing into the ice interior via an opening in the ice such as
a crevasse (fracture on the ice surface) or a moulin
(melt-enlarged vertical shaft in a glacier). Under these
conditions water coexists with ice and the two phases
are close to thermodynamic equilibrium and even small
additions of heat, such as the flux of geothermal heat
from the Earth, can melt the ice at the bottom of gla¬
ciers. Temperate glaciers exist in regions where the
average annual temperatures are not much below 0°C.
These glaciers exist in the temperate and tropical zones
of the Earth, including Europe, North America (exclu¬
sive of the Canadian Arctic and the Brooks Range,
Alaska), and the glaciers of Africa, South America,
New Zealand, New Guinea, and most in central Asia
including the Hindu Kush-Himalaya region.
In areas where average annual temperatures are
well below freezing, the temperature of glacier ice is
also below freezing for much of the ice thickness.
These glaciers are restricted to the polar regions ( e.g .,
Antarctica) or high altitudes (e.g., Alaskan peaks).
If the entire thickness of the glacier is below freezing,
the glacier is frozen to its bed. These glaciers are called
polar or cold-based glaciers and they flow more slowly
than their temperate counterparts because all motion is
accomplished by internal ice deformation. Although
the surface may melt for a short period in summer,
water is found only in or under the ice in restricted cir¬
cumstances (see "Subglacial Caves" section, below).
In subpolar conditions, such as Arctic Sweden, poly-
thermal glaciers may form where the upper portion
of the ice is below freezing, but ice near the bed is
warm. Another condition exists for the two ice sheets
(Greenland and Antarctica) whereby the entire thick¬
ness of the ice is below freezing except at the very bot¬
tom where water may be present. This is partly due to
the insulation against the polar cold provided by the
great thickness of the ice, geothermal heat coming
from the Earth's interior, and heat generated by ice
deformation. Also, the melting temperature of the ice
is lower than 0°C due to the great pressure of the over-
lying mass.
Encyclopedia of Caves.
© 2012 Elsevier Inc. All rights reserved.
354
PROCESSES OF GLACIER
CAVE FORMATION
G
ENGLACIAL CAVES
Caves are common to all glaciers that experience
melting. Caves that are entirely enclosed in the main
body of the glacier are classified as englacial caves,
whereas caves at the glacier bed are subglacial caves.
Regardless of where they are located, caves primarily
form in glaciers by melting. While many of the most
well-known glacier caves have formed in glaciers on
the flanks of volcanoes as a result of geothermal heat
and steam vents, most glacier caves are formed by a
less obvious source of heat—friction. All flowing water
generates heat through friction, but the rate of heat
production is quite small relative to other heat sources,
such as sunlight or air temperature, and self-heating is
typically ignored in most hydrological applications.
However, for water flowing in a glacier, there are no
other sources of heat and because of the close thermo¬
dynamic equilibrium between ice and water, any fric¬
tional heat produced is consumed by melting ice.
Consequently, passageways (caves) in ice enlarge in
summer when surface melt is the greatest and the
water discharge in the conduits is the largest.
Opposing this enlargement is ice creep, acting under
the weight of ice above the passage, which squeezes
the passage closed. As long as the water can melt the
walls, the passage remains open. Larger passageways
drain water from smaller passageways and a network,
much like a dendritic cave system, forms in and under
temperate glaciers. In winter, the surface of the glacier
is typically covered with a cold blanket of snow and
the network is starved of water. The rate of ice creep
dominates melt enlargement and the passageways
close. Near the glacier surface, passageways close
more slowly than those in deep ice because the ice
pressure is reduced close to the surface.
The processes of cave formation in glaciers share
many similarities with that in limestone. Caves form
along hydrologically efficient flow paths, networks of
fractures, or bedding planes that connect a water
source to a discharge point (Gulley et ah, 2009). In gla¬
ciers, a highly permeable feature may be a crevasse.
Ultimately, the morphology of caves in glaciers is
determined by the type of glacier (polar or temperate),
the crevasses that are present, and the magnitude of
meltwater discharge. Generally, increasing the number
of crevasses on a glacier reduces the meltwater dis¬
charge in any one crevasse, thus reducing its capacity
for possible enlargement. In the following section we
discuss the broad categories of englacial caves that
form in different glaciers and then discuss subglacial
caves.
The formation of englacial caves typically requires a
network of fractures to melt-enlarge into caves.
Fractures may include surface crevasses, thrust faults,
or relict crevasses/faults that become separated from
their zone of formation due to ice flow (Gulley et ah,
2009). Surface crevasses drain meltwater and rain from
the glacier surface into the glacier interior. Where a
surface stream drains into a crevasse, a moulin forms.
The crevasse may close due to the ice advecting into a
region of compression, but as long as the stream flow
is maintained, the moulin will survive. In thin regions
of a glacier the moulin may reach the glacier bottom.
French speleologists and glaciologists have made the
deepest known descent of any moulin, reaching a
depth of 205 m below the ice surface of the Greenland
Ice Sheet (Badino, 2007).
Thrust faults can also be exploited by water.
Such faulting occurs in glacier termini when subjected
to extreme compression. A glacier thins toward the ter¬
minus, causing the ice to slow, and the thicker, conse¬
quently faster, ice upstream continues to push, causing
compression. When the compression is sufficiently
large, the ice fractures such that the upper ice moves
over the deeper ice. Low wide caves can form along
these faults as pressurized subglacial water discharges
along the thrust plane. Caves have been mapped along
thrust faults in the polythermal Aldegonda Glacier,
Svalbard, Norway and the Matanuska Glacier, Alaska,
U.S.A. (Fig. 1).
On glaciers where crevasses may be absent from
large areas of melting ice, caves may form from the
incision of surface streams. The streams initially form
slot canyons. Provided the rate of stream incision
FIGURE 1 Englacial conduit formed from surface water exploit¬
ing a thrust fault, Matanuska Glacier, Alaska, U.S.A. Photo by Jason
Gulley.
ENCYCLOPEDIA OF CAVES
GLACIER CAVES
355
FIGURE 2 A shallow cut-and-closure cave in the debris-covered
Khumbu Glacier, Khumbu Himal, Nepal. This portion of the passage
is beneath approximately 20 m of ice. Photo by Jason Gulley.
exceeds that of overall surface melt, closure of the
upper parts of canyon walls by ice creep forms caves
(Fig. 2). These types of caves are cut-and-closure caves
and the environmental requirements for their forma¬
tion generally restrict them to polar or polythermal
glaciers or debris-covered glaciers. Polar glaciers have
slow rates of ice motion and have few, if any, cre¬
vasses. Surface streams have large catchment areas
and have high discharges per unit area where they are
funneled through topographic depressions or along
the margins of glaciers. Because stream discharge, and
hence melt, is concentrated, streams can incise into the
ice much more rapidly than the surface is lowered by
melting. On debris-covered glaciers, surface debris
insulates the underlying ice and melt is concentrated
in stream channels. Creep closure of the canyon walls
isolates the stream from the glacier surface and forms
a cave. The effects of creep closure are more obvious at
increasing ice depths and create a pronounced suture
i
FIGURE 3 A prominent canyon suture formed from ice pressure
squeezing the canyon walls shut is evident in the passage ceiling.
This portion of the passage is beneath about 15 m of ice. Longyear
Glacier, Svalbard, Norway. Photo by Jason Gulley.
in the ceiling running parallel to the passage (Fig. 3).
The depths of cut-and-closure caves are ultimately lim¬
ited by the rate of creep closure relative to the rate of
enlargement. If a cave incises to an ice depth that is
capable of closing a conduit by creep over winter, the
next season's meltwater will back up in the cave and
discharge through any available holes in the cave roof.
If no such holes are available, the cave will fill
completely with water and freeze shut and a new cave
will begin forming nearby. Cut-and-closure caves
include the longest mapped englacial cave—in the
Scott Turner Glacier, in Svalbard, Norway, >2 km long
with a total elevation change >100 m. This cave is
shallow, never more than a few tens of meters below
the ice surface, but has incised to the glacier bed to cre¬
ate a subglacial passageway near the glacier terminus.
On stagnant ice, that which is no longer moving,
glacier karst can readily develop because the deforma¬
tion imposed by ice movement is absent and increases
the lifetime of englacial and subglacial passageways.
Sinking streams, closed depressions, and lakes that
drain rapidly are common features of stagnant ice.
On debris-covered glaciers, such as those in the
ENCYCLOPEDIA OF CAVES
356
G
FIGURE 4 Phreatic-vadose transition in a cave formed by lake
drainage in the Ngozumpa Glacier, Khumbu Himal, Nepal. This por¬
tion of the cave is beneath approximately 20 m of ice. Photo by Jason
Gulley.
FIGURE 5 An N-channel incised into frozen till beneath the thin,
cold-based Rieper Glacier, Svalbard, Norway. This portion of the
cave is beneath 10—15 m of ice. Photo by Jason Gulley.
Himalayas, lakes may form in topographic depressions
on the glacier surface and drain through caves. These
caves are often nearly horizontal and form along per¬
meable debris-filled structures in the ice that connect
lake basins. Often these caves exhibit a keyhole cross-
section (Fig. 4), indicative of initial water-filled condi¬
tions that melt a circular cross-section. Water empties
from the lake and the cave no longer flows at roof-full
condition. Water in the partly full cave no longer melts
the roof and only melts the side and bottom, incising
the cave downward until the lake empties.
Water-filled surface crevasses may penetrate deep
into a glacier by hydrofracturing due to the density
difference between water and ice. A water-filled cre¬
vasse exerts an outward pressure on the walls that
is greater than the pressure exerted by the ice.
Hydrofracturing forms deep, fan-shaped cracks that
may later be altered into a more circular morphology
by concentrating water flow in along one path and
creep closure of marginal areas of the fracture
exposed to less water flow. Exploration of hydrofrac¬
tures in ice that is less than 150 m thick has led to
the discovery of subglacial cave systems in several
glaciers in Svalbard, Norway.
SUBGLACIAL CAVES
Subglacial caves form where englacial passageways
reach the glacier bed or where surface stream or lakes
sink along the ice-bed interface at ice margins. The
common form of subglacial cave is one that is incised
upward into the ice. These form like englacial caves
through the frictional heating of the water that, in this
case, melts the ice roof. Less common are those that
are cut into the bedrock with an ice roof (Fig. 5).
Because much of the extent of subglacial caves is diffi¬
cult to access physically, much of what is known has
been inferred from indirect measurements such as dye
tracing, geophysical measurements, and geochemistry.
Dye tracing has shown that the spatial pattern of sub¬
glacial caves is dendritic. Meltwater sinks into the
glacier surface at many points, primarily through cre¬
vasses or moulins, yet discharges from the glacier at a
few discrete exits. Locally, however, caves may be
anastomotic. Subglacial caves near the terminus, along
the ice margin, and in rare instances where subglacial
caves are accessible via moulins have been observed to
form in the same locations, with very similar morphol¬
ogies, each year.
Subglacial caves can form in polar glaciers that are
frozen to their beds from cut-and-closure caves that
have incised to the glacier bed. Under thin ice, subgla¬
cial caves formed by cut-and-closure caves that reach
the bed are frequently incised in underlying till or
soluble/erodible rock. Subglacial caves can also form
from hot gases that escape from volcanoes covered by
glaciers. Caves formed by escaping volcanic gases are
concentrated around fumeroles where gas escapes and
passageways are organized to efficiently discharge these
escaping gases. Such caves are common to glaciated
volcanic peaks, such as the Cascades (United States)
and Antarctica, and to volcanic terrain in Iceland.
Ice margin caves are commonly seen by even the
most casual tourist. The location where the glacial
stream exits the glacier is an ice marginal cave. Ice
marginal caves also form where streams from valley
walls enter glaciers or where streams from the surface
of the glacier sink along the contact of the ice and
valley wall. In both cases, they provide some access
to the glacier interior that is frequently limited by
flooding or ice thickness.
ENCYCLOPEDIA OF CAVES
GUANO COMMUNITIES
357
EXPLORATION OF GLACIER CAVES
Glacier caves are typically entered soon after
surface melting has ceased sufficiently to permit safe
exploration. However, once melt-enlargement ceases,
the caves begin to close rapidly such that for ice depths
greater than 150 m the caves are frequently closed by
ice creep before they become sufficiently dry. Caves at
depths between 80 and 150 m require exploration as
soon as surface melting has stopped.
There are many hidden hazards in glacier caves.
The most obvious glacier cave entrances are often
where glacier streams discharge from glacier fronts;
however, these entrances are often wide and covered
by thin ice, making them prone to frequent collapse.
Rainfall or an increase in surface temperatures can
cause glacier caves to flood rapidly and catastrophi¬
cally. In cold glaciers, refreezing of the surface of
these floodwaters can create dangerous false floors
when the water level recedes again. Special caution
should be used when exploring volcanic glacier caves
because they can trap deadly gases and have been
implicated in a number of deaths in ice caves. Glacier
caves combine all of the hazards of glacier travel
with all of the hazards of alpine caving. When
approached with experience and a healthy amount
of common sense, glacier cave exploration can be
fairly safe.
Bibliography
Badino, G. (2007). Caves of sky: A journe\/ into the heart of glaciers. :
Graffice Tintoretto (TV).
Benn, D. I., & Evans, D. J. (2010). Glaciers and glaciation. New York:
Oxford University Press.
Gulley, J., Benn, D., Screaton, E., & Martin, J. (2009). Mechanisms of
englacial conduit formation and their implications for subglacial
recharge. Quaternary Science Reviews, 28, 1984—1999.
GUANO COMMUNITIES
Pedro Qnaspini
University of Sdo Paulo, Brazil
GUANO AND ITS IMPORTANCE TO
CAVE COMMUNITIES AND TOSCIENCE
Guano is the accumulation of feces of particular ani¬
mals in time and space. A commonly known example
is seabird guano, which is the accumulation of drop¬
pings of large colonies of seabirds. This has been
exploited for many years as a source of nitrogen com¬
pounds. In caves, guano also occurs, and it comes
from three main different animal sources: birds, bats,
and crickets. It should be stressed that the animals that
add guano into the caves are always recognized as
trogloxenic animals. Bat guano has also been exploited
in the past, especially for the production of gunpowder.
The cave ecosystem lacks primary producers and
completely depends on energy that comes from the
outside; hence, it is a food-poor environment. Actually,
the food supply in caves tends to be low or sporadic in
quantity and poor or variable in quality. It is also clear
that any accumulation of food inside caves (as is a
guano pile) is of great attraction to detritivorous scav¬
enger animals. Other common examples of energy
sources in caves include dead animals and debris
wasted in from the surface (by water, gravity, etc.).
There are some caves where the accumulation of
guano is so large that the caves are considered to be
food-rich places, the so-called eutrophic caves. These
caves are more common in tropical countries but they
occur in all continents. For example, many caves in
Mexico and Brazil are also eutrophic, although in these
countries there are many more caves that are food-poor
(poecilotrophic). In some of these food-poor caves bat
guano deposits are scattered and do not constitute the
main or only base of the trophic chain. Nevertheless, it
is a very important food source. Eutrophic caves tend to
occur where there is an enlarged chance of bat aggrega¬
tion due to the biology of the bats or to physical charac¬
teristics of the cave areas. For instance, in Brazil,
limestone caves tend not to be eutrophic because they
are generally larger and occur in large number in partic¬
ular areas, offering more shelters for bats. In turn, sand¬
stone caves are smaller and occur in smaller numbers
leading to the concentration of bats in the same shelters,
and, therefore, to the concentration of bat guano.
Nevertheless, Poulson and Lavoie (2000) also considered
that bat and cricket guano provide the most important
biotic input of energy to hypogean environments in tem¬
perate zones.
Scattered feces (even of other vertebrates, such as
cats, pumas, dingoes, foxes, otters, rats, agoutis, and
didelphids) may also be used as food by cave animals.
In some cases it may be preferred because such ani¬
mals would avoid the high predation risk that they
would be subjected to in guano deposits, where large
communities may be established. An interesting exam¬
ple of association to feces of other vertebrates was pro¬
vided by Calder and Bleakney (1965), who noticed that
in Nova Scotia bats are not important, and the cave
fauna depends mainly on porcupine dung.
In addition to the availability of food, bats may
have a large influence on the cave climate. In many
caves with a higher concentration of bats, especially
considering large chambers connected to other points
of the cave by narrow/low conduits, temperature
Encyclopedia of Caves.
© 2012 Elsevier Inc. All rights reserved.
358
G
tends to be very high, reaching (and sometimes sur¬
passing) 40°C. Concentration of ammonia is also very
high, making these places very difficult for human
exploration. These are considered bat caves.
Because guano communities are somewhat isolated
and, in general, present a simple structure, they are an
important field for ecological studies. Moreover, the
physical and nutritional properties of the guano vary
according to the kinds of animals producing the piles;
namely, insects, birds, or bats (in turn with several
kinds of feeding preferences) providing an additional
field for comparison of community structures.
Guano has been studied in different places and with
different foci, including mineral and chemical analysis
of bat guano, mathematical simulation of community
dynamics, and extensive faunistic surveys. Guano
communities have been the subject for ecological stud¬
ies in several caves around the world, but more com¬
monly in tropical than in temperate caves, contrary to
what would be expected from the higher development
of temperate speleobiology. Theoretical definitions
about guano inhabitants are based mainly on papers
by Decu and colleagues (focusing on Eastern European
caves) and Poulson and colleagues (relative to North
American caves). Both of these groups started their
research on guano fauna in the 1960s while Gnaspini
and colleagues (focusing on South American caves)
started their research in the 1980s. Their ideas are sum¬
marized in papers by Poulson (1972), Decu (1986), and
Gnaspini (1992), and a general review can be found in
Gnaspini and Trajano (2000).
TYPES OF GUANO
Guano in caves comes from three main different
animal sources: birds, bats, and crickets. Guano of rha-
phidophorid crickets represents a large amount of
food input for cave animals in temperate caves, espe¬
cially those in the United States. Some trogloxenic
crickets ( Hadenoecus subterraneus) leave the caves for
food periodically, and when they come back they
deposit a large amount of feces based mainly on vege¬
tal matter. Cricket guano may also be found in Asian
caves, but, in this case, it should be considered a
by-product of bat guano because these crickets feed
mainly inside caves on the guano of other animals
(Deharveng and Bedos, 2000).
Several birds, such as doves, owls, parakeets, swal¬
lows, swiftlets, and vultures, may nest near cave
entrances, and their droppings may be used as a food
supply in the twilight zone. In some tropical caves, how¬
ever, bird droppings such as those from oilbirds (gua-
charos, Steatornis caripensis ; northern South America and
Caribbean) and swiftlets ( Aerodramus sp.; tropical
oriental caves) may be an important food source. They
colonize the dark zones, and their droppings may form
large piles. Swiftlets feed on insects and are common in
Southeast Asia where their nests are exploited for a tra¬
ditional dish in human cuisine. Oilbirds are common in
northern South American caves, occurring in northern
Bolivia, Colombia, Guyana, Peru, Venezuela, and on the
island of Trinidad. They are known to navigate based
on sound, like bats, but their clicks can be heard by
humans, unlike those of bats. Therefore, they can be
found deep inside caves. They feed mainly on palm
fruits, which they carry inside the caves. Consequently
their guano is comprised of both rejected seeds (with
some pulp adhering) and droppings.
The most important source of animal droppings in
caves throughout the world, however, is bat guano. In
the northern temperate regions and other parts of the
old world, most cave-dwelling bats have an insectivo¬
rous diet. This type of guano is generally finely granu¬
lated and consists of small pieces of insect cuticle,
sometimes with some material still adherent.
In some parts of the old world, for instance, in
African caves, frugivorous bats also deposit guano.
This kind of guano is generally granulated and con¬
sists of seeds and decaying plant matter (Fig. 1). It is
similar to oilbird guano regarding physical and nutri¬
tional characteristics. It is in the Neotropics, however,
that the feeding habits of bats (especially within the
Phyllostomidae) are very diversified. Frugivorous and
insectivorous bats also occur. The most common in
inhabited areas may be the vampire, or hematopha-
gous, bats (especially Desmodus rotundas). Owing to
the high availability of prey in the form of domestic ani¬
mals, and roosts not being a limiting factor in cave
areas, the populations of these bats are very high
FIGURE 1 Close-up of a frugivorous bat guano pile, showing a
Maxchernes pseudoscorpion and the large number of seeds with
adhered digested material. From Gnaspini and Trajano, 2000.
ENCYCLOPEDIA OF CAVES
GUANO COMMUNITIES
359
FIGURE 2 Overview of a somewhat old and dry hematophagous
bat guano pile.
(Trajano, 1985). Their guano is a very dark and smelly
paste (Fig. 2). Although they form small colonies (up to
ten individuals), large carnivorous bats may also
deposit large amounts of guano. These animals feed on
birds and small mammals and occasionally on large
insects, and their guano is formed of large, somewhat
compact droppings full of digested animal material
mixed with feathers and/or hairs (Fig. 3). Although nec-
tarivorous bats (Glossophaginae) are relatively frequent
in caves, guano piles from these bats have not been
identified thus far, probably because the colonies are in
general small and, on account of its liquid nature, the
guano would mix with the substrate.
Sometimes, in the same cave one can find more
than one kind of bat guano pile with some maybe
even mixed with each other. The texture and nutri¬
tional characteristics differ among the various types of
bat guano and these different characteristics will influ¬
ence the associated communities.
Other important characteristics influencing the estab¬
lishment of associated fauna are the substrate and the
place where guano is deposited, because they would
influence the availability of the organic matter (see
Poulson and Lavoie, 2000). For instance, if guano is
deposited on sandy substrates and/or under dripping
water, nutrients may be washed away more rapidly.
THE FAUNA ASSOCIATED WITH GUANO
DEPOSITS—DEFINITIONS
Considering that guano is an important source of
food in caves, there are many animals that directly
feed on this substrate and/or on microorganisms ( e.g .,
bacteria and fungi) that grow on guano. These animals
are called guanophages. To complete the guano commu¬
nities, there are also several predators that feed on the
FIGURE 3 Close-up of a carnivorous bat guano pile, showing a
large concentration of cholevine larvae (at least two species of the
genus Dissochaetus). From Gnaspini and Trajano, 2000.
guanophages. Deharveng and Bedos (2000) also recog¬
nize two types of associated communities: the giant
arthropod community (GAC) and the meso- and
microinvertebrate community (MIC). The latter is
always on or close to guano deposits and formed by
large numbers of species of guanophages and their
predators, and the former consists mainly on one species
of guanophages and mostly on predators, which may
live far from the guano piles and come closer to feed.
Due to the specificity of these communities to the
guano, some authors (see, e.g., Decu, 1986) have
argued that these animals should not be considered
cavernicoles, because they occur in caves just after this
specific source of food, and not because of the cave
environment. Therefore, these would not depend on
caves from an ecological point of view. Because this
would mean that (1) guano inhabitants colonize caves
only because of guano, and would not be there if that
food supply was missing, and (2) they do not depend
on the cave ecosystem nor are they subjected to cave
environmental constraints, Gnaspini (1992) discussed
that, in order to reach and colonize the guano piles fre¬
quently deposited far from known cave entrances,
organisms should be able to orient themselves and
also survive inside the cave. In other words, they must
have an ecological relationship with the caves and
should be considered cavernicoles. In addition, if true,
the same argument to exclude guano dwellers should
have been used to exclude other communities associ¬
ated with specific sources of food in caves. This would
be the case, for instance, in the animals that feed spe¬
cifically on roots in caves, which are an important part
of the troglobitic fauna of Hawaii. Therefore, guano
piles should simply be treated as a substrate, and, as
ENCYCLOPEDIA OF CAVES
360
G
any substrate inside a cave, it would be colonized by
cavernicoles.
Because guano is an important source of food, and,
sometimes, the only source of specific types of food
(such as digested blood or vegetal matter, especially
fruits and seeds), there are some cave-dwelling ani¬
mals that are restricted to guano piles, and some that
occasionally visit this substrate to feed directly on
guano or to prey on guano dwellers. Therefore, follow¬
ing the same rationale of "affinity to the environment"
used to classify cave animals into troglobites, troglo-
philes, and trogloxenes, cavernicolous guano dwellers
can be classified as guanobites, guanophiles, and gua-
noxenes. It should be reinforced that this classification
should be used only for animals inhabiting guano
deposits inside caves.
Guanobites are organisms that, when in caves,
exclusively inhabit guano deposits, and whose entire
biological cycle takes place in this substrate. It should
be noted that guano deposits are discontinuous in
caves, and hence guanobites may be found on any
cave substrate as they move through the cave environ¬
ment to colonize guano piles. Guanobites, however, do
not reproduce or feed in these substrates. Guanophiles
may inhabit and reproduce both in guano piles and in
other substrates of the cave environment. Guanoxenes
may be found feeding and/or reproducing on guano
deposits but depend on other substrate(s) in the caves
to complete their biological cycle. Although theoreti¬
cally possible, this third type seems to be unlikely in
practice.
It should be noted that the relationships with the
cave and with the guano are not necessarily interde¬
pendent. The degree of fidelity to, or dependence on,
the cave environment is neither influenced by nor
influences the relationships of the cavernicoles to the
guano (i.e., their degree of fidelity to the guano envi¬
ronment). For example, an animal restricted, in caves,
to guano piles (thus, a guanobite) may occur in the
epigean environment as well, in this case not being a
troglobite, as expected if the two classifications were
correlated. Therefore, the guano inhabitants should be
classified independently according to their cave and
guano relationships. When analyzing the cave relation¬
ships of an organism, one must take into account its
possible occurrence in the epigean environment, as
well as the possible modifications acquired during col¬
onization and isolation in the hypogean environment,
in order to classify it as a trogloxene, a troglophile, or
a troglobite. Independently, one may also consider
whether or not an organism inhabits substrates other
than guano inside the cave.
The entire biological cycle of a guanobite takes place
in the guano piles inside caves, and, therefore, entirely
in the cave environment. Thus, a guanobite may be
either a troglobite or a troglophile, because both can
complete their entire cycle inside caves. It can never be
classified as a trogloxene, because this group must leave
the caves (and therefore any substrate inside the caves)
to complete its life cycle. This applies also to guano¬
philes, because at least the populations observed in the
guano piles complete their life cycles inside the cave.
Finally, guanoxenes can be trogloxenes, troglophiles, or
troglobites. It is noteworthy that a classification combin¬
ing troglo- and guano-related classifications is very use¬
ful because it gives a precise notion of the ecological¬
evolutionary relationships of the organisms studied.
Troglomorphisms generally are adaptive responses
to the general low predictability and availability of
food in the subterranean environment. Therefore, it
would not be expected to find troglomorphisms in spe¬
cies associated with guano piles (especially not among
guanobites), because these animals would be living
under a large food availability. However, some cases
seem to occur.
COMMUNITY STRUCTURE AND
SUCCESSION IN BAT-GUANO PILES
Like all communities associated with the cave envi¬
ronment, guano communities are simple and com¬
posed of only a few trophic levels. Two levels can be
recognized: the guanophages and their predators. In
turn, the relationships between the animal components
are very complex, because most predators would feed
on any of the items in the guanophage level, and
maybe even within the predator level. Therefore, the
trophic relationship within this community should be
viewed as a food web instead of a food chain.
In temperate regions, the seasonal activity of cave
bats leads to seasonal changes in the structure of
guano piles and, consequently, in the associated fauna
(Poulson, 1972; Decu, 1986; Poulson and Lavoie, 2000).
In general, these bats are active during the summer
and hibernate (in the same or in other caves) in winter.
When bats are present and active, fresh guano accu¬
mulates continuously. As bats reduce their activity or
leave the sites, there is a progressive desiccation of
the deposits, promoting chemical and microclimatic
changes. The microclimatic properties of the guano
depend on the depth and the distance from the center
of the deposit. Because several taxa show microcli¬
matic preferences, a faunistic succession is observed in
the guano deposits. Guano piles are initially alkaline
owing to the concentration of ammonia in the feces.
They subsequently become acid as a result of the
decomposition of alkaline compounds and fermenta¬
tion of the piles.
ENCYCLOPEDIA OF CAVES
GUANO COMMUNITIES
361
On the other hand, Decu (1986), based on the litera¬
ture and his own observations in Cuba, stated that
bats are permanently active in tropical caves. As the
guano input is continuous, there would be no seasonal
differences in the structure and composition of the
guano deposits and the associated fauna. In both cases
guano deposition is predictable, but, whereas in tem¬
perate bat caves deposition is seasonal, in tropical bat
caves such deposition is continuous. This led Decu
and Tufescu (1976) to characterize caves as belonging
to a temperate type when there is seasonal succession,
and to a tropical type when there is not, irrespective
of their geographic location. However, Poulson and
Lavoie (2000) consider that the predictability of bat
guano availability in temperate caves is much smaller
than in tropical caves.
In the Brazilian karstic areas so far studied, continu¬
ous guano deposition usually does not occur because
roost availability is high and the bats form itinerant
colonies, frequently moving between and inside caves.
Therefore, guano piles show temporal changes in these
caves, passing through the same phases of deposition,
and subsequent desiccation after the bat colony leaves
the sites. In contrast to temperate caves, succession in
Brazilian caves is not necessarily (if at all) seasonal. In
fact, the temporal and spatial deposition of guano may
be highly unpredictable. Neotropical bats show regular
activity throughout the year, but can leave a specific
roost at any time, causing the interruption of guano
deposition and consequent succession. Therefore, from
whichever cave region, piles of bat guano are "qualita¬
tively and microclimatically nonuniform ... [and] ...
look like a mosaic of microhabitats that shelter a
mosaic of zoological communities in various evolu¬
tional stages" (Decu, 1986).
THE ASSOCIATED FAUNA—EXAMPLES
As expected, a large variety of animals inhabit guano
piles. A huge list of papers citing communities associ¬
ated with bat guano can be found in the literature, but it
is not the goal of this chapter to present an extensive
review of the available data. For a review and discussion
of guano dwellers in tropical caves, please refer to
Gnaspini and Trajano (2000) and Deharveng and Bedos
(2000). For temperate caves, see, for example, Calder and
Bleakney (1965), Negrea and Negrea (1971), Decu and
Tufescu (1976), and the review by Poulson and Lavoie
(2000). When a broad overview on the guano communi¬
ties is made, we could recognize that some taxa occur
almost everywhere in the world, constituting the basis of
most guano communities. This is especially the case, of
course, when the family level is analyzed: generic and
specific levels would give more differences and would
allow a better biogeographic comparison. These are the
"basic" taxa ( sensu, e.g., Negrea and Negrea, 1971).
In addition, there are some less common but still fre¬
quent groups, which could be called "accessory" taxa
(also sensu Negrea and Negrea, 1971). Of course, faunal
regionalism would also interfere with the taxa observed
in the different places. For instance, although crickets
are commonly associated with guano, rhaphidophorid
crickets occur in North American, European, and
Australian caves, whereas phalangopsid crickets occur in
African and South American caves.
Good examples of basic taxa are laelapid, macrochelid,
oribatid, and uropodid mites; chernetid pseudoscorpions;
entomobriid, hypogastrurid, and isotomid springtails;
tineid moths; carabid, cholevine, histerid, and staphylinid
beetles; and drosophilid, fanniid, milichiid, and phorid
flies. Examples of accessory taxa are planarians; earth¬
worms; other pseudoscorpion families (Cheliferidae,
Ideoroncidae, Neobisiidae); harvestmen; spiders; schizo-
mids; isopods; millipedes; chilopods; psocopterans;
cockroaches; crickets; several other beetle families
{e.g., Aderidae, Curculionidae, Dermestidae, Elateridae,
Hydrophilidae, Nitidulidae, Ptiliidae, Tenebrionidae,
Trogidae); and several other dipteran families {e.g.,
Ceratopogonidae, Dolichopodidae, Empididae, Musci-
dae, Sciaridae, Sphaeroceridae, Stratiomyiidae).
As expected, mites (Acari) are by far the most diver¬
sified group of guano inhabitants. Guano-inhabiting
mites include guanophages and predators. Argasid ticks
may occasionally be found on guano, but they should
not be considered to be true guano-inhabitants because
they drop from their hosts (bats) when well fed, and
stay there for a while before starting to look for new
hosts. Springtails (Collembola) are, together with mites,
the most speciose invertebrates in guano piles, and may
form extremely large populations of guanophages. The
Brazilian species Acherontides eleonorae (Fig. 4) is a good
FIGURE 4 A large concentration of Acherontides eleonorae collem-
bolans, with some scattered mites. Close-up of the region pointed
out in Fig. 2 where the white spots indicate the presence of these
animals.
ENCYCLOPEDIA OF CAVES
362
G
example of a guanobite and is noteworthy in view of
the extremely large populations (hundreds of indivi¬
duals cm -2 ) found on old, dry piles of hematophagous
bat guano from Sao Paulo caves (therefore concentrating
on the borders of the guano piles through succession),
and because they survive and breed when kept in
closed vials with this substrate.
After Collembola and Acari, larvae of tineid
Lepidoptera are among the most ubiquitous guano-
phages in many caves where they make their cases.
Pyralid moths were also frequently recorded. Beetles
(Coleoptera) of the families Dermestidae, Histeridae,
Hydrophilidae, Scarabaeidae, Staphylinidae (includ¬
ing pselaphines), and Tenebrionidae are occasionally
found associated with guano piles, sometimes in large
populations. However, cholevine leiodids may be
considered to be the commonest guano-inhabiting
beetles (Fig. 3). Aderid beetles were frequently found
on swiftlet guano in Southeast Asia, where they form
guanobitic populations (Deharveng and Bedos, 2000).
The most important guanophage flies (Diptera) belong
to the families Drosophilidae, Fanniidae, Milichiidae,
Muscidae, and Phoridae, many of which are probably
guanobites. In Brazil, they show a preference for piles of
hematophagous guano; faniids and muscids are occa¬
sionally found associated with guano from carnivorous
bats in Southern Brazilian caves (Gnaspini and Trajano,
2000). A striking example of a troglobite-guanobite is the
African mormotomyiid Mormotomyia hirsuta, which is a
highly modified species whose larvae live on guano.
In tropical America, bugs (Heteroptera) of the fami¬
lies Cydnidae and Lygaeidae are associated exclu¬
sively with guano from frugivorous bats or from
oilbirds (due to the similar physical and nutritional
characteristics), where they may establish very large
breeding populations (see Gnaspini and Trajano,
2000). Lygaeids were also recorded in frugivorous bat
guano from Asia and New Guinea. In caves these
insects are restricted to guano and thus should be
considered guanobites. Representatives of these fami¬
lies are generally plant eaters (and some are detriti-
vores). This may explain why the cave species occur
only in guano from frugivorous animals.
Cockroaches (Blattaria) are typical tropical cave-
dwellers. These highly opportunistic insects are, as
expected, found in several substrates, including differ¬
ent types of guano, and may establish huge popula¬
tions. For instance, Braack (1989) recorded several
species in a single cave, with the population of one
species reaching 75,000 individuals and of two
other species reaching, each, 40,000—50,000 individuals.
Crickets (Ensifera) are omnivorous cavernicoles obser¬
ved throughout the world, and may concentrate near
and on guano piles. Isopods are also commonly
recorded on guano, sometimes forming large popula¬
tions. In Brazil, for instance, some groups show prefer¬
ence to particular kinds of guano: philosciids in guano
of frugivorous bats from Sao Paulo caves, scleropactids
in guano of insectivorous bats from one Para cave,
and armadillids in guano of carnivorous bats from
Mato Grosso do Sul (Gnaspini and Trajano, 2000).
Diplopods may also be found associated with guano, and
some species may be considered troglobite-guanobites,
because they are troglomorphic.
In addition, other higher taxa were occasionally
collected in or near guano piles. Earthworms (Oligo-
chaeta) may use guano as an extension of their
endogean habitat; they need a moist and nutrient-rich
substrate, such as that found in guano from hematopha¬
gous and frugivorous species (where large populations
of enchytraeid worms were found in Brazil; Gnaspini
and Trajano, 2000). Psocopterans feed on fungi and may
be found in guano. Subulinid gastropods were recorded
in some tropical American caves (see Gnaspini and
Trajano, 2000), and may form guanobitic populations in
Asia (Chapman, 1984). Harvestmen (Opiliones) may
also be found near the piles and are observed occasion¬
ally feeding on guano.
Among the predators, pseudoscorpions may be
found in any kind of guano, preying on the abundant
springtails and mites. A few cases of large populations
were recorded, for example, the Brazilian chernetid
Maxchernes iporangae (Fig. 1), which seems to be
restricted to a single cave where it exclusively inhabits
frugivorous bat guano piles (see Gnaspini and
Trajano, 2000). Spiders (Araneae) such as Pholcidae
and Loxosceles spp. (Sicariidae) sometimes make their
webs over or near guano piles, opportunistically prey¬
ing on guanophages. Barychelid trapdoor spiders
were observed associated to swiftlet guano in Asia
(Chapman, 1984). Whip spiders (Amblypygi) were fre¬
quently recorded in guano from swiftlets in Southeast
Asia (Deharveng and Bedos, 2000). Guanobitic geophi-
lomorph chilopods were recorded in Asia (Chapman,
1984). Reduviid bugs (Heteroptera) are not generally
directly associated to guano but occasionally feed on
guanophages.
It is important to stress that aquatic communities
may also use guano as food. For instance, veliid bugs
were recorded on floating guano particles from
Namibia; aquatic guanophages that feed on hematoph¬
agous bat guano seem to be the main food item of a
troglobitic catfish from Northeastern Brazil; guano
may be nutritionally important for troglobitic salaman¬
ders (Fenolio et ah, 2006); and even whole aquatic com¬
munities may depend on bat guano.
ENCYCLOPEDIA OF CAVES
GUANO COMMUNITIES
363
Acari
Predators outside guano
e.g., Bdellidae
Heteroptera
Macrochelidae
Reduviidae
Pseudoscorpiones
Araneae
e.g., Chernetidae
Ctenidae
Coleoptera
Sicariidae
e.g., Histeridae
Theridiosomatidae
Pselaphidae
Acari
e.g., Uropodidae
Collembola
e.g., Hypogastruridae
Paronellidae
Chilopoda
Lithobiomorpha
Collembola
e.g., Entomobryidae
Isopoda
BAT GUANO
Heteroptera Collembola
Lygaeidae
Cydnidae
e.g., Arrhopalitidae
Isotomidae
Onychiuridae
Psocoptera
Isopoda
Diptera
Drosophilidae
Muscidae
Phoridae
Coleoptera
Cholevidae
Lepidoptera
Tineidae
A A~
Opiliones
Gonyleptidae
Diplopoda
Ensifera
Phalangopsidae
Insectivorous
Frugivorous
Hematophagous
Carnivorous
Other resources
FIGURE 5 Hypothetical food web for bat-guano communities from Sao Paulo caves. Southeastern Brazil. Arrows indicate direction of
energy flow.
THE FOOD WEB—EXAMPLES
As mentioned before, animals inhabiting guano
piles may be guanophages and their predators and
may also inhabit other substrates. Figure 5 shows an
example of a Brazilian food web, where a large variety
of types of guano may be found. Note that some taxa
may be found in any kind of guano (e.g., tineid larvae),
whereas some are restricted to one kind (e.g., flies in
guano from hematophagous bats, and heteropterans and
lithobiomorphans in guano from frugivorous species).
CONSERVATION
Guano mining causes an impact on the bat popula¬
tion, which can move away from the mining site,
resulting in the interruption of guano deposition. This,
in turn, would lead to the abandonment and/or death
of guano inhabitants, especially when guanobites are
considered. This also includes aquatic environments. It
is also clear that any important change in the epigean
environment may lead to the death and/or abandon¬
ment of bats (and/or birds and/or crickets), and, con¬
sequently, the guano inhabitants. Again, guanobites
are important targets, and guanobites-troglobites may
even be extinguished. For instance, logging or forest
cutting may locally extinguish frugivorous bats and
birds, interrupting the deposition of guano inside the
local caves. One important example is the popular use
of poisonous pastes against hematophagous bats in
Neotropical countries because of their menace to cattle.
The paste is applied on the fur of the bats, which, due
to their social behavior, lick each other and die in
large groups. In the following days and months
there is a large increase in energy input due to the
large amount of cadavers deposited in the caves.
However, afterward, energy input drastically decrea¬
ses. Pollution from metals may also represent a threat
to bats.
CONCLUSIONS
In summary, guano is an important source of
food in caves all around the world. There are some
basic taxa that are almost always present in guano
communities in caves, and accessory taxa, which are
somewhat less common but still frequent groups. In
addition, the composition of guano communities
shows some differences related to the kind of guano
(e.g., bird vs. bat guano, guano of frugivorous vs.
ENCYCLOPEDIA OF CAVES
364
G
insectivorous vs. hematophagous bats, etc.). Although
there are large number of papers about guano commu¬
nities in caves, there is still much research to be done
and many important findings have been occurring in
the past decades.
Acknowledgments
Sonia Hoenen and David Culver provided helpful criticisms to the
manuscript. The author has a research fellowship from CNPq
(Conselho Nacional de Desenvolvimento Cientifico e Tecnologico).
Figures 1 and 3 were reprinted/modified from Gnaspini and Trajano
( 2000 ).
See Also the Following Articles
Food Sources
Bats
Diversity Patterns in the Tropics
Bibliography
Braack, L. E. O. (1989). Arthropod inhabitants of a tropical cave
"island" environment provisioned by bats. Biological Conservation,
48, 77-84.
Calder, D. R., & Bleakney, J. S. (1965). Microarthropod ecology of a
porcupine-inhabited cave in Nova Scotia. Ecology, 46, 895—899.
Chapman, P. (1984). The invertebrate fauna of the caves of Gunung
Mulu National Park. Sarawak Museum Journal, 51, 1—22.
Decu, V. (1986). Some considerations on the bat guano synusia.
Travaux de VInstitut de Spe'ologie "Emile RacovUd," 25, 41—51.
Decu, V., & Tufescu, M. V. (1976). Sur l'organisation d'une biocenose
extreme: la biocenose du guano de la grotte "Pestera lui Adam"
de Baile Herculane (Carpates Meridionales, Roumanie). Travaux
de VInstitut de Spe'ologie "Emile Racoviviid," 15, 113—132 (in French).
Deharveng, L., & Bedos., A. (2000). The cave fauna of Southeast Asia.
Origin, evolution and ecology. In H. Wilkins, D. C. Culver &
W. F. Humphreys (Eds.), Subterranean ecosystems (pp. 603—632).
Amsterdam: Elsevier Press.
Fenolio, D. B., Graening, G. O., Collier, B. A., & Stout, J. F. (2006).
Coprophagy in a cave-adapted salamander; the importance of bat
guano examined through nutritional and stable isotope analysis.
Proceedings of the Royal Society B, 273(1585), 439—443.
Gnaspini, P. (1992). Bat guano ecosystems. A new classification and
some considerations, with special references to Neotropical data.
Me'moires de Biospeologie, 19, 135—138.
Gnaspini, P., & Trajano, E. (2000). Guano communities in tropical
caves. In H. Wilkins, D. C. Culver & W. F. Humphreys (Eds.),
Subterranean ecosystems (pp. 251—268). Amsterdam: Elsevier Press.
Negrea, A., & Negrea, S. (1971). Sur la synusie du guano des grottes
du Banat (Roumanie). Travaux de VInstitut de Spe'ologie "Emile
RacovUd," 10, 81 — 122 (in French).
Poulson, T. L. (1972). Bat guano ecosystems. Bulletin of the National
Speleological Society, 34(2), 55—59.
Poulson, T. L., & Lavoie, K. H. (2000). The trophic basis of subsurface
ecosystems. In H. Wilkins, D. C. Culver & W. F. Humphreys
(Eds.), Subterranean ecosystems (pp. 231—249). Amsterdam:
Elsevier Press.
Trajano, E. (1985). Ecologia de populafoes de morcegos cavernicolas
em uma regiao carstica do sudeste do Brasil. Revista Brasileira de
Zoologia, 2(5), 255—320 (in Portuguese).
GYPSUM CAVES
Alexander Klimchouk
Ukrainian Institute of Speleology and Karstology,
Simferopol, Ukraine
OCCURRENCE OF GYPSUM AND TYPES
OF GYPSUM KARSTS
Gypsum (CaS 04 ■ 2H 2 0) is one of the most common
evaporite rocks, that is, inorganic rocks formed by
chemical precipitation in a concentrated solution. As
gypsum is much more soluble than limestone, it is
much less common in outcrops. This is one of the rea¬
sons why gypsum karst has received relatively little
appreciation in the past in the mainstream of karstol¬
ogy, and has commonly been considered of limited
significance. However, since the 1970s it has been
increasingly realized that karst processes in gypsum
operate extensively when it occurs below various types
of cover beds within the upper few hundred meters of
sedimentary sequences, that is, in intrastratal karst set¬
tings that are very common over cratonic regions
(Klimchouk et ah, 1996). Various engineering problems,
such as ground instability and construction damage
due to collapse and subsidence, leakages from canals
and reservoirs, and complications during mining
operations readily arise in gypsum karst areas due to
both the presence of preformed caves and fast devel¬
opment of new conduits.
The variety of depositional settings in which gyp¬
sum can form ranges from deep-water, to subaerial,
to lacustrine, so the mode of occurrence of gypsum
and the associations of lithofacies is quite diverse.
Evaporites are most susceptible to recrystallization,
dissolution, and replacement among common rocks.
Gypsum deposited in the shallow subsurface is con¬
verted to anhydrite with burial, and back to gypsum
as the sulfate reenters the shallow subsurface realm
and is washed in cooler and fresher fluids. Gypsum
and anhydrite occur widely as rare to frequent inter¬
beds in sequences containing halite, carbonates, and
siliciclastic sediments, but also as massive beds. Indi¬
vidual beds of gypsum with thicknesses of 10—40 m
are fairly common, although sequences of evaporite/
carbonate/siliciclastic rocks may exceed 1000 m and
more in thickness. Further details of gypsum geology
can be found in published comprehensive accounts
(i.e. Warren, 2006).
The karst that develops in gypsum has many
features in common with carbonate karst, but also has
significant peculiarities imposed by the specific fea¬
tures of evaporite geology and dissolution processes
Encyclopedia of Caves.
© 2012 Elsevier Inc. All rights reserved.
GYPSUM CAVES
365
(Klimchouk, 2002; Klimchouk et al., 1996). In character¬
izing gypsum karst and caves, it is particularly impor¬
tant to refer to the overall settings of karst
development, which are conveniently generalized by
the evolutionary classification of karst types. Types of
karsts are viewed as successive stages in geological/
hydrogeological evolution, from deposition through
burial to reemergence, between which the major
boundary conditions, the overall circulation pattern,
and extrinsic factors and intrinsic mechanisms of karst
development appear to change considerably (see the
Figure 3 in this book's article, "Speleogenesis,
Hypogenic"). The different types of karsts are marked
by characteristic styles of cave development, which are
particularly distinct in gypsum.
Syngenetic karst, which commences in freshly depos¬
ited gypsum, is limited in extent and does not present
any significant caves.
Intrastratal karst is considered to develop within
rocks already buried by younger strata, where karstifi-
cation is later than deposition of the cover rocks. It is
by far the predominant karst type in gypsum. As a
consequence of standard denudation and uplift on the
continents, the buried rocks are shifted with time into
progressively shallower positions. Karstification may
be initiated at any of the stages of intrastratal develop¬
ment en route back to the surface. The entire sequence
of intrastratal karst types (stages) includes (in the
order as they potentially evolve) deep-seated karst, subja¬
cent karst, entrenched karst, and denuded karst. The
respective settings differ by the degree of separation of
soluble units from the surface and by structural and
hydrogeological conditions for karstification. In terms
of groundwater dynamics, deep-seated karst is always
confined (or semiconfined). In subjacent karst the
major hydrogeological confinement is locally breached
by erosion of the cover so that hydrodynamics changes
through to vadose and water table. Further erosional
incision leads to the establishment of entrenched karst,
where major valleys incise below the bottom of the
karst aquifer and largely or entirely drain the gypsum
beds that remain capped with protective insoluble
cover for the most part of an area. Denuded karst is
former intrastratal karst that develops where the solu¬
ble rock has been almost completely exposed from
beneath the cover, but with substantial inheritance of
solution porosity from the preceding stages of intras¬
tratal development.
Where poorly permeable clays or shales surround
gypsum in a stratified sequence, intrastratal karst may
not develop at all until the gypsum bed is exposed
to the surface. Open karst is karst that commences
only after exposure of the soluble rock, without any
significant intrastratal karstification. While both open
and denuded karst represents exposed settings, styles
of cave development differ dramatically in these types
of karsts. Paradoxically, conduit development in
gypsum in open karst is somewhat hindered by rapid
dissolution of gypsum. Overall, the true open karst
type in gypsum is commonly characterized by limited
karstification, while karst types of the intrastratal
group frequently demonstrate a very high degree of
karstification.
IMPLICATION OF EQUILIBRIUM
CHEMISTRY AND DISSOLUTION
KINETICS TO SPELEOGENESIS
IN GYPSUM
Chemical equilibrium and reaction kinetics for
gypsum dissolution differ from those of calcite,
imposing some important consequences for speleo¬
genesis. The solubility of gypsum and calcite depends
on conditions different for each mineral. The solubil¬
ity of gypsum is roughly one order of magnitude
greater than the solubility of calcite under typical
CO 2 pressures found in karst. However, this differ¬
ence can be much greater under certain conditions
(see Fig. 6B in this book's article, "Speleogenesis,
Hypogenic"). The gypsum solubility can be consider¬
ably enhanced or renewed at the presence of other
salts in groundwaters, by anaerobic reduction of sul¬
fates in the presence of organic matter, by dedolomiti-
zation of intercalated dolomite layers, and by some
other factors. This is particularly important for hypo-
gene speleogenesis in intrastratal settings.
From the perspective of speleogenesis, the most
important difference of dissolution processes between
limestone and gypsum lies in kinetics. The dissolution
of calcite, the principal constituent of carbonate rocks,
is controlled mainly by the reaction at the rock surface.
Experimental studies of dissolution rates for lime¬
stones suggest that relatively high rates drop rapidly
when a solution reaches about 60—90% saturation
because kinetic reaction changes its order in this region
(White, 1977; Palmer, 1991). The nonlinear dissolution
rate is low and this kinetic switch is of the ultimate
importance for speleogenesis in carbonates as consid¬
erable dissolution can continue, although at slow rates,
along entire flow paths. A positive feedback loop
between discharge and the rate of fissure growth
makes the mechanism self-accelerating. The rapid
enlargement stage begins when the water is able to
pass through the entire conduit while preserving con¬
siderable undersaturation. This represents the break¬
through event resulting in a boost of the growth rate,
provided that hydrogeological settings permit increas¬
ing discharge.
ENCYCLOPEDIA OF CAVES
366
G
In contrast to limestone, dissolution of gypsum is
controlled mainly by diffusion across a boundary
layer. It was long assumed that the dissolution reac¬
tions of gypsum follow a linear rate law and no switch
to slow kinetics occurs with an increase in concen¬
tration. Hence slow but uniform enlargement through¬
out long initial flow paths cannot occur to open them
up. However, recent experiments have shown that
gypsum dissolution does slow at roughly 95% satura¬
tion (Jeschke et al., 2001), which allows some slow
dissolution at very close to saturation. In any case,
enlargement of initial tight flow paths in gypsum to
permit a fast dissolution regime (the breakthrough con¬
ditions) seems to be much more difficult than in lime¬
stones. Caves may form only where gradients are high
enough, or initial widths of openings are large enough,
and/or where flow distances are short enough to enable
breakthrough conditions within a feasible time. These
conditions do not occur frequently in exposed settings.
The result is that speleogenesis in open gypsum karst
settings is limited, localized, and does not extend to
considerable depth (Klimchouk, 2002).
Because of the high solubility and fast dissolution of
gypsum (which also depends on flow velocity), early
conduit development in this rock is extremely sensitive
to variations of boundary conditions such as fissure
aperture widths and hydraulic gradient (Birk, 2002).
The range of variations, between which the speleoge-
netic initiation of conduits is attained either very fast
or virtually never, is very narrow. This explains why,
on one extreme, gypsum formations may remain virtu¬
ally untouched by cave development in many cases,
and on the other extreme conduits may develop fast
along favorable flow routes.
High dissolution rates in gypsum and their depen¬
dence on flow velocity cause the flow/dissolution
feedback in gypsum speleogenesis to be stronger than
in limestones. Before the breakthrough event, disparity
of dissolution rates between alternative flow paths in
gypsum grows faster than in limestone. After break¬
through, enlargement rates in the successful conduits
increase more dramatically in gypsum than in lime¬
stone. These differences determine that speleogenetic
competition between alternative flow paths is much
stronger in gypsum than in limestone and that a "run¬
away" development of the favorable paths is more
pronounced (Klimchouk, 2002).
SPELEOGENESIS IN DIFFERENT TYPES
OF KARSTS
The above-mentioned peculiarities of gypsum disso¬
lution manifest themselves best in open karst settings.
Precipitation that falls onto the exposed gypsum rocks
contributes greatly to surface denudation but it has a
limited potential to create caves because the surface
runoff and dispersed infiltration get almost saturated
very fast. No epikarst is known to form in open gyp¬
sum karst. Instead, a kind of sealing crust develops on
exposed gypsum in arid and semi-arid climates due to
recrystallization in the uppermost layers caused by
repeated wetting and drying. In the saturated (phre¬
atic) zone of unconfined aquifers, speleogenetic initia¬
tion of initial tight flow paths in gypsum is very slow
and it may not occur within reasonable geological
time. Speleogenesis in open gypsum karst settings
depends on the availability of initial pathways wide
enough for breakthrough (undersaturated though-
flow) to occur quickly. It finds and exploits a few
favorable paths (if there are any), while other potential
paths languish their growth. After breakthrough, pas¬
sages quickly develop to accommodate the highest
possible discharge, and water table or vadose condi¬
tions are quickly established. Linear or crudely branch¬
ing caves form (Fig. 1), particularly where flashy
concentrated recharge occurs, rapidly adjusting to the
FIGURE 1 Typical sketches, plans, and profiles of epigene gyp¬
sum caves formed in unconfined settings: (A) in the open karst of
Sorbas, Spain; feeding dolines are indicated by dots ( adapted from
Calaforra, 1998); (B) and (C) in the open karst of the Ekeptze-Gadyk
massif. North Caucasus, Russia ( adapted from Ostapenko, 2001); (D)
the Spipola-Acquafredda system in the open karst of the Emilia
Romagna, Italy ( adapted from Grimandi, 1987).
ENCYCLOPEDIA OF CAVES
GYPSUM CAVES
367
contemporary geomorphic setting and available
recharge. Most of the dissolution occurs where water
first enters the gypsum, or is focused along streambeds
during high-flow events. No deep phreatic conduits
develop in unconfined gypsum aquifers. In the vadose
zone, cave development concentrates along wide verti¬
cal percolation paths, forming vertical pitches, and
along free stream courses often following minor
insoluble interbeds. Caves of this type may have sev¬
eral levels of subhorizontal passages connected by ver¬
tical pits.
Open gypsum karsts are found mainly in arid and
semi-arid environments, and where fairly thick gyp¬
sum units were surrounded by impervious sequences
prior to exposure. The best examples of caves formed
in open karst type are documented in the Central
Apennines and Sicily in Italy (Neogene gypsum), in
Sorbas (Neogene gypsum) and Vallada (Triassic gyp¬
sum) regions in Spain, in New Mexico and in
Oklahoma, in the United States (Permian gypsum),
and in the Pinega region of Russia (Permian gypsum).
They include the longest gypsum caves of this type:
Kulogorskaja-Troja (Pinega; 14.3 km long). Jester Cave
(Oklahoma; 11.8 km long), Spipola-Aquafredda (Central
Apennines; 11km long and 118 m deep; Fig. ID), the
Gueva de Aqua (Sorbas; 8.35 km long), and the deepest
gypsum cave in the world, Tunel dels Sumidors
(Vallada; 210 m deep). Other significant gypsum caves
of this type are explored in the Diebel Nador and
Oranais areas in eastern and western Algeria; in the Ar
Rabitat/Bir area in northwest Libya; in central and
northern Somalia; in some mountain areas in the North
Caucasus in Russia; and in Central Asia in Tajikistan.
In intrastmtal settings, conditions and styles of spe-
leogenesis change in the course of reburial, geomor¬
phic development, and hydrogeological evolution.
Caves can initiate and develop in confined conditions
(deep-seated karst; hypogene speleogenesis). Subjacent
karst represents transitional conditions from confined
to unconfined ones, where hypogene and epigene spe¬
leogenesis may operate depending on the locally pre¬
vailing flow regime, but hypogene speleogenesis still
dominates. The further development of preformed
hypogene cavities and the contemporaneous epigene
speleogenesis can continue through unconfined condi¬
tions (entrenched and denuded karst).
In deep-seated gypsum karst, there are several common
favorable situations and respective mechanisms for cave
initiation and development. The most common is trans¬
verse speleogenesis in a stratified sequence, where the
gypsum bed is sandwiched between good aquifers such
as clastic or dolomite beds in a multistory leaky con¬
fined system, with diffuse input from below and
upwelling flow driven by head gradient across the gyp¬
sum bed (Fig. 2A). Where laterally connected fracture
FIGURE 2 Initiation and development of hypogene conduits and
cavities in gypsum (adapted from Klimchouk, 2000): (A) transverse spe¬
leogenesis in a fractured gypsum bed sandwiched between aquifers;
(B) transverse speleogenesis along a single fracture zone cross-cut¬
ting an otherwise low-fissured bed; (C) lateral contact speleogenesis
driven by natural convection in the case of a massive nonfractured
bed underlain by a prolific aquifer.
systems exist in the gypsum, reticular maze caves can
form by aggressive waters rising from the underlying
aquifer through many discrete points (feeders; Fig. 3).
Where fractures are arranged in several horizontal
levels due to textural/structural inhomogeneities or the
presence of minor interbeds, multistory networks
develop (see Fig. 7 A in the article "Speleogenesis,
Hypogenic"). Reticular, rectilinear, or polygonal maze
patterns guided by fractures are common in gypsum
(Fig. 3), while anastomotic and spongework maze pat¬
terns are not.
The transverse hypogene speleogenesis across a sol¬
uble bed in a multiple aquifer system has an important
mechanism that suppresses strong flow/dissolution
feedback and allows many conduits to develop. When
some favorable flow paths across the gypsum achieve
the breakthrough, their growth accelerates, but this
quickly diminishes the head gradient across the gyp¬
sum in the vicinity of the successful conduits. As the
hydraulic resistance of the conduit becomes smaller
than that of the adjacent aquifers, discharge through
the conduit is now controlled by the hydraulic
ENCYCLOPEDIA OF CAVES
368
FIGURE 3 Distribution of point feeders through networks of master passages in transverse hypogene maze caves. Red dots indicate
feeders; subvertical conduits at the lower story are shown in blue; connecting passages at the upper level are shown in yellow. Left: Ozerna
Cave in the Neogene gypsum. Western Ukraine (adapted from Klimchouk, 1990). Right: Coffee Cave in the Permian gypsum. New Mexico, U.S.A.
(adapted from Stafford et al., 2008). Lower level passages locally form feeding subsystems that connect to the master level through point feeders.
conductivity of the least permeable lithological unit in
the system and by the boundary conditions (external
hydraulic control), but no longer by the diameter of
the conduit. Unless and until the boundary conditions
change, the flow and enlargement rates in the conduit
remain roughly constant at some level. The positive
feedback loop is no longer the determinant of conduit
development. Because the vertical head gradient
between the aquifers is still maintained, although
diminished, at some lateral distance apart from the
successful transverse conduit, alternative conduits do
not languish but continue to grow and eventually
reach the breakthrough, either to the upper aquifer or
laterally to the conduit that had been "broken
through" earlier.
On the general background of upwelling hydraulic
communication, flow pattern commonly acquires local
lateral components at various levels because of vertical
heterogeneities in hydraulic conductivity. These are
caused by an imperfect connection of vertically super¬
imposed fracture systems (see Fig. 7 in the article,
"Speleogenesis, Hypogenic"), the presence of minor
insoluble intercalations, and overall discordance in
plan positions of ultimate recharge and discharge
points. The lateral flow components cause merging
individual transverse passage clusters into laterally
extensive cave systems, giving a false impression of
generally lateral cave-forming flow.
Outstanding examples of transverse hypogenic
maze caves are the giant gypsum caves in the Neogene
gypsum of Western Ukraine. Five of them occupy the
top ranks in the list of the longest gypsum caves in
the world (Table 1), totaling over 514 km of passages.
The longest gypsum cave Optymistychna is currently
mapped for 232 km and is the second longest cave in
the world after the Mammoth—Flint-Ridge system in
Kentucky, U.S.A. Other prominent examples include
network caves encountered by mines in the Paleogene
gypsum of the Paris basin, France; the maze caves in
the Neogene gypsum of Estremera, Madrid basin,
Spain; in the Permian gypsum of the Pecos Valley
Region in New Mexico, U.S.A.; and in the Pinega and
Fore-Ural regions of Russia.
Where cross-bed fractures in gypsum are scarce or
where water rises into the gypsum not diffusely but
through localized fracture zones in the underlying for¬
mation, isolated passages or clusters of a few passages
form (Fig. 2B). Good examples are the first hypogenic
caves in the Messinian (Neogene) gypsum in Italy,
recently encountered by the Moncalvo underground
quarry (Vigna et al., 2010). Interception of these caves
caused a massive inrush of the pressurized waters into
ENCYCLOPEDIA OF CAVES
GYPSUM CAVES
369
TABLE 1 The World’s Longest (>2000 m) and Deepest (>100 m) Gypsum Caves
Name
Longest Gypsum Caves of the World
Length, m Country, Region
Rock Age
Opty misty chna
232,000
Ukraine, Western
Neogene
Ozerna
131,400
Ukraine, Western
Neogene
Zoloushka
92,000
Ukraine, Western
Neogene
Mlynki
36,000
Ukraine, Western
Neogene
Kristal'na
22,610
Ukraine, Western
Neogene
Kulogorskaja-Troja
16,250
Russia, Pinega
Permian
Jester
11,800
U.S.A., OK
Permian
Spipola-Aquafredda
10,500
Italy, Emilia-Romagna
Neogene
Olimpijskaja-Lomonosovskaya
9110
Russia, Pinega
Permian
Slavka
9100
Ukraine, Western
Neogene
Verteba
8550
Ukraine, Western
Neogene
Agua, cueva de
8350
Spain, Sorbas
Neogene
Cater Magara
7300
Syria
Neogene
Park's Ranch
6595
U.S.A., NM
Permian
Konstitutzionnaja
6130
Russia, Pinega
Permian
Kungurskaya Ledjanaja
5700
Russia, Fore-Ural
Permian
Mushkarova Jama
5050
Ukraine, Western
Neogene
Selman
4794
U.S.A., OK
Permian
Severny Sifon
4617
Russia, Pinega
Permian
Ordynskaya
4600
Russia, Fore-Ural
Permian
Kumichevskaja
4520
Russia, Pinega
Permian
Zolotoj Kljuchik
4380
Russia, Pinega
Permian
Covadura
4245
Spain, Sorbas
Neogene
Estremera, cueva de
4000
Spain, Madrid
Neogene
Crystal Caverns
3807
U.S.A., NM
Permian
Double Barrel
3724
U.S.A., NM
Permian
Scrooge
3700
U.S.A., NM
Permian
Umm al Masabih
3593
Libya
Jurassic
Gostry Govdy
3570
Ukraine, Western
Neogene
Simfonia
3240
Russia, Pinega
Permian
Pekhorovskaja
3200
Russia, Pinega
Permian
Carcass
3164
U.S.A., NM
Permian
Martin
3150
U.S.A., NM
Permian
Abisso LU.S.A.
3000
Italy, Emilia-Romagna
Neogene
Leningradskaja
2950
Russia, Pinega
Permian
Wimmelburger Schlotte
2840
Germany
Permian
Horseshoe Valley
2760
U.S.A., OK
Permian
( Continued )
ENCYCLOPEDIA OF CAVES
370
G
TABLE 1 (Continued)
Name
Longest Gypsum Caves of the World
Length, m Country, Region
Rock Age
Pshashe-Setenay
2690
Russia, N. Caucasus
Jurassic
Pinezhskaja Terehchenko
2600
Russia, Pinega
Permian
Vodnaja
2600
Russia, Pinega
Permian
Jansill / Driftwood
2562
U.S.A., NM
Permian
River Styx
2562
U.S.A., TX
Permian
Fanning Ranch
2557
U.S.A., NM
Permian
Yubuleynaya
2555
Russia, Pinega
Permian
Atlantida
2525
Ukraine, Western
Neogene
Eras'kina 1-2
2500
Russia, Pinega
Permian
Ingh. Ca' Siere
2500
Italy, Emilia-Romagna
Neogene
Triple Engle Pit
2485
U.S.A., NM
Permian
10-years LSS
2450
Russia, Pinega
Permian
Bukovinka
2408
Ukraine, Western
Neogene
Nescatunga
2398
U.S.A., OK
Permian
Segeberger-Kalkhoehle
2360
Germany, S. Harz
Permian
Coffee Cave
2321
U.S.A., NM
Permian
Hyaenlabyrint
2310
Somalia
Paleogene
Pekhorovskij Proval
2262
Russia, Pinega
Permian
Geograficheskogo obshchestva
2150
Russia, Pinega
Permian
Ugryn'
2120
Ukraine, Western
Neogene
Re Tiberio
2110
Italy, Sicily
Neogene
Hay's
2037
U.S.A., NM
Permian
Kulogorskaja-5
2035
Russia, Pinega
Permian
Popova
2032
Russia, N. Caucasus
Jurassic
Michele Gortani
2015
Italy, Emilia-Romagna
Neogene
Name
Deepest Gypsum Caves of the World
Depth, m Country, Region
Rock Age
Tunel dels Sumidors
210
Spain, Vallada
Triassic
Abisso LU.S.A.
204
Italy, Emilia-Romagna
Neogene
Pozzo A
>200
Italy, Emilia-Romagna
Triassic?
Triple Engle Pit
134
U.S.A., NM
Permian
Corail, sima del
130
Spain, Sorbas
Neogene
Covadura
126
Spain, Sorbas
Neogene
Campamento, sima del
122
Spain, Sorbas
Neogene
Aguila, sima del
112
Spain, Vallada
Triassic
Rio Stella-Rio Basino
100
Italy, Sicily
Neogene
AB 6
100
Russia, N. Caucasus
Jurassic
ENCYCLOPEDIA OF CAVES
GYPSUM CAVES
371
the mine. After several months of pumping, the water
level was lowered, giving access to 5—7-m- wide pas¬
sages with distinct hypogene morphology.
In highly soluble aquifers such as gypsum, natural
convection circulation and gravitational separation of
water due to density differences play important roles
in speleogenesis, especially in hypogenic speleogen-
esis. As forced-flow regimes in confined settings are
commonly sluggish and water with lesser density
enters the gypsum from below, free convection pat¬
terns powered by density gradients are widely opera¬
tive in hypogenic speleogenesis in gypsum. Various
morphological effects of buoyancy dissolution are well
represented in hypogenic gypsum caves, including
keyhole sections, rising-wall channels, ceiling half¬
tubes, cupolas, and domepits. In mixed convection
systems, buoyancy dissolution effects are particularly
pronounced during the mature stage of speleogenesis,
where considerable conduit space has been created
and vertical hydraulic gradients across the gypsum
bed are diminished.
Speleogenesis at the base of the gypsum due to
buoyancy dissolution may operate even without guid¬
ing cross-bed and forced transverse flow across the
unit. Where the gypsum is underlain by a prominent
aquifer, the aggressive water attacks the gypsum from
the bottom, and when it gets loaded with dissolved
sulfates, it is replaced by less dense waters, returns to
the basal aquifer, and outflows with the regional flow
(Fig. 2C). In this way vast isolated irregular chambers
and passages, or ramiform cavities, can form at the
bottom of thick gypsum sequences. Outstanding exam¬
ples of this kind of speleogenesis include numerous
giant Schlotten-type caves in the Zechstein (the Upper
Permian) gypsum encountered by mines in the South
Harz region of Germany at depths up to 400 m
(Kempe, 1996).
Reemergence of soluble units to the shallow subsur¬
face and local breaching of the hydrogeological con¬
finement in the course of uplift and erosion signifies
the stage of subjacent karst, marked by drastic changes
of boundary conditions for speleogenesis. At the
beginning of this stage discharge through preformed
hypogenic caves increases in areas of the breaching,
with respective increase of dissolution rates and vol¬
ume of cavities. With the subsequent establishment of
vadose zone and water-table conditions dissolution is
localized along the water table and, where surface
streams are swallowed, along underground stream-
beds. Dissolution along the water table is favored by
distinct differences in density and aggressiveness
between the top 5—10-cm layer and the bulk of water
in partially submerged cave passages. This leads to
expressed horizontal notching and widening of pas¬
sages (Fig. 4A,B,C). In passages located below the
FIGURE 4 Modification of hypogene gypsum caves in uncon¬
fined (water table) conditions. (A) Basic unmodified fissure-like pas¬
sage. (B) Passage with horizontal wall notches and inclined facets
developed by dissolution along the water table. (C) Diagram illus¬
trating natural convection flow pattern in a passage under water-
table conditions (exclamation mark and red arrows indicate the most
aggressive water at the top layer of water) and evolution of the cross
section. (D) Map of the 92-km-long Zoloushka Cave in Western
Ukraine; the dashed lines in the profile inset indicate the water-
table positions before and after the quarrying ( map courtesy of the
Chernivtsy Speleological Club; from Andrejchouk, 2008). (E) A typical
passage at the upper level in Zoloushka Cave showing horizontal
notching due to dissolution at the water table. Photos A and B courtesy
of A. Kananovich; photo E h/ the author.
water table, natural convection effects lead to the
development of characteristic tip-down triangular
cross-sections, with flat ceilings (laugdecke in German),
combined with inclined wall facets.
ENCYCLOPEDIA OF CAVES
372
G
A good example of the hypogene maze cave in the
subjacent karst stage is the longest explored underwater
gypsum cave Ordinskaya in the Fore-Ural, with over
4.5 km of passages mapped by cave divers (Fig. 5).
Divers also mapped more than 100 active feeders at the
bottom of passages through which upward recharge
still occurs to the cave from the underlying limestone
aquifer. Another outstanding example is the maze of
Zoloushka Cave in Western Ukraine, in which 92 km of
passages have been mapped since 1976 (Fig. 4D). The
position of the water table in the area during the
Holocene is marked by extensive horizontal notching
and widening in passages at the upper story (Fig. 4E).
With the open-pit quarry operation and groundwater
withdrawal since the late 1950s the water table has
dropped more than 20 m (Fig. 4D, inset profile), which
made the huge maze cave accessible for direct
exploration.
FIGURE 5 The map of Ordinskaya Cave in the Fore-Ural, the lon¬
gest underwater gypsum cave. The hypogenic cave is presently
under conditions of subjacent karst. Red circles indicate active
feeders at the bottom of passages mapped in the central part of the
cave (one of which is shown on the inset photo (A)). Inset photo (B)
shows the entrance of the cave and the entrenching Kungur River.
Modified and compiled from the map and the photos by P. Sivinskykh
(2009).
Deepening of erosional valleys below the gypsum
bottom signifies the stage of entrenched karst. Most pre¬
formed caves become relict, although those within the
reach of high levels of nearby rivers can be subject to
backflooding, which greatly contributes to further
solution widening of passages. A characteristic feature
of many entrenched gypsum karsts are vertical solu¬
tion pipes (also known as chimneys or comins) that per¬
forate gypsum sequences and often superimpose on
the relict hypogene caves, being formed by aggressive
percolation waters leaking through the vadose zone
from perched aquifers above.
Surface denudation may lead to almost complete
removal of the protective cover in some areas of
instrastratal karst regions, which establishes denuded
karst type. Caves inherited from the preceding stages
progressively disintegrate through collapsing. This is
why areas of denuded gypsum karst demonstrate
much denser doline distribution than open karsts.
Contemporary cave development under unconfined
conditions in subjacent, entrenched, and denuded
karst types generate single-conduit or crudely branch¬
ing caves, very much in the same way as in open karst.
However, the presence of the insoluble cover provides
more aggressive recharge from the surface via dolines
and ponors, and the presence of inherited cave poros¬
ity facilitates more rapid epigene cave development. In
the regions of intrastratal gypsum karst, contemporary
epigene caves coexist with (and often superimpose
onto) relict hypogene caves, demonstrating strikingly
different patterns and mesomorphology (Fig. 6). Being
evolutionary stages of intrastratal karst development,
subjacent, entrenched, and denuded karst types often
intersperse in an area or lie adjacent to each other
and to deep-seated karst within the same aquifer in
block-fault and monoclinal tectonic settings. Such var¬
iations are due to differentiated rates of neotectonic
uplift. Remarkable examples include the Pinega and
Pre-Ural regions in Russia, Western Ukraine, the
South Harz in Germany, and the Gypsum Plain in
New Mexico, U.S.A.
SUMMARY
Gypsum caves include some of the world's longest
caves, although the deepest caves in this rock barely
reach 200 m in depth. There are several characteristic
patterns and styles of caves that form in gypsum,
distinctly related to particular settings of karst devel¬
opment. These settings change regularly with the
hydrogeological and geomorphological evolution.
In confined intrastratal settings (deep-seated karst)
the following types of caves form in gypsum by
ENCYCLOPEDIA OF CAVES
GYPSUM CAVES
373
FIGURE 6 Patterns and morphology of hypogene (left column,
(A), (C), and (E)) and epigene (right column, (B), (D), and (F)) gyp¬
sum caves in the Pinega region in north Russia. (A) Simfonia Cave
(hypogene transverse maze cave); (B) Kumichevka-Vizborovskaya
Cave (epigene stream cave); (C), (E) Simfonia Cave;
(D) Lomonosovskaya Cave; (F) Golubinsky Proval Cave. Cave maps
from Maikov et al., 2001; photos by Yu. Nikolayev.
different mechanisms of hypogene speleogenesis,
depending on lithostratigraphic, structural, and hydro-
geological conditions:
1. Reticular 2D and 3D (multistory) maze caves
2. Isolated passages along prominent disruptions
3. Isolated irregular chambers and passages at the
base of gypsum
The morphology of these caves can be considerably
modified on the subsequent stages of unconfined intra-
stratal karst development, particularly in the water-
table conditions in subjacent and entrenched karsts.
The common patterns of gypsum caves that form
contemporaneously in unconfined intrastratal settings
(subjacent, entrenched, and denuded karst types)
include:
4. Linear and crudely branching through-flow caves
5. Vertical solution pipes fed by perched aquifers in
cover beds
In open karst, linear and crudely branching
through-flow caves are the predominant type. They
form close to the water table and in the vadose zone.
No deep phreatic conduit development occurs in open
gypsum karst.
See Also the Following Article
Ukraine Giant Gypsum Caves, Speleogenesis, Hypogenic
Bibliography
Andrejchouk, V. N. (2007). Zoloushka cave. Simferopol-Sosnowiec:
Silezian University — Ukrainian Institute of Speleology and
Karstology (in Russian).
Birk, S. (2002). Characterisation of karst systems by simulating aqui¬
fer genesis and spring responses: Model development and appli¬
cation to gypsum karst. Tubingen Geowissenschaftliche Arbeiten,
Reine C, 60.
Calaforra, J. M. (1998). Karstologia de yesos. Almeria, Spain: Universidad
de Almeria.
Grimandi, P. (1987). Grotta della Spipola. Ipoantropo (Italy), 5, 51—64.
Jeschke, A., Vosbeck, K., & Dreybrodt, W. (2001). Surface controlled
dissolution rates of gypsum in aqueous solutions exhibit nonlin¬
ear dissolution kimetics. Geochimica et Cosmochimica Acta, 65(1),
27-34.
Kempe, S. (1996). Gypsum karst of Germany. In A. Klimchouk, D.
Lowe, A. Cooper & U Sauro (Eds.), Gypsum Karst of the World (pp.
207—224) International Journal of Speleology Theme issue 25
(3-4).
Klimchouk, A. B. (2002). Evolution of karst in evaporites.
In F. Gabrovsek (Ed.), Evolution of karst: From prekarst to
cessation (pp. 61—96). Postojna-Ljubljana, Slovenia: Zalozba
ZRC.
Klimchouk, A. B., Ford, D. C., Palmer, A. N., & Dreybrodt, W. (Eds.),
(2000). Speleogenesis: Evolution of karst aquifers Huntsville, AL:
National Speleological Society.
Klimchouk, A. B., Lowe, D., Cooper, A., & Sauro, U. (Eds.), (1996).
International Journal of Speleology 25(3—4), 307.
Maikov, V. N., Gurkalo, E. I., Monakhova, L. B., Shavrina, E. V.,
Gurkalo, V. A., & Frants, N. F. (2001). Karst and caves of the
Pinezhje. Moscow: Ecost Association (in Russian).
Ostapenko, A.A. (2001). Underground karst forms in the sulphate depos¬
its of the Western Caucasus. PhD thesis, Krasnodar: Kubansky
University (in Russian).
Palmer, A. N. (1991). Origin and morphology of limestone caves.
Geological Society of America Bulletin, 203(1), 1—21.
Sivinskykh, P. (2009). Features of geological conditions of the
Ordynskaya Cave, Fore-Urals, Russia. In A. B. Klimchouk &
D. C. Ford (Eds.), Hypogene speleogenesis and karst hydrogeology of
artesian basins. Simferopol: Ukrainian Institute of Speleology and
Karstology.
Stafford, K. W., Land, L., & Klimchouk, A. (2008). Hypogenic speleo¬
genesis within seven rivers evaporites: Coffee cave, eddy county.
New Mexico. Journal of Cave and Karst Studies, 70(1), 47—61.
Warren, J. K. (2006). Evaporites: Sediments, resources and hydrocarbons.
Berlin: Springer.
White, W. B. (1977). Role of solution kinetics in the development of
karst aquifers. In J. S. Tolson & F. L. Doyle (Eds.), Karst hydrogeol¬
ogy (pp. 503—517). Huntsville, AL: UAH Press.
ENCYCLOPEDIA OF CAVES
374
G
GYPSUM FLOWERS AND RELATED
SPELEOTHEMS
William B. White
The Pennsylvania State University
Caves provide excellent environments for the growth
of mineral crystals. Cave conditions of temperature and
moisture remain constant for long periods of time.
Some 256 different minerals have been identified from
caves, although most are quite rare (Hill and Forti,
1997). Most of the speleothems found in caves are com¬
posed of calcium carbonate. However, minerals com¬
posed of sulfates of various kinds are found in dry
caves. The crystal growth habits of the sulfate minerals
are quite different from the habits of carbonate miner¬
als, and the resulting speleothems have a quite differ¬
ent appearance. Sulfate mineral crystals tend to grow
as large, curved fibrous bundles having the appearance
of flower petals as well as other shapes. The most com¬
mon of the sulfate minerals is gypsum, CaSC >4 • 2H 2 0,
and the commonly appearing speleothems are known
as gypsum flozvers.
EVAPORITE MINERALS IN CAVES
Calcium sulfate occurs in three stages of hydration.
Anhydrite is CaSC> 4 , the anhydrous salt. It is stable
at temperatures above 58°C and at high salinities.
Anhydrite is common in sedimentary deposits where
the high ionic strength of seawater lowers the activity
of water and stabilizes anhydrite structure. Anhydrite
is not stable in the cave environment because of the
lower temperature and the presence of freshwater.
CaSC >4 • Y 2 H 2 O is ordinary plaster but occurs only
rarely in nature as the mineral bassanite. Gypsum,
CaSC >4 ■ 2H 2 0, is the fully hydrated form. Gypsum is
monoclinic with unit cell parameters: a = 5.679 A,
b = 15.202 A, c = 6.522 A, and (3 = 118.4°. Single crystals
of gypsum are water-clear. These transparent masses
are called selenite. Selenite can be easily cleaved into
plates and slabs perpendicular to the b-crystallographic
axis, the {010) cleavage plane.
Most crystals grow at different rates in different crys¬
tallographic directions. The fast growth direction in
gypsum is along the crystallographic c-direction.
Because of the differences in growth rate, gypsum often
takes the form of long acicular crystals and fibers.
In addition to gypsum, two other sulfate minerals
have similar crystal structures and also form crusts
and "flowers": mirabilite, Na 2 S0 4 ■ 10H 2 O, and epso-
mite, MgSC >4 • 7H 2 0. These minerals are highly water
soluble, so they appear only in very dry caves.
Mirabilite is unstable in the outside atmosphere. If mir¬
abilite crystals are removed from the cave, the mineral
loses its water of crystallization to become thenardite,
Na 2 S0 4 . The large water-clear crystals of mirabilite col¬
lapse into a white powder. The behavior of epsomite is
unusual. Unlike mirabilite, which loses all ten water
molecules at once, epsomite loses water stepwise
where the first step is MgS0 4 • 7H 2 0 (epsomite) <->
MgSCh ■ 6H 2 0 (hexahydrite) + H 2 0. The conditions of
temperature and water vapor pressure for this reaction
are very close to cave conditions, so small shifts in
these parameters will cause the mineral to hydrate or
dehydrate, thus changing mineralogy depending on
small seasonal changes in the cave environment.
Other sulfate minerals that occur in caves include the
double salts bloedite, MgS0 4 ■ Na 2 S0 4 ■ 4H 2 0, and
eugsterite, CaS0 4 • Na 2 S0 4 ■ 2H 2 0. Celestine, SrS0 4/
occurs as a minor mineral associated with gypsum
deposits where seepage waters are rich in Sr 2+ . Typically
celestine appears as pale-blue crystals embedded in a
white gypsum crust. In caves where the sulfate ion is
derived from oxidation of sulfide minerals or from the
oxidation of H 2 S, there may occur aluminum-containing
sulfate minerals such as alunite, KAb,(S0 4 ) 2 (0H) 6 . Some
caves are associated with ore deposits. The oxidation
and hydration of the metal sulfides of the ore deposits
by reaction with cave waters produces a complex suite
of iron, copper, and other metal sulfate minerals
(Alpers et al. r 2000).
SPELEOTHEMS
Gypsum and the other sulfate mineral speleothems
exhibit a great variety of shapes depending on growth
rate and the source of water.
A. Bulk crystals
1. Soil gypsum
2. Dentate gypsum
3. Gypsum needles
B. Gypsum crusts
1. Granular crusts
2. Fibrous crusts
C. Fibrous gypsum
1. Gypsum flowers
2. Angel hair
3. Cave cotton
Gypsum can grow in clastic sediments to form a
variety of crystals which can reach dimensions of many
centimeters. These are sometimes elongated and other
times irregular masses. The surfaces are rough and
rarely display crystal faces, although the interior may
be clear selenite. The presence of some moisture films
on gypsum in air-filled cave passages can result in
large-faceted crystals. Bulk gypsum crystal clusters can
reach dimensions of meters and have been termed
Encyclopedia of Caves.
© 2012 Elsevier Inc. All rights reserved.
GYPSUM FLOWERS AND RELATED SPELEOTHEMS
375
gypsum chandeliers. Gypsum needles are straight, clear
crystals that grow by extrusion from the soil in some
caves (Fig. 1). At the base of each needle, where it is
inserted into the soil, is a V-shaped notch. Polarized
light microscopy shows that the needles are twins, with
the {100| twin plane emerging from the tip of the V and
forming the central plane of the needle. The V-notch,
termed a reentrant twin, provides an optimum site for
the deposition of new crystal from moisture in the soil.
Thus, the gypsum needles grow from the base and are
extruded from the soil as the crystals grow. Growth at
the reentrant V maintains a linear form so that gypsum
needles do not curve in the manner of many other gyp¬
sum speleothems. Gypsum needles have been observed
with lengths of more than one meter.
Gypsum crusts are formed on cave walls from
minute amounts of moisture seeping through pores in
the limestone rock. Some crusts have a sugary appear¬
ance and are composed of sand-sized grains of gyp¬
sum. Others have a satin appearance in cross section
and are composed of fibrous crystals arranged perpen¬
dicular to the crust. Crusts have been observed to
grow over historical names that had been scratched on
the cave walls. They grow irregularly because of the
variation in the permeability of the bedrock. Locations
where growth is faster produce blister-like bulges in
the crusts, and these may erupt to form gypsum flow¬
ers. As crusts continue to grow, they often break free
of the wall and peel away, exposing bare bedrock and
later new crusts.
Most spectacular of the gypsum speleothems are
the gypsum flowers (Fig. 2). Gypsum flowers range in
size from a few centimeters to tens of centimeters,
often with the "petals" elaborately curved. Under
the optical microscope, the petals of the gypsum flow¬
ers are seen to be bundles of fibers, each fiber a sepa¬
rate gypsum crystal. Gypsum flowers grow from the
same seepage waters that form gypsum crusts. The
solutions move faster in more porous zones and thus
the crystals grow faster near the center of the zone.
FIGURE 1 Gypsum needles, Cumberland Caverns, Tennessee.
Photo by the author.
The faster growing crystals tend to push past the
slower growing fibers at the side, thus causing the
overall mass of fibers to curve outward, producing
the curved flower petals. The fibers that make up
gypsum flowers are relatively coarse, perhaps frac¬
tions of a millimeter to a millimeter in diameter.
Much smaller and longer fibers occur as the form
known as angel hair. The fibers that make up angel
hair appear to be single crystals with lengths of a
meter or more with a cross section of a millimeter or
less. Still smaller fibers occur in a speleothem called
cave cotton. As the name implies, cave cotton is a mass
of tangled fibers generally with fiber lengths of a few
tens of centimeters but diameters of tens or hundreds
of micrometers. The thin crystals flex easily so that
the mass appears much like a ball of cotton.
Mirabilite, epsomite, and other sulfate minerals
have similar crystal habits although different crystal
structures on the atomic scale. As a result, "flowers" of
mirabilite and epsomite are also found in extremely
dry cave passages. However, the high solubility of
these minerals means that they dissolve easily in any
available seepage. Stalactites of epsomite (Fig. 3) and
mirabilite (Fig. 4) occur when there is sufficient seep¬
age to form liquid droplets. Otherwise, the minerals
occur as curved bundles of fibers or fibrous crystals. It
is difficult to distinguish the water-soluble minerals
from gypsum by visual inspection only, although they
can be distinguished by taste.
SOURCES AND DEPOSITIONAL
MECHANISMS OF EVAPORITE MINERALS
There are multiple sources for the gypsum and
other sulfate minerals that occur in limestone caves.
FIGURE 2 Gypsum flowers, Cumberland Caverns, Tennessee.
Photo by the author.
ENCYCLOPEDIA OF CAVES
376
G
FIGURE 3 Epsomite stalactites on mixed sulfate mineral crust.
Cottonwood Cave, New Mexico. Photo by the author.
1. Gypsum beds that occur within the limestone
2. Gypsum derived from the oxidation of pyrite that
occurs within the limestone
3. Gypsum derived from oxidation of pyrite and other
sulfide minerals in overlying rock formations
4. Gypsum derived from oxidation of H 2 S
5. Gypsum derived from the hydration of anhydrite
pods within the limestone
The chemistry of gypsum deposition is exceedingly
simple. Evaporation of a solution containing Ca 2+ and
SOf - causes them to combine to form gypsum.
Ca 2+ + SOf" + 2H 2 0 <-► CaSCh • 2H 2 0
The solubility of gypsum is approximately 2400 mg/L
at 10°C making it relatively easy for gypsum to dissolve
in one part of the groundwater flow system and be trans¬
ported to a site of deposition elsewhere. Beds of primary
gypsum occur in many limestone sequences; thus they
can provide a source of gypsum that requires only trans¬
port to its final site of deposition.
Pyrite, FeS 2/ occurs widely distributed in many sedi¬
mentary rocks. It is unstable in wet, oxidizing environ¬
ments. Although the details are complicated, the overall
reaction is:
FeS 2 + 7/2 0 2 + H 2 0 <-► Fe 3+ + 2S0 2 + 2H+
Ferric iron is highly insoluble in mildly alkaline
karstic waters and precipitates as Fe(OH) 3 so that iron
FIGURE 4 Water-clear mirabilite stalactites hanging on massive
gypsum crystals. Turner Avenue, Mammoth Cave, Kentucky. Photo
by the author.
rarely migrates far from the site of the original pyrite.
The reaction can be more complicated because Fe 3+
can also break down more pyrite with the release of
additional acidity. Microorganisms often catalyze these
reactions. The sulfuric acid produced by pyrite oxida¬
tion can either react with limestone directly to produce
gypsum which is then transported in solution, or the
reaction can be delayed until the reactants reach the
walls of the cave passage where gypsum is deposited
in situ. In either case the combined reaction is:
H 2 0 + 2H+ + SO 2 + CaC0 3 <-► CaS0 4 • 2H 2 0 + C0 2
Some limestones contain nodules of anhydrite incor¬
porated within the limestone. When these are exposed
by the formation of cave passages, the anhydrite reacts
with water to form gypsum.
Some caves, particularly those in the Carlsbad area
of New Mexico, have been formed by the dissolving
action of sulfuric acid rather than by carbonic acid,
which is the usual mechanism in shallow groundwater
caves. The source of sulfuric acid is the oxidation of
hydrogen sulfide, H 2 S, which seeps upward from
ENCYCLOPEDIA OF CAVES
GYPSUM FLOWERS AND RELATED SPELEOTHEMS
377
nearby petroleum fields until it encounters oxygen-
rich groundwater. The sulfuric acid dissolves the lime¬
stone and also can deposit quantities of gypsum which
can then be redistributed to form gypsum speleothems
(Hill, 1987).
It is difficult to determine the details of gypsum
deposition because the solutions from which the crys¬
tals grow seep so slowly from pores in the rock that no
collectable liquid is observed. Evidence for relatively
rapid growth is provided by the observation of gyp¬
sum crusts overlying historic signatures that had been
scratched in the cave wall. Early Americans who
inhabited the Mammoth Cave region mined Mammoth
Cave and nearby Salts Cave for gypsum and other
salts. In some places where the ceiling had been
scraped down, the bedrock now shows growths of
mirabilite that must have formed during the several
thousand years since the paleomining operation.
Bibliography
Alpers, C. N., Jambor, J. L., & Nordstrom, D. K. (2000). Sulfate
minerals: Crystallography, geochemistry and environmental sig¬
nificance. Reviews of Mineralogy and Geochemistry, 40. 608 pp.
Hill, C. A. (1987). Geology of Carlsbad Cavern and other caves in the
Guadalupe Mountains, New Mexico and Texas. New Mexico
Bureau of Mines and Mineral Resources, 117. 150 pp.
Hill, C., & Forti, P. (1997). Cave minerals of the world (2nd ed.).
Huntsville, AL: National Speleological Society.
ENCYCLOPEDIA OF CAVES
HELICTITES AND RELATED
SPELEOTHEMS
Donald Q. Davis
National Speleological Society
SUB AERIAL HELICTITES
Helictites are elongated speleothems that, unlike sta¬
lactites, may grow in any direction. Upward-growing
helictites have sometimes been called heligmites, but
there is little logical basis for this distinction, because
helictites do not occur as separate "up" and "down"
forms. Helictites may be straight, smoothly curving, or
even spiral ( helical , the root meaning of helictite),
but in most cases they twist and turn erratically.
Accordingly, the alternative names erratics, eccentrics,
or eccentric stalactites have been used by some authors.
Helictites are usually composed of calcite or aragonite,
more rarely other minerals. They occur in a great range
of sizes, from hair-thin and a fraction of a centimeter
long to several centimeters wide and more than a
meter in length. All share the common characteristic of
a narrow central canal of capillary size.
Because helictites are so conspicuous and unusual
in appearance, many (often fanciful) theories have
been proposed for their origin. By growing artificial
helictites of sodium thiosulfate. Huff (1940) demon¬
strated that hydrostatic pressure feeding capillary flow
was the true mechanism. In natural carbonate helic¬
tites, the tip is extended by deposition of calcium car¬
bonate around the central pore as the outflowing
moisture evaporates or loses carbon dioxide. Moore
(1954) subsequently explained helictite curvature by a
combination of effects of impurities, crystallographic-
axis rotation, and stacking of wedge-shaped crystals.
These factors take precedence over gravity because the
rate of flow is too slow to form a hanging drop at the
tip. Increased flow can cause helictites to convert to
soda-straw stalactites, and decreased flow, vice versa.
Varieties including filiform (hair-like), vermiform
(worm-like), and antler' (forking) helictites have been
defined on the basis of shape and size. Aragonite helic¬
tites may be beaded (Fig. 1), consisting of a string of
conical beads of radiating fibrous crystals. The larger
ends of the cones may face either the attached end of
the helictite or the free end, but the orientation is usu¬
ally consistent within each chain. This bizarre-looking
beaded structure is relatively rare and has never been
explained.
ANTHODITES
Anthodites (Fig. 2) were defined by Henderson (1949)
from Skyline Caverns, Virginia, as "clusters of slender,
branching, tubular formations originating near a com¬
mon point and radiating outward and downward." The
term antliodite, which simply means "flower rock," has
come to be used in conflicting ways, with many writers
applying it to branching aragonite sprays that lack central
canals. For the latter speleothem, the term frostivork,
which was used in the Black Hills (South Dakota, U.S.A.)
before 1900, has much earlier priority. The first edition of
Cave Minerals of the World called frostwork "acicular
anthodites" and speleothems of the Skyline Caverns type
"quill anthodites." The second edition (Hill and Forti,
1997) treated anthodites and frostwork as separate spe¬
leothem types (following a discussion exchange in the
National Speleological Society Bulletin).
Anthodites, in Henderson's original sense, are not
very clearly distinct from helictites. Occurrence in clus¬
ters in which they radiate or curve away from a com¬
mon base seems to be the only distinct morphological
trait characterizing "anthodites." Typical helictites are
usually more independent and less smoothly curvilin¬
ear. These macroscopic differences reflect the growth
habit of the original primary mineral. Anthoditic
Encyclopedia of Caves.
379
© 2012 Elsevier Inc. All rights reserved.
380
H
FIGURE 1 (A) Spectacular display of beaded helictites. (B) Detail
of beaded helictites. Photos courtesy of Kevin Doumey. Used with
permission.
morphology is generally seen in examples that origi¬
nated as aragonite; helictitic morphology in those orig¬
inating as calcite. Alternation between mineralogies,
causing overgrowth of one mineral by the other, as
FIGURE 2 Anthodites in Skyline Caverns, Warren County,
Virginia. Photo courtesy of William B. White. Used with permission.
well as conversion of aragonite to calcite, results in
intermediate or transitional forms that are difficult to
classify and interpret.
FROSTWORK
Frostwork (Fig. 3) consists of sprays of acicular ara¬
gonite needles, often intricately branching. Like helic¬
tites, these needles are not gravitationally oriented and
may grow in any direction. Flowever, because it lacks
internal feeder channels, frostwork is genetically dis¬
tinct from helictites and from Henderson's anthodites,
although all of these may be found in proximity to
each other.
Frostwork apparently grows from seepage of a thin
film of carbonate-bearing moisture over the crystal sur¬
faces, although growth from aerosols has also been
suggested. It is usually found in sections of caves
below dolomitic bedrock and where evaporation pre¬
vails. These conditions encourage enrichment of mag¬
nesium in wall moisture, facilitating aragonite growth.
Relatively high cave temperature is also favorable but
not essential. When evaporation has concentrated the
solution sufficiently that magnesium minerals can also
precipitate, the aragonite needles may be tipped with
blobs of hydromagnesite moonmilk. Individual frost¬
work "trees" may grow up to half a meter long.
Composite masses can form even larger conical stalag-
mitic growths of intertwined aragonite, which are
often hollow where drips have passed through the
center. Such hollow aragonite stalagmites have been
called logomites in the Black Hills caves. Where not
soiled, frostwork is almost always snowy white, rarely
incorporating colored contaminants. It is among the
most aesthetically impressive cave decorations.
ENCYCLOPEDIA OF CAVES
HELICT1TES AND RELATED SPELEOTHEMS
381
FIGURE 4 Active subaqueous helictites in a pool in Lechuguilla
Cave, New Mexico. Photo courtesy of Kevin Downey. Used with
permission.
FIGURE 3 Aragonite frostwork tree in Lechuguilla Cave,
New Mexico. Photo courtesy of Bnllman/Downey/Widmer. Used with
permission.
CAVE SHIELDS
Shields, also called palettes, are disk-shaped spe-
leothems consisting of paired circular or oval plates
separated by a medial crack, which is an extension of a
fracture in the rock substrate. Like helictites, they may
be oriented in any direction. The largest are more than
3 m wide. Shields and helictites, though different in
appearance, are closely related genetically. From a geo¬
metric viewpoint, shields can be regarded as flattened,
biaxial "helictites" in which growth occurs along the
margin of a plane rather than the end of a line. Shields
and helictites are similar in being fed by seepage
under hydrostatic pressure; an active shield in France
was observed jetting five arcs of water from its edge.
To sustain flow from a shield's extended margin
requires a higher rate of discharge than in a helictite.
Shields often grow helictites from their upper plates,
and draperies from overflow from the lips of the lower
plates.
The medial crack in a shield is usually of capillary
size along the edge, but may be wider, up to several
millimeters, internally. This may be caused by crystal-
growth wedging along the active edge pushing the
plates slightly apart. Some have suggested that flexing
by Earth tides prevents the cracks from closing up;
however, self-generated wedging may be all that is
really required. Shields that have split open, showing
the internal surfaces, display concentric growth rings,
and in some cases, also closely spaced horizontal ribs,
2 mm or so wide, that presumably reflect former fluc¬
tuating water levels inside the shield.
Shields are sometimes confused with flowstone can¬
opies undermined by sediment removal; canopies usu¬
ally curve and have no medial crack. Shields are found
in fewer caves than helictites, but are often abundant
where they do exist. They are probably favored by low
primary permeability in the substrate rock, forcing
seepage to be localized along fractures; highly perme¬
able walls are more likely to grow helictites. Welts are
incipient shields—linear excrescences along cracks
where flow has not become confined to particular
sections.
SUBAQUEOUS HELICTITES
Until 1987, all helictites were assumed to grow sur¬
rounded by air. In that year, spectacular displays of
spaghetti-like helictites were discovered in Lechuguilla
Cave, New Mexico, United States, growing from the
undersides of shelfstone and from subaqueous calcite
crust in pools (Fig. 4). They are usually 2—5 mm wide,
but sometimes more than 1 cm, and up to 30 or more
centimeters long. By the year 2000, more than 30 sites
with active or "dead" subaqueous helictites had been
ENCYCLOPEDIA OF CAVES
382
H
recorded in Lechuguilla Cave, which remains almost
the only cave in which they are known (a few small
ones have been found in Virgin Cave in the same area).
Like subaerial helictites, subaqueous helictites have
tiny central canals, but the mechanism of capillary seep¬
age under hydrostatic pressure, with CO 2 degassing or
evaporation at the tip, does not apply in the pool envi¬
ronment. Instead, subaqueous helictites have invariably
been found where blocks or crusts of gypsum are in
contact with flowstone upflow from the pool basins. It
is now generally accepted that the subaqueous helictites
were produced by the common-ion effect, in which
stringers of gypsum-enriched water flowed into pools
already saturated with calcium bicarbonate, triggering
deposition of the less soluble calcium-mineral species
(calcite) where the calcium-rich solutions mixed. Water
analyses at the type locality were consistent with this
mechanism (Davis et al., 1990). The initial result is
development of shelfstone and crust barriers between
the incoming and ambient waters, with further inflow
being restricted to small pores around which rings of
calcite grow. Extension of these rings into the pool cre¬
ates the helictites. Subaqueous helictites grow generally
downward, as is appropriate to the greater density of
the entering gypsum-rich solution, though some turn
upward toward the ends.
Lechuguilla Cave is almost the only known sub-
aqueous-helictite site because it is exceptionally well
supplied with gypsum left over from its sulfuric-acid
speleogenesis, some of it in contact with flowstone¬
growing seepage. This is a rare and unstable situation.
SUBAQUEOUS “HELICTITE BUSHES”
In certain passages of Wind Cave, South Dakota,
U.S.A., there are clusters of intertangled helictitic cal¬
cite speleothems (Fig. 5), typically about 0.5 m or less
long, but a maximum of 2 m. They almost always
extend upward from discontinuities in floor crust,
though the largest clump grows downward from a
passage-end pinch. All yet observed appear to be inac¬
tive. These were informally called helictite bushes by the
cavers who observed them, and were assumed for
many years to be subaerial helictites of unusual form.
After the discovery of Lechuguilla Cave's subaque¬
ous helictites, Davis (1989) examined the Wind Cave
bushes and concluded, from several kinds of evidence,
that they represent another type of subaqueous spe-
leothem, developed below the water table at a time
between the end of speleogenesis and the drainage of
the passages. They are quite distinct from all other
helictites, either subaerial or subaqueous. They are fre¬
quently strap-shaped and irregularly flattened in cross
section (in contrast to the cylindrical outline typical of
FIGURE 5 Helictite bush of subaqueous origin in Wind Cave,
South Dakota. Calcite rafts demonstrate former submergence. Photo
courtesy of Kevin Downey. Used with permission.
helictites), and often branch dendritically. Their inter¬
nal canals are not capillary tubes, but relatively large,
irregular conduits that may be more than 1 cm wide
and that more or less mirror the convolutions of the
speleothem's cross section. It is unlikely that such
structures grew upward in a subaerial environment.
The distribution of helictite bushes shows no corre¬
lation with sources of downward-seeping water. They
are largely confined to a prominent dip-oriented pas¬
sage trend that bisects the main part of the maze cave
and are always near the lowest level in any particular
section. The typical upward growth suggests that the
incoming fluid was less dense than the surrounding
ambient water (in contrast to the Lechuguillan case).
Davis (1989, 1991) proposed that helictite bushes grew
by interaction of near-surface groundwater, possibly
via the common-ion effect, with thermal fluid rising
from depth along a fracture set. In this scenario, helic¬
tite bushes would be more closely related to geyser-
mites and submarine "white smokers" than to other
helictite types. They may be low-energy analogs of the
subaqueous tufa towers and "sand tufa" of Mono
Lake, California (Davis, 2004).
Chemical and isotopic analyses of helictite-bush frag¬
ments by LaRock and Cunningham (1995) failed to con¬
firm a deep-seated origin of the hehctite-growing fluid, but
did indicate growth in a cooling thermal environment at
ENCYCLOPEDIA OF CAVES
HYDROGEOLOGY OF KARST AQUIFERS
383
33—42°C, approximately 200,000 years ago. Their speci¬
mens showed a multiple-layered structure: an inner
tube of crystals with c-axes parallel to the speleothem's
axis, an outer coating of "palisade" crystals at right
angles, and a transitional layer between. The transition
and palisade layers were repeated a number of times in
some examples, perhaps recording cycles of emergence/
submergence. Fossil microbial filaments, most abundant
in the inner layer, suggest a microbiological role in bush
growth.
LaRock and Cunningham (1995) found no analytic
evidence for common-ion-effect involvement, and sug¬
gested that the calcite was deposited via CO 2 degassing
at the overlying water surface. They proposed that the
flow inside the bush branches was driven not by fluid
rising from depth, but by water-table fluctuations, caus¬
ing a pressure gradient toward the passage to develop
in water trapped behind the passage-lining crust when
the water level in the open passage dropped faster than
that behind the crust. Their interpretation does not
explain why the bushes are largely confined to a nar¬
row passage trend, whereas the crust lining is much
more extensive in the cave, or why the phenomenon
has not been seen in other caves with subaqueous crust
linings. These unique speleothems need further study.
Bibliography
Davis, D. G. (1989). Helictite bushes: A subaqueous speleothem?
National Speleological Society Bulletin, 51(2), 120—124.
Davis, D. G. (1991). Wind Cave helictite bushes as a subaqueous spe¬
leothem: Further observations. Geer, 19(1), 13—15.
Davis, D. G. (2004). Helictite bushes, tufa towers, and inverted rim-
stone dams. Geo 2 , 31(2), 13—19.
Davis, D. G., Palmer, A. N., & Palmer, M. V. (1990). Extraordinary
subaqueous speleothems in Lechuguilla Cave. National Speleological
Society Bulletin, 52(2), 70—86.
Henderson, E. P. (1949). Some unusual formations in Skyline
Caverns, Virginia. National Speleological Society Bulletin, 11, 31—34.
Hill, C. A., & Forti, P. (1997). Cave minerals of the world (2nd ed.).
Huntsville, AL: National Speleological Society.
Huff, L. C. (1940). Artificial helictites and gypsum flowers. Journal of
Geology, 48(6), 648-659.
LaRock, E. J., & Cunningham, K. I. (1995). Helictite bush formation
and aquifer cooling in Wind Cave, Wind Cave National Park,
South Dakota. National Speleological Society Bulletin, 57(1), 43—51.
Moore, G. W. (1954). The origin of helictites, Occasional Paper No. 1.
Huntsville, AL: National Speleological Society.
HYDROGEOLOGY OF
KARST AQUIFERS
William B. White
The Pennsylvania State University
An aquifer, according to standard definitions, is a rock
unit capable of storing and transmitting water. Both
properties are necessary. The open spaces within the
rock mass provide the storage space. These open
spaces—pores, fractures and, in the case of karst aqui¬
fers, pipe-like conduits—must be interconnected so that
water can flow through. In order to appreciate the
special complications of aquifers that contain cave-size
openings, it is necessary to first describe the more
common porous media aquifers. We then proceed to
the special properties of karst aquifers. These are pre¬
sented as a conceptual model, that is, a picture or car¬
toon of the parts and pieces of a karst aquifer, how
they interrelate, and how they function. The purpose
of a conceptual model is to set a foundation on which
more precise and quantitative descriptions of aquifer
behavior can be built.
POROUS MEDIA AQUIFERS
In porous media aquifers, pores between mineral
grains are the open spaces available for the storage of
water. Porosity is defined as the ratio of the volume of
pores to the volume of bulk rock and is usually
expressed as a percentage. Pore space by itself is insuf¬
ficient. There must also exist pathways for the water to
move through the pores so that overall the rock will
transmit water. The ability of a rock to transmit water
is referred to as its permeability. It is possible to have a
rock, for example, a volcanic pumice, that has a high
porosity but an almost zero permeability. Many types
of rocks can be effective aquifers, but not all of them.
Shales typically have both low porosity and low per¬
meability. Rocks that do not transmit water are called
aquicludes. They frequently serve as barriers or confin¬
ing layers in the movement of groundwater.
In addition to the pore spaces between mineral
grains, many rock units have been fractured so that
the bulk rock is laced with cracks, known as joints, and
more sparsely with clusters of joints, known as frac¬
tures or joint swarms. In addition, there may be partings
where the rock has separated along bedding planes.
The term fracture is used loosely for all of these
mechanical openings. Fractures also act as pathways
for groundwater movement so that one can speak of
matrix permeability referring to the interconnected pores
and fracture permeability referring to joints, larger frac¬
tures, and bedding plane partings. In some rocks, for
example, fractured granites and fractured massive
basalts, fracture permeability completely dominates
matrix permeability.
Aquifers are usually named with the name of the
rock formation in which they occur. Rock formations
extend indefinitely along the beds but the beds have a
definite thickness. The aquifer thickness enters into
various calculations but the lateral extent does not.
Encyclopedia of Caves.
© 2012 Elsevier Inc. All rights reserved.
384
H
Some aquifers are local; others extend over hundreds
of kilometers. Descriptions of the principal aquifers in
the United States have been compiled into the Ground
Water Atlas by the U.S. Geological Survey (United
States Geological Survey, 1992—1998).
Groundwater in porous media aquifers moves very
slowly, typically a few meters per year. The velocity is
proportional to the hydraulic gradient according to
Darcy's law
The proportionality constant, K, has units of velocity
and is called the hydraulic conductivity. The hydraulic
conductivity is related to properties of the rock, prop¬
erties of the fluid, and the strength of gravity that
provides the ultimate driving force moving the water.
If the cross-sectional area. A, of the bit of aquifer being
examined is specified, Darcy's law also describes the
volume of water moving across the specified area per
unit time. Combining these ideas
most porous media aquifers, groundwater and sur¬
face water are disconnected. In a rainstorm, some
water soaks into the ground as infiltration, I. Some
water runs off over the land surface to enter creeks
and rivers, which quickly drain the water away as
runoff, R. Some water is lost back to the atmosphere
by evaporation and by the transpiration through the
leaves of plants. Generally, these factors are taken
together as evapotranspiration, E. Over long time
periods, these factors must balance
P — E = I + R
Surface water flows in streams and rivers at veloci¬
ties of fractions of a meter per second whereas ground-
water moves at velocities of meters per year. The two
can be treated as almost separate entities. The same is
not true in karst aquifers.
KARST AQUIFERS
Q = vA
-A
N d 2 p g dh
77 dl
The parameter d is the mean diameter of the mineral
grains and N is a geometric factor to take account of the
irregular shape of the grains. Nd 2 is the permeability, a
property of the rock. Permeability and hydraulic con¬
ductivity are often confused. Because both measure the
ability of the rock to transmit water, they are sometimes
used—incorrectly—interchangeably. Hydraulic conduc¬
tivity has units of velocity; permeability has units of
area. Darcy's law is valid for other fluids (such as petro¬
leum) as well as water. The properties of the fluids are
described by their density, p, and viscosity, rp both
functions of temperature. The gravitational force is
described by g, the acceleration due to gravity.
Darcy's law is one of the most powerful statements
in all of hydrology. It underlies most of the theory of
groundwater flow. When combined with statements
of the conservation of water within a unit volume of
the aquifer, the result is a differential equation that
describes the groundwater flow field. Many pages of
standard textbooks on groundwater hydrology deal
with the boundary conditions and solutions of these
equations. The results take the form of flow fields,
techniques for testing water wells, prediction of con¬
taminant plumes, and many others. One of the most
fundamental distinctions between karst aquifers and
most porous media aquifers is that in important parts
of the karst aquifer, Darcy's law does not work.
Another aspect of karst aquifers that differs
remarkably from normal aquifers is in the relation¬
ship between groundwater and surface water. In
The Triple Porosity Model
The internal structure of a karst aquifer contains three
types of permeabilities and is thus sometimes described
by what is called the triple porosity or triple permeability
model.
1. Matrix permeability is due to the pore spaces
within the bedrock. Matrix permeability in karstic
aquifers is not intrinsically different from the matrix
permeability in porous media aquifers.
2 . Fracture permeability is due to joints, joint swarms,
and bedding plane partings. Again, these are not
intrinsically different from fractures in sandstone,
granite or other massive rocks except for their
tendency to be enlarged by dissolution of the
bedrock.
3. Conduit permeability consists of solutionally
generated caves or conduits in the aquifer bedrock.
Flow of water in conduits is similar to flow in pipes.
The flow regime may be turbulent. The flow
velocities are in the range of centimeters per second
compared with meters per year in porous media
aquifers. Water is confined to the conduits with
little dispersion. As a result, contaminants are not
dispersed to form a plume. Instead, they are carried
down the conduit to a spring with little dilution
except for incoming tributary passages.
The properties of each component of the permeabil¬
ity are summarized in Table 1. Many Paleozoic lime¬
stones and dolomites have very little matrix porosity,
while young limestones that have not been subject to
the heat and pressure of deep burial are quite porous.
ENCYCLOPEDIA OF CAVES
HYDROGEOLOGY OF KARST AQUIFERS
385
TABLE 1 Characteristics of the Three Components of the Triple-Permeability Model
Permeability
Aperture
Travel Time
Flow Mechanism
Guiding
Equation
Distribution
Matrix
|im—mm
Long
Darcian flow field: laminar
h - rpL
f pg(Nd 2 )
Continuous medium
Fracture
10 pm—10 mm
Intermediate
Cube law: mostly laminar, may be
nonlinear components
a = c g 3
i, f b
Localized but statistically
distributed
Conduit
10 mm—10 m
Short
Darcy—Weisbach: open channel and
pipe flow turbulent
i fLv 2
h f~w
Localized
Surface Streams/
Conduit Flow
GRANITES DOLOMITES SANDSTONES AQUIFERS
FIGURE 1 The three types of aquifer permeabilities
expressed as a triangular diagram. Normal aquifers
would lie along the bottom of the triangle. Karst aqui¬
fers are those in which conduit permeability is an
important component.
Limestones and dolomites tend to be hard, brittle rocks
and are often highly fractured.
All aquifers can be categorized by the triple porosity
model, as illustrated in Figure 1. Various nonkarstic
aquifers such as fractured sandstones, fractured granites,
and gravel aquifers lie along the base of the triangle
where there is little or no contribution from dissolution
processes in the development of the aquifer. Dense, brit¬
tle carbonate rocks form aquifers that lie along the side
of the triangle between conduit flow and fracture flow.
Aquifers in young carbonates such as the Floridan aqui¬
fer contain flow systems that are a mix of matrix and
conduit flow. There are aquifers, such as the important
Edwards Aquifer in central Texas, that contain contribu¬
tions from all three types of permeabilities.
Conduit Permeability
The feature of karst aquifers that makes them dis¬
tinct from other aquifers is the presence of conduits,
which act as networks of pipes carrying water rapidly
through the aquifer. The degree of conduit develop¬
ment is extremely variable. The connection of the
conduits to the other types of permeabilities is also
highly variable. As a result, there is a range of aquifer
types, as illustrated in Figure 2, with very different
roles played by the conduit system.
At one end of the sequence (A) is the downstream
reach of a sinking stream with its catchment mostly as
surface streams on noncarbonate rock. In this case
the conduits take the form of a drainage tunnel. Water
moves through the tunnel at velocities similar to those in
surface streams, in one end and out the other. There are
only weak connections between the water flowing in
the conduit and the fracture and pore water in the sur¬
rounding rock. When a larger fraction of the drainage
basin is karstic, sinkhole drains and smaller conduits con¬
verge to form master trunk systems (B). Master trunk and
branchwork conduits form where the flow of water is
under moderate to high hydraulic gradients. These con¬
duit systems behave hydraulically much like surface
streams except that they happen to be underground.
ENCYCLOPEDIA OF CAVES
386
Surface Water *+
► Ground Water
M - Allogenic Recharge
Travel Time
FIGURE 2 Sketches showing various types of conduit permeabilities: (A) Single conduit with mainly surface water input; (B) branchwork
conduits with sinkhole and sinking stream inputs; (C) anastomotic maze of interconnected tubes developed along bedding plane partings;
(D) network maze developed along vertical fractures; (E) permeability consisting of solutionally widened fractures; and (F) matrix permeabil¬
ity only. Flow is most localized at (A) and most diffuse at (F). Travel times increase left to right. The sequence moves from a surface stream
with a roof at (A) to a porous-medium groundwater aquifer at (F).
When gradients are low or where the free flow of
the conduit system is impeded, water circulates along
multiple pathways, gradually dissolving out either
networks of joints or multiple loop openings along
bedding plane partings. Dissolution along fractures
leads to network mazes (C); circulation along bedding
plane partings leads to anastomotic mazes (D). If
master trunk and branchwork conduit systems can be
likened to surface streams, network and anastomotic
mazes can be likened to swamps. Like swamps, much
water can move through these multiple pathways but
at very low velocities. The sequence continues with
aquifers of less and less karstic character to fracture
aquifers (E) and porous media aquifers (F). Real aqui¬
fers can be much more complicated depending on
their geology and how they have evolved over time.
The flow through conduits is generally turbulent
whereas the flow in fractures and pores is generally
laminar. Flow velocity increases with conduit size.
Under usual hydraulic gradients, the onset of turbulence
occurs when conduit apertures reach about 1 centimeter.
We may take one centimeter—either the width of
a solutionally widened fracture or the diameter of a
tubular conduit—as the threshold for conduit flow
behavior. Continuous flow paths through the rock with
an effective width of one centimeter or greater would be
considered conduit permeability. Smaller openings
would be fracture or matrix permeability.
Caves are not exactly equal to conduits. Caves are
explored by humans, which gives them a minimum
dimension of roughly 0.5 m whereas with a minimum
aperture of 1 centimeter there are many conduits that
are too small to be called caves (or, at least, are too
small to be explored by humans). Conduits are orga¬
nized to provide continuous flow paths because for
them to function as permeability in the aquifer, water
must be able to move continuously from input to out¬
put. Caves are often occluded and are discontinuous;
caves are fragments of conduits. Active conduits are
located near or below the water table. The dry caves
that lie above the active zone of groundwater
circulation are fragments of abandoned conduits.
Water flows through conduits either as open chan¬
nel flow (a stream with a free air surface), or as pipe
flow (a completely water-filled tunnel). A single con¬
duit may have sections with open stream passage and
sections that are water-filled. Exploration of cave
sumps by scuba divers has shown that conduits
ENCYCLOPEDIA OF CAVES
HYDROGEOLOGY OF KARST AQUIFERS
387
BASE-
LEVEL
I STREAM
i
FIGURE 3 A conceptual model for a carbonate aquifer showing various inputs and flow paths. Any particular groundwater basin may or
may not contain all of these features.
frequently undulate in the vertical plane with alternat¬
ing free surface streams and sumped passages.
Karst Groundwater Basins and Groundwater
Flow Systems
Because of the rapid transmission of water through
conduits and because of the rapid drainage of surface
water into the aquifer through sinkholes and surface
streams, it is necessary to combine both surface water
and groundwater characteristics. The result is the
concept of a groundwater basin which includes both
surface and subsurface components. Combining water
sources with the internal characteristics of the aquifer
and with the discharge of water back to surface routes
produces the standard model of karstic aquifers illus¬
trated in Figure 3.
The standard model contains the following sources
of recharge:
1. Dispersed (or diffuse) infiltration. Rainfall onto
the karst surface infiltrates through the soils,
the epikarst, underlying fractured bedrock, to
eventually reach the water table and the active
conduit system. This component is not intrinsically
different from infiltration in any aquifer.
2. Allogeneic recharge from surface streams. Mostly, these
drain from nearby drainage basins on noncarbonate
rocks. Allogeneic recharge enters the aquifer at
well-defined swallets. Sometimes these are cave
entrances, pits, or debris piles in blind valleys, and
sometimes they are just sections of streambed
where water is gradually lost. Flood flow in
allogeneic basins injects large pulses of water into
the aquifer on short time scales. There is little
or no filtration where surface streams become
groundwater. Dissolved contaminants, suspended
contaminants, and even massive objects such as
cans, old tires, household garbage, and dead
animals can be swept into the aquifer during
periods of high flow.
3. Internal runoff. The land surface above karst aquifers
is often pocked with closed depressions of various
sizes. Some of these are soil mantled, some have
exposed bedrock, and a few expose an underlying
underground stream—often called karst windows.
Some closed depressions have collapses in the
bottom where soil piping has occurred. During
moderate storms, the rainfall infiltrates through the
soils into the underlying aquifer through fractures
and bedding plane partings. However, there are
often no surface runoff channels in sinkhole
topography. Overland flow during storms drains
to the bottoms of the closed depressions to make its
way into sinkhole drains and thus directly into the
underlying aquifer. Storm runoff can pond in the
ENCYCLOPEDIA OF CAVES
388
H
closed depressions and the seepage pressure can
induce soil piping failures. Soil piping failures inject
volumes of clastic sediment into the aquifer along
with any contaminants that may have accumulated
in the sinkhole.
Piracies and Spillover Routes
Unlike surface water basins, groundwater basins rarely
have rigorously fixed boundaries. Escape routes along
conduits that cross the basin boundaries into adjacent
drainage basins are common. These are known as piracy
routes and are most common in locations where the
groundwater basin seems to be defined by the overlying
surface water basin. Overflow routes on the land surface
converge to a master stream that is often the location of
the spring from which the groundwater discharges.
However, certain surface streams that sink underground
within the surface water basin may cross surface divides,
sometimes under ridges or mountains, and appear in
springs far outside the original basin. Such piracies can
be identified by tracer studies, but these must be
conducted with great care because dyes injected into
swallow holes may reappear in completely unexpected
places.
Karst aquifers evolve over time. New conduits are
formed at depth and old conduits are drained and
abandoned. The old conduits may have converged on
a spring in a particular surface water basin. The new
conduits may have shifted the flow path so that the
present system drains to a new spring which may
be in a different surface water basin. Under low-flow
conditions, the present-day basin may be well behaved
with all recharge draining to the active spring. During
flood flow, however, rising water levels within the
conduit system may activate some of the abandoned
conduits. The result is that floodwaters are diverted
into the old drainage system. Such spillover routes are
also a common feature of karst aquifers.
The Karst Water Table
In a porous medium aquifer, rainfall infiltrates
through the soil into the underlying bedrock. It moves
vertically through the pore spaces in the rock, displac¬
ing the air that occupied the pores. This portion of the
aquifer is known as the vadose or unsaturated zone. The
infiltrating water continues to move vertically until it
reaches a portion of the aquifer where all pore spaces
are water-filled. This portion of the aquifer is known
as the phreatic or saturated zone. The interface between
the vadose and phreatic zones is the water table. In a
porous medium aquifer, the water table is a surface
that stands higher under topographic highs and lower
under topographic lows. The water table slopes from
topographic highs to surface streams which are the
locations of groundwater discharge. The slope of the
water table is supported by the hydraulic resistance
for groundwater moving through the pore spaces of
the rock.
In a karst aquifer, the open pipes of the conduit sys¬
tem have a very low hydraulic resistance. The presence
of conduits creates a trough in the water table. The
water table in the fracture and matrix permeability in
the nearby bedrock slope toward the conduits rather
than to the surface streams. The conduit system acts as
a drain for the entire aquifer. Sinkholes and sinking
streams allow very rapid input of storm water into the
aquifer. The groundwater trough fills quickly, often
flooding the conduit systems.
In terrain with low to medium relief, the karst
water table is well defined. The location of the water
table can be determined by measuring the elevation of
standing water in wells. If there is a sufficient density
of wells, the standing water elevations can be con¬
toured to construct a map of the water table surface. In
high relief terrain, the position and even the existence
of the water table is more difficult to define. Recharge
water from rainfall and snow melt in mountainous
regions often descends through the vadose zone as
waterfalls in shafts and steep gradient free-surface
streams which may cross over each other. The
presence of even a large free surface stream does
not necessarily mark the location of the water table.
Furthermore, sumped passages may occur that are
perched above regional base levels, and these also do
not necessarily indicate the location of the water
table either.
CHARACTERISTICS OF
KARST AQUIFERS
Discharge Characteristics: Hydrographs
Karst aquifers usually discharge through a relatively
small number of large springs. The discharge, turbid¬
ity, and chemistry of water from karst springs provide
information on the generally inaccessible processes
upstream in the groundwater basin.
Surface drainage basins respond very rapidly to
storms. A sudden intense storm will send a flood
down the valley of the surface stream. Depending on
the degree of development of the conduit system,
groundwater basins may respond much the same way.
If there is a well-developed conduit system, storm
inputs through sinking streams and internal runoff
into sinkholes will move rapidly through the system.
The discharge at the spring will increase rapidly and
ENCYCLOPEDIA OF CAVES
HYDROGEOLOGY OF KARST AQUIFERS
389
Sept Nov Jan Mar May July
FIGURE 4 Sketches of (A) a fast response hydrograph showing
individual storm peaks and (B) a slow response hydrograph showing
only seasonal wet and dry periods. The horizontal scale is the water
year, which begins in September in the eastern United States.
the water may become muddy. Other springs, draining
groundwater basins with more poorly developed
conduit systems, will have a more subdued response.
These responses can be described by measuring the
time variation of discharge at the spring and construct¬
ing spring hydrographs (Fig. 4). Fast response springs
are those with open conduit systems so that the rise and
fall of the storm hydrograph is short compared with the
average spacing between storms. The storm flow from
the spring rises, peaks, and falls back close to base flow
before the next storm arrives. The discharge from slow
response springs may increase during wet seasons and
decrease during dry seasons, but the system responds
too slowly for the spring to be impacted by individual
storms. There are, of course, intermediate-response
springs with some storm response but with the hydro¬
graphs of individual storms smeared together.
Spring hydrographs provide several parameters
for characterizing the aquifers that feed the spring.
The ratio of peak discharge to base flow discharge,
Qmax/Qbase/ ' s a measure of the "flashiness" of the aqui¬
fer. The recession of the hydrograph often has an expo¬
nential form. It can be fitted to a function of the form
Q = Qo expf-f/r)
where Q is spring discharge in volume per unit time,
t is time, and r is the response time of the aquifer.
Plots of hydrograph recession curves for many highly
karstic groundwater basins show two (or more) seg¬
ments, a fast response time on the order of a few days,
and a slow response time on the order of 20—30 days.
One interpretation is that the fast response represents
the draining of the conduits while the slow response is
the draining of fractures and small openings that feed
into the conduits.
Carrying Capacity of Conduit Systems
When surface streams flowing on noncarbonate
rocks reach the contact with the carbonates, the amount
of surface-stream flow lost to the carbonate aquifer
depends on the carrying capacity of the conduit system.
If the conduit system is poorly developed so that it can
carry only a portion of the base flow, the stream will
remain on the surface across the carbonate rock. The
flow in the surface stream will be smaller than expected
for the area of the basin because part of the flow will be
lost into the carbonate aquifer. At the next stage, the con¬
duit system can carry base flow but not storm flow. At
this stage, the surface stream will go dry during the dry
season. With further enlargement of the conduit system,
small storms can also be accommodated. At this stage,
the surface stream will be dry most of the time and will
carry water only during flood flow. Because of disuse,
the streambed will become overgrown with vegetation,
banks will cave in, and sinkholes may develop along the
stream channel. At the final stage, when the conduit sys¬
tem can carry the most extreme flood flows, the surface
channel will eventually disappear. Upstream from the
swallet, surface flow on noncarbonate rocks will con¬
tinue to deepen the channel, resulting in a blind valley.
Base Flow/Area Relationships
The boundaries of the groundwater basins that
subdivide karst aquifers are determined by a variety of
evidence—geological constraints, tracer tests from
sinkholes and sinking streams to springs, and explored
cave passages—but the basin boundaries cannot be
measured directly. In most regions, the base flow
emerging from the spring that drains a groundwater
basin is proportional to the basin area.
Q = kA
The proportionality constant, k, varies depending on
the hydrogeology of the basin, particularly the develop¬
ment of the conduit system. Basins with well-developed
conduit systems will drain effectively between storms
so that base flows will be small. Basins with poorly
ENCYCLOPEDIA OF CAVES
390
H
developed conduit systems will store water more effec¬
tively so that base flows are maintained. A serious dis¬
agreement between basin area calculated from base
flow and basin area determined from the interpreted
boundaries is an indication of the presence of piracy
routes or of serious error in interpreting the boundaries.
Clastic Sediments
The solutionally widened fractures and conduits in
karst aquifers are sufficiently wide to permit the trans¬
port of clay, silt, sand, and gravel and even materials
up to the size of cobbles and boulders. More impor¬
tantly, the velocity of moving water is sufficient to
drag solid particles through the aquifer.
Clastic sediments are transported mainly during
flood flow episodes. Very small particles, mainly clay,
are transported in suspension, making the water turbid
or muddy. When water velocities decrease following a
storm event, the sediments in suspension settle out,
leaving a layer of mud. Coarser particles, sand and
gravel, are transported as bedload. That is, the parti¬
cles are rolled along the bottom of the conduit by the
tractive force of the water flowing over them. Bedload
transport requires a minimum velocity, which depends
on particle size, to place the particles in motion. Banks
of silt, sand, and gravel are frequently found in caves.
Streambeds in caves are often armored with a layer of
gravel and cobbles that move only during flood flows.
Transport of clastic sediment is an essential part
of the functioning of the karst aquifer. In completely
developed aquifers where all surface streams sink
underground, the insoluble material carried by the
sinking surface streams must be carried through the
conduit system and washed out through the springs.
If this did not happen, the conduit system would even¬
tually become completely clogged with sediment and
drainage would be forced to return to the surface.
WATER-SUPPLY AND WATER QUALITY
Water Wells in Karst
Groundwater for domestic or industrial use is
extracted by drilling wells. Pumping water from the
wells lowers the water level in the well (a quantity
known as drawdoum), thus creating a local hydraulic
gradient that forces water from the surrounding rock
to flow toward the well. For any given aquifer, there
will be a maximum quantity of water that can be
pumped from the well determined by the balance
between the pumping rate and the drawdown. Water
supply wells are located to provide the maximum
quantity of water with minimum drawdown.
The optimum placement of wells in karst aquifer is
difficult to determine. A well drilled into a flooded
conduit would provide large quantities of water with
negligible drawdown, but the water might be of ques¬
tionable quality. In contrast, wells drilled into unfrac¬
tured carbonate rock are often dry holes or at best
provide only small quantities of water. For those aqui¬
fers with fracture permeability consisting of vertical
fractures, wells drilled on fractures or on fracture
intersections provide the best yields. For those aquifers
with fracture permeability consisting of bedding plane
partings, these partings provide high permeability
zones that will be tapped regardless of the detailed
location of the well.
Water Quality
In the highly populated and developed countryside
of today, one learns quickly not to drink from creeks
and rivers. Municipal water companies draw water
from surface sources, but these supplies are filtered,
treated, and chlorinated before being released into
public water supplies. Water from wells and springs,
on the other hand, is often thought to be "pure."
This statement is not true in general and it is excep¬
tionally untrue when applied to karst aquifers. Karst
springs are the outlets draining open conduit systems.
Conduits receive portions of their water from sink¬
holes and sinking streams. There is little or no filtra¬
tion. Any surface contaminants that enter the conduit
system are flushed through to the springs with little
dilution. Karst springs should be regarded with the
same suspicion as surface streams.
Groundwater in carbonate aquifers tends to contain
high concentrations of magnesium and calcium. These
ions. Mg 2 + and Ca 2+ , make the water "hard." Hard
water is not a health risk, but it does require more soap
for washing and laundry. It tends to precipitate cal¬
cium and magnesium carbonates in hot-water heaters
and other plumbing fixtures and may require water
softeners for many uses. Wells drilled into fractured
carbonate aquifers are also at risk. Open fractures can
allow contaminants to migrate down from the land
surface. Groundwater in the fractures generally flows
down-gradient to the conduit system. Pumping on
wells in fractures may reverse the hydraulic gradients
and cause water to flow from the conduits to the wells,
bringing contaminants to the wells.
Contaminant Transport
Because of high velocity flow restricted to the con¬
duit system, contaminants in karst aquifers generally
do not spread out into a plume as they do in normal
ENCYCLOPEDIA OF CAVES
HYDROTHERMAL CAVES
391
aquifers. Different categories of contaminants have
different modes of transport within karst aquifers.
Types of contaminants are listed below:
1. Water-soluble compounds. Inorganic salts and some
organic compounds are soluble in water. These
move with the water in the aquifer and appear at
the springs possibly in a somewhat diluted form.
In the case of a spill, water-soluble compounds will
reach the spring in about the same time as the
travel time of the water, often a matter of hours or
at most a few days.
2. Light nonaqueous phase liquids (LNAPLs). Gasoline,
fuel oil, home heating oil, and related hydrocarbons
are less dense than water and are only slightly
soluble. These compounds will float on the water
table and will float on tree-surface underground
streams. However, they tend to pond behind sumps
and tend to be trapped in pockets in the ceilings of
water-filled conduits. Rising water levels can force
LNAPLs upward along fractures where fumes can
enter homes and other buildings.
3. Heavy nonaqueous phase liquids (DNAPLs). Chlorinated
hydrocarbons such as trichloroethylene (TCE) and
perchloroethylene (PCE)—botli used as solvents,
degreasers, and dry cleaning agents—as well as
polychlorinated biphenyls (PCB) and many other
compounds are denser than water and are only
slightly soluble. These materials are sometimes
trapped in the epikarst but when they enter the
aquifer, they tend to sink to the lowest water-filled
passages or become trapped in the clastic sediments
that occur in the conduits. As a result, spills of
DNAPLs often never reappear at springs and remain
trapped in the karst aquifer for long periods of time.
4. Metals. Metallic elements such as chromium, cadmium,
lead, and mercury are highly toxic, while others such
as copper and zinc are less so, and these can be carried
into karst aquifers either as ionic species in solution or
as solid particles of various sorts. Because of the
alkaline chemistry of karst waters, some metals are
precipitated, some are adsorbed on clay particles, and
some are incorporated into the manganese and iron
oxides/hydroxides that form coatings on cave stream
sediments. Metal transport in karst aquifers, therefore,
involves a very complex chemistry that is not easily
generalized.
5. Pathogens. Viruses, bacteria, protozoa, and larger
organisms are easily transported into karst
aquifers because of the large solution openings
and the absence of filtering. Most common of
these are the fecal coliform family of organisms
and fecal streptococci. These organisms are
indications of contamination by sewage or animal
waste. Giardia lamblia is of most concern among
protozoa. It is released in a cyst form in animal
feces and is present in many surface waters.
Sinking streams carry the stable cysts into the
subsurface. Die-off is slow underground so that
karst waters remain contaminated far from
surface inputs.
See Also the Following Articles
Modeling of Karst Aquifers
Sinking Streams and Losing Streams
Springs
Water Tracing in Karst Aquifers
Bibliography
Ford, D., & Williams, P. W. (2007). Karst hydrogeology and geomorphology
and hydrology. Chichester, U.K.: John Wiley.
Milanovic, P. T. (1981). Karst hydrogeology. Littleton, CO: Water
Resources Publications.
Milanovic, P. T. (2004). Water resources engineering in karst. Boca
Raton, FL: CRC Press.
Palmer, A. N. (1991). Origin and morphology of limestone caves.
Geological Society of America Bulletin, 103( 1), 1—21.
Palmer, A. N. (2007). Cave geology. Dayton, OH: Cave Books.
United States Geological Survey (1992—1998). Ground water atlas of
the United States. Hydrologic investigations atlas, 730-B-M, 12 folio
volumes.
White, W. B. (1988). Geomorphology and hydrology of karst terrains.
New York: Oxford University Press.
White, W. B. (2002). Karst hydrology: Recent developments and open
questions. Engineering Geology, 65(2), 85—105.
White, W. B. (2007). Groundwater flow in karst aquifers. In J. W. Delleur
(Ed.), Handbook of groundwater engineering (2nd ed., pp. 21-1—21-47).
Boca Raton, FL: CRC Press.
HYDROTHERMAL CAVES
Yuri Dublyansky
Innsbruck University, Austria
DEFINITION
The term hydrothermal karst defines a process of
dissolution of cavities in the rocks under the action of
hot waters. This definition, though quite simple, is some¬
times difficult to apply, because it requires another defi¬
nition of which water should be called hot or thermal. In
hydrological studies any water that is appreciably
warmer (5°C or more) than the surrounding environ¬
ment is called thermal. This type of definition is fairly
satisfactory when it is applied to a still-active process in
areas with a moderate climate. It is difficult to apply,
however, to those settings where neither resurgence
Encyclopedia of Caves.
© 2012 Elsevier Inc. All rights reserved.
392
H
temperatures nor annual climatic averages are applicable
(i.e., to deep-seated waters tapped by boreholes or to fos¬
sil karst process). Conventionally, the temperature of
20° C is considered to be the lower limit of the hydrother¬
mal environment. Although some meteoric karst sys¬
tems in hot arid climates may exceed this limit without
any thermal input, most hydrothermal cave systems
relate to hypogenic sources of energy (i.e., internal heat
of the Earth). Formed due to the action of rising waters,
hydrothermal systems typically lack any genetic
relationships to recharge from overlying or directly adja¬
cent surfaces. Caves must be uplifted with the rock and
intersected by surface erosion by common karst, or by
drilling, mining, or quarrying to be discovered and stud¬
ied. Hydrothermal karst is a special case of hypogene
karst.
The concept that some caves could have been
formed by ascending thermal waters rather than by
cold descending, gravity-driven ones was first put
forth as early as in the mid-nineteenth century. It has
been suggested that Pb-Zn ores in some of the
European deposits in carbonate rocks were emplaced
in dissolution cavities and that these cavities owe
their existence to the same solutions from which, at
later stages, the ores were deposited. In his "A Treatise
on Metamorphism", Charles Van-Hise (1904) provided
explanations for how and why hydrothermal solutions
move advectively through the rocks, and what causes
their aggressiveness. He conjectured that most hydro-
thermal solutions originate as common meteoric
waters that become heated during their circulation
deep in the Earth's crust. These early works laid a
foundation for the concept of hydrothermal karst.
SETTINGS OF HYDROTHERMAL KARST
It is convenient to subdivide hydrothermal karst
settings into the following three categories: endokarst,
deep-seated hydrothermal karst, and shallow hydro-
thermal karst.
Endokarst
It is a well-known fact that the rate of karst devel¬
opment decreases with increasing depth. However,
deep drilling for oil and gas reveals that solutional
porosity of carbonate rocks at depths of 4—5 km may
be as great as 18—28%, and the porosity of aluminosil¬
icate rocks may be as great as 30—35%. The pressures
of fluids at these depths are typically greater than the
hydrostatic ones and may approach lithostatic values.
At those levels, where the pressure exceeds the
strength of the rock, pores and cavities may exist only
if they are filled with high-pressure fluid, which pre¬
vents them from failure. The process of formation
of such cavities is termed endokarst (Andreychouk
et al., 2009). The size of endokarstic cavities does not
normally exceed several centimeters. Though this
type of karst apparently does not produce traversable
caves, it may play a significant role in creation of
deep-seated reservoirs for hydrocarbons.
Deep-Seated Hydrothermal Karst
This covers a range of depths (approximately 0.3 to
4.0 km) where the temperature gradients are relatively
small, pressures are close to hydrostatic, and the influ¬
ence of the temperature changes at the Earth's surface
is practically absent. Processes of cave excavation and
cave infilling occur in response to the change in physi¬
cochemical parameters of fluids moving toward the
Earth's surface, such as the decrease in temperature
and pressure. It is thought that dissolution related to
the elevated content of carbonic acid is the leading fac¬
tor in initiation and enlargement of caves in the deep
setting. Recent studies, however, revealed the impor¬
tant role of dissolution processes involving sulfuric
acid. At certain specific combinations of bedrock solu¬
bility and frequency of fractures, dissolution is accom¬
panied by collapse. Characteristic collapse breccias are
known in many fossil hydrothermal systems related to
Pb-Zn ore deposits in dolomitic rocks ( e.g ., in Silezia,
Poland).
Shallow Hydrothermal Karst
The shallow setting describes processes developing
near the free surface of the thermal water—both
below and above it. In this zone the pressures are low
(down to atmospheric) and temperatures may range
from boiling to just slightly exceeding the ambient
ones. The temperature gradients may be significant,
which leads to the appearance of some specific and
powerful processes, like thermal convection and
condensation corrosion. Also, this is a zone where
upwelling thermal waters meet colder oxidized mete¬
oric waters and atmosphere. This may induce specific
reactions and processes such as H 2 S oxidation, mixing
corrosion, and cooling corrosion. The caves formed in
such a setting commonly exhibit extremely diverse
morphologies. Characteristic dimensions of individ¬
ual caves are appreciably greater than those of
the deep-seated hydrothermal caves. Also, shallow
hydrothermal karst speleothem types are much more
varied than those of the deep-seated karst.
All enterable active hydrothermal caves (those contain¬
ing hot waters) are examples of shallow hydrothermal
ENCYCLOPEDIA OF CAVES
HYDROTHERMAL CAVES
393
karst. Examples of such caves are known in many places
including Turkmenistan (Bakharden Cave, with a lake
with temperature of 35—37° C), Italy (Grotta Giusti,
32—34°C), Mexico (Sistema Zacaton, 29—32°C), and
Hungary (Molnar Janos Cave, 18—24°C).
CHEMISTRY OF FLUIDS AND PROCESSES
OF CAVE EXCAVATION
Hydrothermal caves associated with the CO 2 and
caves formed by waters containing H 2 S are considered
two major classes of hydrothermal (hypogene) caves.
Dissolution by Rising Thermal Water (CO z )
Thermal waters rising from significant depth are
commonly saturated with C0 2 , which may originate
from metamorphism of carbonate rocks and igneous
activity. The solubility of C0 2 in water depends on
both temperature and pressure. Water saturated with
respect to C0 2 at deep levels (e.g., 2—4 km) becomes
supersaturated as it rises toward the surface. Hence,
C0 2 must exsolve in the gaseous phase and leave the
system. Rising carbonic thermal waters also cool down.
Due to inverse relationships between carbonate solubil¬
ity and temperature, they may acquire and maintain
aggressiveness, even at decreasing C0 2 levels. The sol¬
ubility of CaC0 3 increases evenly along the ascending
fluid path, but near the land surface (or water table) it
drops drastically. Such nonlinear behavior leads to the
appearance of two geochemical zones: a zone of car¬
bonate dissolution at depth and a zone of carbonate
precipitation closer to the surface.
Oxidation of Sulfides (H 2 S)
Sulfuric waters become aggressive when their dis¬
solved H 2 S oxidizes on contact with oxygen-rich
waters or air to form sulfuric acid. Conversion of H 2 S
to H 2 SO 4 produces a sharp increase in dissolution. The
effect is attenuated when C0 2 generated by the H2SO4-
CaC0 3 reaction is degassed. In two settings, hydrogen
sulfide oxidation is an important speleogenetic pro¬
cess. The first is subaqueous dissolution of carbonates
near the water table. Large rooms of the Carlsbad Cavern
in New Mexico, U.S.A., are believed to be formed this
way. The second setting is dissolution of and subse¬
quent replacement of calcite by gypsum and its conse¬
quent removal above the water table. The mechanism,
termed replacement corrosion, was suggested for caves
of the Big Horn Basin in Wyoming, U.S.A.
Dissolution Due to Mixing of Waters
(CO z and H 2 S)
Solutional aggressiveness can be renewed or enhanced
by mixing of waters with contrasting chemistry, particu¬
larly those differing in CO 2 and II 2 S content or salinity.
Mixing of waters having different temperatures produces
a similar effect, due to contrasts in C0 2 contents. This
effect, known as mixing corrosion, is thought to be respon¬
sible for the development of network maze cave systems
in Budapest, Hungary, for example, Pal-volgy, Szemlo-
hegy, and Ferenc-hegy.
Hydrothermal Karst in Noncarbonate Rocks
The mechanisms described above pertain to the
most common variety of hydrothermal karst develop¬
ing in carbonate rocks. Besides, hydrothermal caves
have been reported from silicate rocks (quarzite, scarn,
jasperoid, quartz veins), sulfate rocks (gypsum, anhy¬
drite), rock salt, and even from massive sulfide ores.
Hydrothermal Karst Related to Oxidation
of Sulfide Ores
A substantial amount of heat can be released when
infiltrating oxygen-rich waters react with sulfide ores.
This may lead to both increased temperatures and
enhanced carbonate aggressiveness of waters passing
through ore bodies. Those thermally and chemically
modified waters may then attack carbonate rocks to
produce a specific type of hydrothermal karst. This
type of hydrothermal karst is commonly triggered by
mining activities that may facilitate access of waters to
the ore bodies.
Dissolution in Subaqueous and
Subaerial Settings
Many large hydrothermal caves and cave systems
have formed in subaqueous conditions, that is, below
the water table. In addition, significant development
of caves can also occur in a subaerial setting, above
the open surfaces of underground thermal lakes by a
mechanism of condensation corrosion. Moisture evapo¬
rating from the lake surface moves upward and con¬
densed on cooler bedrock. The resulting water film
becomes aggressive by dissolving gaseous C0 2 ; the
condensate attacks the bedrock and creates cavities
with characteristic cupola-like morphology (Fig. 1).
This speleogenetic mechanism can be very efficient,
but requires a number of prerequisites, such as ele¬
vated temperature of water (promoting evaporation)
and sustained gradient of temperatures between the
ENCYCLOPEDIA OF CAVES
394
H
FIGURE I Solutional cupolas in an ascending channel presum¬
ably formed in a subaerial setting by condensation corrosion. Kraus
cave, Austria. Photo by L. Plan. Used with permission.
vapor phase and cave wall (promoting condensation).
Importantly, by this mechanism, cavities can develop
above the underground lake water in which it is
saturated with respect to bedrock and is not capable of
dissolving it.
MORPHOLOGY OF HYDROTHERMAL
CAVES
Hydrothermal karst produces a large variety of cave
morphologies with cave sizes ranging from solution
enlarged pores to extensive cave systems with total
mapped passage length exceeding 100 km. The most
common morphologic types of hydrothermal caves are
discussed below.
Solution Porosity
Solution enlarged pores are commonly observed
in borehole cores from depths up to 4—5 km in geother¬
mal areas and oil fields. They may form extensive layers
of rock in which solution voids account for as much as
5—15% of the entire rock volume. In these places such
horizons become parts of oil and gas reservoirs. Layers
of solution-enhanced transmissivity appear where the
movement of the fluids is slow, of the order of a few
millimeters per year or less, and where hydraulic struc¬
tures which could concentrate the flow (faults and
fissures) are absent. Such zones of enhanced porosity
may become the inception horizons for future caves.
Isometric Rooms
Roughly spherical pockets or rooms with diameters
ranging from 0.5 to 8.0 m were reported from Khod
Koniom Cave in the Crimea. The cavities, lined with
crystals of hydrothermal calcite (temperature of forma¬
tion 40—85° C) and filled with red clay are truncated
by a later vertical vadose cave at a depth of 80—200 m.
Similar caves occur in Kighizstan, where medieval
excavations in the Birksu mercury mine (Turkestan
Range) uncovered a series of near-spherical rooms
3—8 m in diameter. The rooms coalesce in two- and
three-dimensional clusters following the bedding
planes and minor faults. Originally, the rooms were
entirely filled by massive hydrothermal calcite contain¬
ing veinlets of cinnabar, but the ore was removed by
medieval miners.
Individual Chambers
Individual chambers are distinguished from iso¬
metric rooms discussed above by their larger dimen¬
sions, ranging from tens to hundreds of meters. Caves
of this type are composed of one or several large indi¬
vidual chambers. The latter commonly have a length
of 100—200 m, a width of 30—60 m, and a height of
80 m. Examples of caves belonging to this type are
Bakharden Cave in Turkmenistan (Kopet-Dag Range),
Novoafonskaya Cave in Abkhazia (Caucasus), Karani
Cave in the Crimea, Ukraine, and Champignons Cave
in Provence, France. Another spectacular example of
individual chambers are cenotes Caracol and La Pilita
(Mexico), which have nearly spherical shapes and di¬
ameters of 80 to 100 m and are filled with moderately
thermal water (29.6 and 31.6°C).
Single-Conduit Caves
As implied by the name, the single-conduit caves are
composed of a single long, typically tube-shaped pas¬
sage, which can be aligned horizontally, parallel to the
water table, or at a steep angle to it. Horizontal single
conduit caves are exemplified by the hydrothermal
Hellespont, Spence, and Kane Caves (Wyoming, U.S.A.),
which represent nearly horizontal, tube-shaped con¬
duits 60—600 m long. They have developed in a vadose
setting, where ascending H 2 S-bearing fluid comes in
contact with the air, resulting in replacement corrosion.
Thermal springs discharge through the three above-
mentioned caves, and several inactive caves having
similar morphology are known in the region. Steeply
dipping single-conduit caves develop in a phreatic set¬
ting. Examples of such caves are Grotte de Chat
(France), Pozzo del Merro (Italy; the deepest explored
underwater cave of the world), and cenote Zacaton
(Mexico).
ENCYCLOPEDIA OF CAVES
HYDROTHERMAL CAVES
395
FIGURE 2 Examples of the presumably mono¬
genic hydrothermal bush-like caves with cupo¬
las: Satorko-puszta Cave in the Pilis Mountains
and Batori Cave in Buda Hills, Hungary. Maps
by M. Juha'sz, P. Borke, and L. Krirpat.
Bush-Like Caves or Caves with Cupolas
Such caves typically consist of a basal chamber from
which a branching pattern of rising passages develops.
The branches are composed of coalesced spherical cupo¬
las whose typical size is 0.5—1.5 m, and cupolas termi¬
nate most of the branches. Such caves are known in
Hungary (e.g., Satorko-puszta and Batori caves; Fig. 2)
and in the Azerous Mountains in northern Algeria.
These types of caves are thought to be due to the deliv¬
ery of hot water to a single input point at the base of car¬
bonate rock having low fissure density. Bush-like caves
are an example of monogenetic hydrothermal karst,
although the exact mode of formation for such caves is
not yet understood. One hypothesis ascribes their forma¬
tion to the convective movement of water in the phreatic
zone, whereas another model (preferred) invokes the
convective movement of moist air above the hot-water
table coupled with condensation corrosion.
Phreatic Maze Caves
Maze caves are the most common types of hydro-
thermal cave systems. Among them network caves.
anastomotic caves, spongework caves, and ramiform
caves are distinguished.
Network caves are angular grids of intersecting
passages formed by widening of nearly all major
fractures within favorable areas of soluble rocks.
Two-dimensional rectilinear maze systems are created
where rising water is trapped in densely jointed
carbonate rock below a relatively impervious bed.
Examples of such a pattern are Cserszegtomaji-kut and
Acheron-kut caves in Hungary, developed in Triassic
dolomite under the cover of Miocene sandstone.
Multistory rectilinear network caves are more com¬
mon. Examples of these are caves of the Buda Hills,
Hungary (Pal-volgy, Szemlo-hegy, Ferenc-hegy) and
caves of the Black Hills in South Dakota (Wind and
Jewel caves). Many of the carbonate-hosted lead-zinc
ore deposits (the so-called Mississippi Valley type)
exhibit a network pattern of solutionally widened frac¬
tures later filled—partly or entirely—with sulfides
(Jefferson City Mine in Tennessee and Devil's Hole
Mine in the United Kingdom).
Anastomotic caves consist of curvilinear tubes that
intersect in a braided pattern. They usually form a
ENCYCLOPEDIA OF CAVES
396
H
two-dimensional array along a single favorable parting of
low-angle fracture. Less common three-dimensional var¬
iants follow more than one geologic structure. An example
of a supposedly hydrothermal three-dimensional anasto¬
motic cave is the Pobednaya Cave in Kirhgizstan, with its
1.5 km of very narrow, tube-shaped crawlways.
Spongework caves consist of interconnected cavities of
varied size in a seemingly random three-dimensional
pattern. Such caves appear to form by the coalescing
of intergranular pores and minor interstices. Ramiform
caves consist of irregular rooms and galleries wander¬
ing three-dimensionally with branches extending out¬
ward from the main areas of development. Passage
interconnections are common, producing a continuous
gradation with spongework and network caves. Josef-
hegy Cave in the Buda Hills, Hungary, is a good exam¬
ple of the ramiform pattern.
Small-Scale Morphology
Hydrothermal caves commonly exhibit characteris¬
tic morphology of cave elements (e.g., cave passages,
rooms, as well as yet smaller morphological features,
such as rock relief of cave walls). Among those
features Klimchouk (2007) has defined the morphologic
suite of rising flow. The suite comprises three major
components: (1) feeders (inlets); (2) transitional wall
and ceiling features; and (3) outlet features. Feeders
correspond to sites of input of rising water into hypo-
genic caves. Typical feeders are vertical or subvertical
conduits, either individual or forming small networks.
Transitional wall and ceiling features form with a
considerable role of buoyant effects (upward-focused
dissolution associated with rising limbs of free
convection cells). They include such small morphologi¬
cal forms as rising wall channels, ceiling channels
(half-tubes), and ceiling cupolas. Outlet features represent
channels of varying morphology, connecting the cave
to the next upper story or discharging water out of the
cave-development zone. Typically they are represented
by cupolas and domepits (vertical tubes) that rise from
the ceilings of cave passages and rooms. The suite
comprises morphological features created by rising
water, which may or may not be thermal.
Nevertheless, identification of these morphological
elements may be an important first step in determining
the hydrothermal origin of a cave.
CAVE DEPOSITS
Mineralogy
Calcite is the most common mineral of hydrother¬
mal caves developed in carbonate rocks. In addition.
such minerals as quartz, barite, fluorite, and sulfides
are commonly reported from the deep-seated hydro-
thermal caves. (Ore-related hydrothermal karst, where
the list of minerals can be quite large, is not considered
here.) Shallow hydrothermal karst caves rarely contain
any "exotic" minerals.
Character of Cave Deposits
Large euhedral (that is, having perfect crystallo¬
graphic shape) calcite crystals, aggregates, thick crusts,
and sediments reflecting stable hydrodynamic condi¬
tions are common in deep-seated hydrothermal caves.
Individual calcite crystals can be as large as 10—30 cm
and even 100 cm; gypsum crystals can be even larger,
reaching in some cases the length of several meters
(the most striking example is the 11-m-long gypsum
crystals in Naica cave, Mexico; these largest-ever-
found natural crystals were formed at 52° C).
The morphology of calcite crystals in deep-seated
hydrothermal caves is normally simple, dominated by
a scalenohedron (dogtooth spar) sometimes combined
with an obtuse rhombohedron. In contrast, the depos¬
its of shallow caves commonly reflect a more dynamic
environment. Crusts are typically less thick (except
in the immediate vicinity of the water table, where
massive mammillary crusts called cave clouds may
form) and euhedral crystals are rare. The size of crys¬
tals in aggregates ranges from several millimeters to a
few centimeters. The dominant crystal morphology is
a combination of a scalenohedron and a prism with
the crystal tip blunted by obtuse rhombohedron (nail-
head spar). Minerals might be contaminated by clay,
which indicates that the paleowaters were dynamic
enough to carry the particulate matter. In addition to
subaqueous deposits, two more types of speleothems
occur in shallow hydrothermal karst: waterline depos¬
its (rafts, folia, cave cones), and subaerial deposits (e.g.,
cave popcorn). These two types are also common in
cold karst.
HYDROTHERMAL CAVE LIFE
A peculiar cave life is known to exist in some active
hydrothermal caves, the most striking example being
Movile cave (Romania). The cave hosts slightly ther¬
mal (20.9°C) water rich in dissolved hydrogen sulfide.
The hydrogen sulfide is used as an energy source
by bacteria to fix inorganic carbon, thus producing a
food base for a highly evolved chemoautotrophic (i.e.,
based on chemosynthesis rather than photosynthesis)
ecosystem. As many as 48 species of cave-adapted
terrestrial and aquatic invertebrates, 33 of which are
ENCYCLOPEDIA OF CAVES
HYDROTHERMAL CAVES
397
endemic to this ecosystem, have been identified in
Movile. The cave did not have a natural entrance:
it was intercepted by an artificial shaft at 18 m below
the surface. It is thought that life in the cave has been
separated from the outside environment for the past
5.5 million years.
IDENTIFICATION OF
HYDROTHERMAL_CAVES
Unequivocal evidence of the hydrothermal origin of
a cave can be provided by (1) the presence of thermal
waters in a cave (obviously, this indicator is only rele¬
vant in presently active hydrothermal-karst caves);
(2) the presence of hydrothermal minerals deposited
on the cave walls (hydrothermal character of minerali¬
zation is inferred from mineralogical and/or geochem¬
ical evidence; it is also necessary to demonstrate that
the two speleogenetic stages, cave development and
mineralization, are related to the same process); and
(3) the presence of isotopic alteration of the cave walls
(character and degree of alteration must correspond to
the water/rock interaction at elevated temperature,
which is inferred on the basis of isotopic calculations).
Evidence suggesting the hydrothermal origin of a
cave includes characteristic morphological features,
indicating very slow water movement, free convection
of water, and active enlargement of caves in a subaer¬
ial setting (air convection and condensation corrosion
mechanism). Additional suggestive evidence includes
a lack of morphological features and deposits indicat¬
ing running water environment; a lack of association
with the surface features of "conventional" karst; and
spatial association with hydrothermal activity (includ¬
ing extinct one).
REGIONAL EXTENT
Hydrothermal origins have been established for
many large cave systems of the world: for example,
caves in the United States (South Dakota, Wyoming,
Montana), Austria, England, France, Italy, Poland, Czech
Republic, Slovak Republic, Hungary, Romania, Ukraine
(Crimea), Caucasus, Kyrgyzstan, Israel, Iran, Algeria,
Namibia, Mexico, Mauritania, and South Africa.
See Also the Following Article
Speleogenesis, Hypogenetic
Bibliography
Andreychouk, V., Dublyansky, Y., Ezhov, Y., & Lysenin, G. (2009).
Karst in the Earth's Crust: its distribution and principal types.
Sosnowiec (Poland) — Symferopol (Ukraine): University of Silesia —
Ukrainian Institute of Speleology and Karstology.
Audra, P., Hoblea, F., Bigot, J.-Y., & Nobecourt, J.-C. (2007) The role
of condensation-corrosion in thermal speleogenesis: Study of a
hypogenic sulfidic cave in Aix-Les-Bains, France. Acta Carsologica,
36/2,185-194.
Collignon, B. (1983). Speleogenese hydrothermale dans les Bibans
(Atlas Tellien—Nord de l'Algerie). Karstologia, 2, 45—54 (in French).
Dublyansky, Y. (1997). Hydrothermal cave minerals. In C. Hill &
P. Forti (Eds.), Cave minerals of the world (2nd ed., pp. 252—255).
Huntsville, AL: National Speleological Society.
Dublyansky, Y. (2000). Hydrothermal speleogenesis: Its settings
and peculiar features. In A. Klimchouk, A. Palmer, D. Ford &
W. Dreybrodt (Eds.), Speleogenesis: Evolution of karst aquifers
(pp. 292—297). Huntsville, AL: National Speleological Society.
Dzulynsky, S., & Sass-Gustkiewicz, M. (1985). Hydrothermal karst
phenomena as a factor in the formation of the Mississippi Valley-
type deposits. Handbook of strata-bound and stratiform ore deposits,
Vol. 13 (Pt. 4), 391-439.
Egemeier, S. J. (1981). Cavern development by thermal waters.
National Speleological Society Bulletin, 43, 31—51.
Ford, D. C., & Williams, P. W. (1989). Karst geomorphology and hydrology.
London: Unwin Hyman.
Ford, T. D. (1995). Some thoughts on hydrothermal caves. Cave and
Karst Science, 22, 107—118.
Klimchouk, A. (2007). Hypogene speleogenesis: Hydrogeological and
morphogenetic perspective (Special Paper 1). Carlsbad, NM: National
Cave and Karst Research Institute.
Palmer, A. N. (1991). Origin and morphology of limestone caves.
Geological Society of America Bulletin, 103, 1—21.
Sarbu, S. M., Kane, T. C., & Kinkle, B. K. (1996). A chemoautotrophi-
cally based cave ecosystem. Science, 272(5270), 1953—1955.
Takacz-Bolner, K., & Kraus, S. (1989). The results of research into
caves of thermal water origin. Karszt e's Barlang (Hungary) (Special
Issue), 31-38.
Van-Hise, C.R. (1904). A Treatise on Metamorphism. Monographs of
the United States Geological Survey, v. 47, Washington: Govt.
Print. Off.
ENCYCLOPEDIA OF CAVES
ICE IN CAVES
Aurel Per§oiu and Bogdan P. Onac
University of Suceava, Romania
University of South Florida
Ice in caves is a common occurrence in the Northern
Hemisphere during the winter when the temperature
drops to 0°C and below. These cold temperatures
extend inside the cave for a few hundred meters, caus¬
ing the drip water (more rarely flowing water) to
freeze and form seasonal ice stalactites, stalagmites,
and crusts. The life of these speleothems is short. The
speleothems melt during summer when both geother¬
mal heat and warmer outside airflow raise the cave
temperature above 0°C. This freezing/thawing cycle
rejuvenates every winter. Under favorable conditions
and particular cave settings, however, ice survives the
melting season, and accumulates into large masses of
perennial ice (Fig. 1). In contrast to glacier caves (i.e.,
caves within glaciers or at the contact between glaciers
and the rock beneath them), rock caves hosting peren¬
nial ice are known as ice caves. However, in some
places, caves with seasonal ice or even lacking ice but
having a colder climate than the external environment,
were inappropriately called ice caves.
Whereas seasonal ice forms a variety of speleothems,
perennial ice occurs mostly in the form of stratified ice
blocks (and to a minor extent, stalagmites and domes), in
many ways similar, but not identical, to mountain gla¬
ciers. Thus, we consider that the term glacier can be
extended to perennial ice accumulations in caves, pro¬
vided that a series of conditions are met: large, strati¬
fied ice blocks (a few meters in thickness), well-
marked seasonal mass balance fluctuations and active
flow. The last point is especially important, as flow of
cave ice has been rarely described. The cave walls
can restrict flow (e.g., Focul Viu Ice Cave, Romania;
Velika ledena jama v Paradani, Slovenia), or in some
cases, the flow of ice is merely an artifact of the ice
lying on an inclined surface (e.g., Eisriesenwelt,
Austria). In some cases, however (i.e., Scarigoara Ice
Cave, Romania), active flow is present, with the ice
moving under its own weight, generating glacier-
specific features: push-moraines, ice tongues, folds
(when obstacles are encountered), and so on. If
choosing to name an underground ice accumulation
glacier, one should make sure that the use of this
term does not bring in the confusion between ice
caves and glacier caves (Yonge, 2004).
Ice caves have been described from various climatic
zones, extending from subpolar (Canada, Norway, and
Siberia) to the Mediterranean, with a few occurrences in
the hot arid climates of the southwestern United States
and central Turkey. A peculiar combination of climatic,
hydrologic, and (cave) morphologic factors is required
for the genesis and preservation of ice in caves, in con¬
ditions that are otherwise not favorable for its forma¬
tion (i.e., too warm or too dry climates). A series of
mechanisms have been proposed for accumulation of
ice in caves, but often the mechanisms of cave cooling (a
prerequisite for ice genesis) and those leading to cave ice
formation were confused. Cooling of the cave can be
achieved by (1) trapping of cold air (Fig. 2A); (2) unidi¬
rectional ventilation (Fig. 2B); (3) evaporative cooling;
or (4) as a consequence of the cave's geographic setting
in the subpolar climatic zone. Most of the cave ice forms
by freezing of water, with sublimation of moisture and
snow diagenesis playing less important roles.
Caves at high latitude (e.g., Canada, Norway, and
Siberia) or high altitude (e.g., Altai Mountains) are
often under permafrost conditions. As a consequence,
the air temperature is below 0°C, which provides
favorable conditions for ice preservation in the cave.
Permafrost conditions, however, restrict water inflow
and therefore only a limited amount of ice (in summer)
can form by congelation. Perennial ice in caves, such
as those in Canada's Yukon Territory, is formed by
both congelation and desublimation processes. In sum¬
mer, moist air entering the cave desublimates to form
thick hoar ice deposits on the cave walls, which.
Encyclopedia of Caves.
399
© 2012 Elsevier Inc. All rights reserved.
400
i
depending on the cave's morphology and topoclimatic
conditions can either melt, with the resulting water
subsequently freezing to form congelation ice ( e.g., in
the Tsi-tche-han Cave, Canada) or accumulate on the
floor to form massive ice (e.g., Serendipity Cave,
Canada). Hoar deposits on cave walls can also be
found in caves outside the permafrost zone (e.g.,
Sairisoara Ice Cave and Dobsinska l'adova jaskyna.
FIGURE 1 The upper part of the ice block in the Great Hall of
Scari§oara Ice Cave and various ice speleothems (stalactites, domes).
Photo by Aurel Perqoiu.
Slovakia), but the deposition mechanism is slightly
different. The cold air entering these caves promotes
sublimation of ice and desublimation of the resulting
moisture on the undercooled walls and/or ice (Fig. 3A).
This hoar ice melts in early summer and the resulting
water drains away before the initiation of the freezing
in autumn.
Trapping and diagenesis of snow have been proposed as
mechanisms for ice formation in caves, but thus far, no
detailed descriptions and studies have been put forward.
Ice in Q5 Cave in Vancouver Island (Yonge, 2004) and in
Avenul din Albele (Retezat Mountains, Romania) might
have formed in this way. A number of caves throughout
Europe have snow accumulations near their entrances,
but the conversion to ice via snow metamorphosis and
freezing of meltwater inside the snow pack was not
reported. An explanation could be the reduced thickness
of the snow layer (up to 4—5 m), which prevents snow
diagenesis through weight compaction and also permits
through-flow of water before freezing.
Freezing of liquid water in an undercooled cave envi¬
ronment is the most common mechanism of cave ice
formation. As mentioned above, cooling can be
achieved through different mechanisms (points (1)—(3)),
but generally, it is a combination of them.
Trapping of cold air is the most common mechanism
of cave cooling, leading to cave ice formation. It is most
Winter
^outside < Tcave
Summer
Toutside > Tcave
Ml Ice block
I I Snow
l—»l Warm air flux
l~~^ I Cold air flux (A)
FIGURE 2 Schematic diagram showing the most
common cave cooling mechanisms. (A) Trapping of
cold air; (B) unidirectional ventilation.
ENCYCLOPEDIA OF CAVES
ICE IN CAVES
401
FIGURE 3 Ice speleothems. (A) Hexagonal hoar ice crystals ( photo by Aurel Per§oiu); (B) complex ice stalagmite morphology ( photo courtesy
of Iosif Viehmann ); (C) cryogenic calcite on ice surface ( photo fry Bogdan P. Onac); (D) overview of the 18-m-high ice wall in Little Reservation of
Scari§oara Ice Cave. Photos courtesy of Claudiu Szabo and Gheorglie Pra\ild. Used with permission.
common in caves with a single entrance (or multiple
entrances situated in the upper part of the cave) and
descending passages (Fig. 2A), in which cold air sinks
in winter, displacing the warm air, which is pushed out
through the same entrance (if the diameter is large
enough; e.g., Scarisoara and Bor tig ice caves, Romania)
or through other entrances (e.g., Focul Viu Ice Cave).
The direct cooling under the influence of sinking cold
air is accompanied by evaporative cooling, induced by
evaporation of moisture from the walls (Wigley and
Brown, 1976). In summer, the air inside the cave is
colder and thus heavier than the external air and is
trapped inside the cave, hence the name of the phenom¬
enon. However, "trapping" of cold air as the origin of
ice caves is a rather misleading term, as the presence of
the cold air inside the cave in summer is a consequence
of the presence of ice itself, rather than a cause for it.
Cooling begins in early autumn (when air temperature
outside the cave becomes lower than inside and cold
airflow propagates into the cave) and lasts until late
spring (the actual length depends on the general cli¬
matic conditions outside the cave, being longer in
colder climates and shorter in warmer ones). In sum¬
mer, however, the inflow of cold air ceases, and the
cave air temperature begins to rise, due to the heat con¬
duction through the walls and air column in the
entrance shaft(s) and the warmer percolating water.
Thus, the heat delivered to the cave leads to ice melting
and hence cooling of air, so that cave atmosphere is
maintained at 0°C and the cold air remains trapped
inside the cave.
Cooling by ventilation (chimney effect) is the third
most important mechanism of cave cooling below
freezing point. It occurs in caves with (at least) two
entrances situated at different elevations, as a conse¬
quence of temperature contrast between the cave and
the exterior. These differences result in pressure vari¬
ance at the lower entrance, which in turn triggers a
dynamic ventilation of the cave. If the temperature
inside the cave at the lower entrance is higher than
outside, cold air will be advected to the cave and
warm air will be pushed out through the upper
entrance (in winter); while if the cave air temperature
is below the outside one (in summer), warm air will
move off the cave through the lower entrance and cold
air will be advected through the upper one (Fig. 2B).
These mechanisms of air circulation through the cave
lead to cooling of the lower entrance and ice genesis
by water freezing in winter, while in summer melting
of ice will maintain negative air temperatures and thus
prevent inflow of warm air.
The first cold air wave inside the cave usually pro¬
motes evaporation of water from the walls and con¬
densation of moisture from air. Next, water freezes;
ENCYCLOPEDIA OF CAVES
402
i
this is a multiple-stage process, forming different types
of ice. Occasionally, liquid water (from the summer
melting season), pooling on top of older ice ( e.g., in
Scarigoara and Bortig ice caves) will freeze from top to
bottom to form lake ice. In a second stage, if tempera¬
ture is close to (but still below) CPC and water has time
to spread out, dripping water freezes on top of this
lake ice to form a thinner layer of floor ice. When tem¬
peratures drop below —2°C, dripping water will freeze
to form ice stalagmites and mounds. If the negative air
temperatures persist for longer periods, freezing condi¬
tions will occur in the ceiling of the caves, with subse¬
quent formation of ice stalactites and curtains. The
freezing of carbonate-rich water causes the expulsion
of CO 2 from the solution and subsequent precipitation
of cryogenic carbonates (Fig. 3B). These carbonates are
easy to diagnose as they show a marked shift toward
extremely high 8 13 C values (up to +18%o VPDB). A mix¬
ture of cryogenic carbonates and other organic and
inorganic impurities accumulates at the bottom of the
lake ice layer, forming, together with the ice, an annual
couplet, similar to varves in lakes. Most of the glaciers
(see above) in caves in Europe are formed by the accu¬
mulation of such couplets, although the annual lami¬
nation is usually partly destroyed by prolonged and
strong summer melting. Accumulation over longer
periods of time leads to relatively thick ice blocks (up to
24 m in Scari§oara Ice Cave), whose bases are close or
slightly below the pressure melting point of ice. As a
consequence, ice becomes plastic and steadily flows,
with rates varying between 0.5 and 30 mm per year.
Differentiated basal melting rates also contribute to
variations in the speed and direction of flow, as do
constrictions imposed by the surrounding cave walls.
Therefore, the ice can bend and fold or spill over steps
in the bedrock to form ice tongues. Moving of ice is
accompanied by rock polishing (e.g., in Scarisoara Ice
Cave a 3-m-high band of rock with different textures
can be seen above the present-day ice level, indicating
the former height of the ice block), moraine formation
(mixture of rocks, cryogenic calcite, and tree trunk
remains), frost shattering of the rock walls in the vicin¬
ity of the ice block, and patterned ground arrange¬
ments. A peculiar phenomenon associated with ice in
caves is the formation of cryogenic cave micropearls,
which form through precipitation of CaCC >3 in sum¬
mer, from carbonate-rich waters, and also in winter,
during the freezing of water and loss of CO 2 , as
described above.
Apart from the large ice blocks, the most interesting
cave ice formations are the ice stalagmites. They form in
the so-called periglacial cave environment, at a certain
distance from the larger ice blocks, where the cold air
halo centered on the ice block partly vanishes, leaving
the air temperature near the cave ceiling above 0°C
allowing the dripping water to enter the cave and sub¬
sequently freeze on the overcooled ground. Similar ice
stalagmites can also form in the coldest parts of caves,
before the freezing of the ceiling, as well as near the
entrances of unglaciated cave. These stalagmites
display a wide range of morphologies, from simple
candle-style to more complex ones as a result of air
temperature and variations in relative humidity (Fig. 3C).
A particular type of ice stalagmite, known as bamboo
or thermoindicator, consists of successions of perfectly
transparent ice knobs and narrow sections of opaque
ice (Fig. 4A). The narrow portions are formed at lower
temperatures when dripping water freezes instantly,
favoring the growth in height. The knobs are created
when higher temperatures at the apex of stalagmites
allow dripping water to flow down the stalagmite
before freezing, thus increasing their diameter. This
growth mechanism was confirmed by daily measure¬
ments and observations performed on stalagmites in
the Great Reservation of Scarisoara Ice Cave between
February 9 and March 23,1965 (Fig. 4B).
SCIENTIFIC SIGNIFICANCE
OF ICE IN CAVES
Short descriptions and papers related to ice caves
are scattered throughout various journals and books
published since the beginning of the nineteenth cen¬
tury. However, it was only over the past 50 years that
ice caves stimulated complex scientific investigations.
Most of the earlier studies targeted the ice in caves or
the caves per se (mostly climatic, morphologic, and gla-
ciologic studies). The more recent investigations are
aimed at an integrated assessment of the processes in
ice caves. In most cases, the ultimate goal is obtaining
information on past climatic changes from proxies in ice
such as pollen, water stable isotopes, stratigraphy,
chemical properties of ice and the impurities within,
and so on (Racovita and Onac, 2000; Yonge, 2004).
The organic matter preserved within the ice layers of
many caves is generally well-suited for radiocarbon age
dating, so that a very precise, absolute chronology can be
tied to the ice stratigraphy. With a few notable exceptions
(e.g., Eisriesenwelt, ~4000 years; Scarisoara, ~ 10,000
years (Per§oiu, 2011), the radiocarbon ages of all the other
ice blocks (cave glaciers) from Europe and North
America indicate they are younger than 1000 years.
Supported by stratigraphic data, the existing ages thus
far point toward a complex interaction between cave ice
and climate over this time period.
Most of the past climate studies were examining the
oxygen and hydrogen stable isotope variation in ice.
The isotopic composition of ice mirrors that of mete¬
oric water before freezing inside the cave. There is a
ENCYCLOPEDIA OF CAVES
ICE IN CAVES
403
strong correlation between outside temperature and
the oxygen isotopic composition in the ice precipita¬
tion. By combining the ages of the ice at different
depths with high-resolution stable isotope analyses
along the core, a detailed record of climate changes in
the cave's region (mean-annual temperature, source of
moisture, etc.) since ice began its accumulation can be
derived.
At Scari§oara Ice Cave, studies of pollen, micro- and
macrocharcoal, and macrofossil recovered from cave ice
provided valuable paleoecological data. The collected
data offered an accurate picture of past vegetation
dynamics and composition at both local and regional
scale, along with clear signals of human impact over the
past 1000 years.
Traces of industrial pollution were documented
from ice caves in Romania and Montenegro. This field
of study has great potential, although the great mobil¬
ity of the chemical species in ice during melting/
refreezing events may modify the original signal.
Studies in the fields of crystallography and mine¬
ralogy have revealed a wealth of data related to the
presence of ice in caves along with the description
of rare minerals such as monohydrocalcite or ikaite
associated with these environments. Furthermore, iso¬
topic and mineralogical investigations in ice caves
FIGURE 4 Bamboo ice stalagmites and their tem¬
perature record. Photo by Aurel Perqoiu; graph adapted
from Racovitd and Onac (2000).
80 cm
enabled the identification of cryogenic cave carbonates
(Fig. 3B).
A peculiar phenomenon associated with ice in caves
is the formation of cryogenic cave micropearls, which
form through the precipitation of CaCC >3 from carbon¬
ate-rich waters by either ordinary degassing of CO 2
when cave temperature is above 0°C or during freez¬
ing of water and loss of CO 2 , as described above.
Cave tourism has rapidly developed in the past sev¬
eral years. Tourism in ice caves has increased the pres¬
sure on the fragile cave environment and the ice. An
extreme example in this regard is the complete melting
of ice in Kungur Ice Cave (Russia) in the mid-1980s.
Since then, it has beautifully recovered due to ade¬
quate new management solutions. This led to the
growth of cave management—related research in ice
caves, mainly the study of microclimate and the inter¬
action between visitors and the cave environment.
Body heat from visitors and the output from lights
contribute heat to the cave environment, which become
greatly intensified when doors are added to a cave,
causing a change in the ventilation regime. All these
affect the melting of ice by sublimation or/and con¬
densation processes. Such climate studies allow long¬
term predictions on how the presence of visitors and
other physical modifications might impact the cave
ENCYCLOPEDIA OF CAVES
404
i
microclimate. Cave managers can use this information
to better maintain the appropriate environmental con¬
ditions in their ice caves.
On Earth, caves are hosted by almost every major
type of rock that has a soluble component. Caves are
natural traps for sediments, unique life environments,
and valuable repositories for various forms of paleocli-
matic and paleoenvironmental information since their
deposits are protected from destructive processes act¬
ing on the surface. Extraterrestrial caves on other pla¬
nets ( e.g ., lava tubes on the moon. Mars, etc.) may host
ice deposits, which in turn represent radically different
life conditions for some microorganisms (Boston,
2004). Understanding all aspects related to Earth's ice
accumulation and preservation in caves (even when
developing under deserts) and their particular micro¬
biological setting could ultimately have a paramount
importance in deciphering the history of celestial bod¬
ies. Investigating minerals, microorganisms, or other
materials trapped within ice in caves could give valu¬
able clues in the formation of these celestial bodies.
Furthermore, a wide diversity of cold-adapted bacteria
and fungi can live and grow while frozen in ice at tem¬
peratures approaching —30°C. Studying such microor¬
ganisms capable of metabolizing while under higher
pressure and very low temperatures may help our
understanding of ice caves as extreme life environ¬
ments and shed light on possible similar life forms that
could exist on Mars, Jupiter's moon Europa, or other
extraterrestrial planets (Boston, 2004).
Bibliography
Boston, P. J. (2004). Extraterrestrial caves. In J. Gunn (Ed.),
Encyclopedia of cave and karst science (pp. 355—358). London:
Fitzroy-Dearborn Publishers.
Pergoiu, A. (2011) Palaeoclimate significance of perennial ice accu¬
mulations in caves: an example from Scarigoara Ice Cave,
Romania. PhD thesis. University of South Florida, Tampa, USA.
Racovifa, G., & Onac, B. P. (2000). Scariqoara Glacier Cave.
Monographic Study. Cluj-Napoca, Romania: Editura Carpatica.
Wigley, T. M. L., & Brown, M. C. (1976). The physics of caves.
In D. T. Ford & C. H. C. Cullingford (Eds.), The science of speleol-
ogy (pp. 329—358). London: Academic Press.
Yonge, C. (2004). Ice in caves. In J. Gunn (Ed.), Encyclopedia of cave
and karst science (pp. 435—437). New York: Fitzroy-Dearborn
Publishers.
INVASION, ACTIVE VERSUS PASSIVE
D an L. Danielopol and Raymond Rouch
Austrian Academy of Sciences , ^Centre National de la
Recherche Scientifique France (Retired)
Why and how subterranean habitats were invaded
by surface-dwelling animals is one of the themes that
has fascinated generations of speleologists. The subter¬
ranean realm is colonized actively and/or passively
by aquatic and terrestrial fauna. Ecological evidence
and arguments for evolutionary scenarios that support
various models of colonization are provided in this
article. Modern evolutionary ecological research offers
new perspectives to better understand the invasibility
of subterranean systems by both terrestrial and aquatic
fauna.
INTRODUCTION
The topic of subterranean life is closely related
to the question of why and how organisms penetrate
and further settle, sometimes in large numbers, in an
apparently inimical environment, that is, where space
and energy can be very constraining for the develop¬
ment of flourishing organismic populations. Caves
accessible to humans were one of the first habitats to
be intensively explored over the past 200 years. This
led to the discovery of a diverse aquatic and terrestrial
troglobitic fauna, sometimes without direct relation¬
ship to the living surface dwelling animals. This
discovery stimulated naturalists to propose various
explanations for the subsurface habitat colonization by
surface-dwelling animals.
Students working on subterranean animals used
the term invasion to stress that the subsurface realm
is extensively and, in many cases, massively colonized
by surface-dwelling animals which later adapt to sub¬
terranean life conditions. The invasion modalities are
grouped in two major categories: (1) active coloniza¬
tion, which is mainly related to the environmental cues
as perceived by the surface-dwelling organisms, and
(2) passive colonization where animals arrive into
and settle within the subsurface environment either by
the force of or by the chance of the environmental
dynamics.
Scenarios explaining the active and / or passive inva¬
sion of subterranean animals were proposed during
the late nineteenth century as well as during a large
part of the twentieth century by naturalists interested
in the evolutionary processes that generated the
troglobitic and/or stygobitic fauna. At the beginning
of the twentieth century Neo-Lamarckians, such as
E.G. Racovitza and R. Jeannel in Europe and A.M. Banta
in North America, favored the view of an active inva¬
sion of the subterranean realm by animals that could
not survive at the Earth's surface either because of cli¬
matic changes or because of strong predator pressures.
These animals retreated into subterranean habitats
which offered them a kind of stable refugium against
epigean environmental constraints. A similar view is
shared by A. Vandel (1965) who is an exponent of
Encyclopedia of Caves.
© 2012 Elsevier Inc. All rights reserved.
INVASION, ACTIVE VERSUS PASSIVE
405
the organicist view of subterranean evolution, which
considers that old species, like senescent individuals
unable to live in an epigean dynamic environment,
retire into subterranean habitats. An origin of troglo-
bites due to passive invasion was proposed at the
end of the nineteenth century by E.R. Lankester who
considered that the troglobitic evolution could start
with the organisms that accidentally colonized cave
habitats. Most of these kinds of hypotheses were not
followed by rigorous scientific tests.
During the second part of the twentieth century the
authors assisted in an increased interest in the ecology
of subterranean animals. Biologists proposed various
scenarios explaining the origin of the diverse hypo-
gean fauna. As compared to the previous generation of
scenarios, these are more carefully documented and
offer better possibilities to test their plausibility. In this
article, a brief review of this information is subse¬
quently presented within its evolutionary and ecologi¬
cal context.
EVOLUTIONARY SCENARIOS
Models Based on Active Dispersal
The Climatic-Relict Model
This assumes that surface-dwelling animals actively col¬
onize subterranean habitats during periods when the envi¬
ronment changes drastically. Such events happened
during the Pleistocene period in the Northern Hemisphere.
Dry and/or cold climate at the soil surface, especially in
karstified areas, is considered an important climatic
stressor that constrained various animal groups to invade
underground systems. The fraction of organisms that
could not adapt to subterranean environments became
extinct while the new troglobionts or stygobionts built a
relictual fauna. Various terrestrial and aquatic hypogean
animals from Europe, North America, Africa, and
Australia are considered to fit this model.
The Adaptive-Shift Model
Preadapted surface-dwelling organisms are able to
colonize unconstrained subsurface habitats where
they progressively adapt to the new environment.
The invasion of terrestrial fauna and their subsequent
evolution in caves formed in volcanic regions such
as the Hawaiian Islands is an example proposed by
F.G. Howarth (1983). For marine anchihaline caves
a similar scenario called the zonation model was
proposed by T.M. Iliffe (1986).
The Active Colonization Model
This scenario, well documented by R. Rouch and
D. L. Danielopol (1987), considers the possibility of
colonization of subterranean environments by pre¬
adapted and/or generalist organisms. This is an active
process which does not occur during periods of envi¬
ronmental stress. The authors consider this model as
having a wide generality for the subterranean inva¬
sion. The process of invasibility following the scenario
of Rouch and Danielopol is decoupled from the specia-
tion process, while in the adaptive-shift model both
processes play an additive role.
Colonization through Marine Shallow-Water
Ecotones
The penetration route within the subterranean aquatic
caves follows either directly through the entrance of the
caves or through submarine karstic springs or porous
clastic sediments along coastal areas.
Transit of Marine Fauna through Epigean
Limnic Systems
This is an alternative for the colonization of caves
and/or interstitial habitats by marine benthic animals.
It occurs through the adaptation of euryhaline animals
in a first step to surface inland waters and further
by the invasion of subterranean waters. The latter sce¬
nario applies to crustaceans that colonized the Dinaric
Karst.
Colonization from the Deep Sea
Bathyal and/or abyssal fauna displaying biological
convergent traits with troglobitic fauna could colonize
shallow marine caves, especially on volcanic islands.
It was proposed that deep-sea animals—sponges, crus¬
taceans, or fishes—could, during that time, penetrate
through the crevices of volcanic or karstic rocks up to
the anchihaline shallow caves.
Escape from Epigean Predators and/or
Strong Competitive Pressure
The invasion of subterranean habitats by animals,
especially small invertebrates that tried to escape to
stronger predation and/or competition, is a scenario
envisaged by generations of biospeleologists. It is
hypothesized that many crustaceans invaded subterra¬
nean aquatic habitats in order to escape the predation
and/or competition pressure of other animal groups
such as insects. For instance, E.G. Racovitza (1907), at
the beginning of the twentieth century, suggested that
the troglobitic isopod Spelaeoniscus debrugei took refuge
in caves because of the incomplete capacity to protect
itself by rolling up into a ball when predators attack.
Because the isopod's antennae are left out during the
rolling up of the body (Fig. 1), it was assumed that
these crustaceans were easily preyed on by epigean
invertebrates.
ENCYCLOPEDIA OF CAVES
406
i
FIGURE 1 Spelaeoniscus debrugei, a troglobiont isopod from North
Africa. From E.G. Racovitza (1907). Used with permission.
Models Based on Passive Dispersal
Passage of Epigean Animals during
Massive Surface Water Infiltration into
Subterranean Systems
There is a whole series of possibilities for passive
migration of fauna into the subsurface. In alluvial plains
along river channels during floods, epigean animals
are transported by surface water which later infiltrates
the subsurface, taking animals with it. One example
is the periodic invasion of crustacean Cladocera from
the backwaters of the Danube wetlands (in Austria)
into the aquifer existing below the floodplain. Surface
streams that flow into karstic systems, for example,
through sinkholes or infiltrates through macro- and
microchannels, allow the invasion of subterranean karst
by a diverse fauna. The dispersal of this fauna can con¬
tinue underground into large karstic systems through
the complex drainage systems. It is hypothesized that
the highly diverse stygobitic crayfish fauna of Florida
originated through this process.
The Regression Model
Dutch biologist J. Stock developed a general model for
the invasion of marine benthic fauna into inland subterra¬
nean waters based on the eustatic regressive sea level
movements. Stock noticed that various crustacean groups
with stygobitic representatives, such as the amphipod
Ingolfiellidea, are presently known from sites located on
ancient marine paleocoasts. He assumed also that once
animals colonized subsurface coastal marine habitats
such animals become thalassostygobionts that minimally
dispersed into other geographic areas. During marine
regressive phases, the coastal thalassostygobionts, instead
of following the regression of the marine water, remained
in place and progressively adapted to inland nonmarine
subterranean water (brackish and later freshwater) which
replaced the marine water. Such animals become limnos-
tygobionts. In this way the invasion of subterranean
waters happens through a so-called passive regression
process and by vicariant events, that is, the inland origi¬
nally marine form is stranded and respectively isolated
from the marine population. Stock's model has since
been extended to more detailed scenarios, such as the
two-step evolutionary model of J. Notenboom, C. Boutin,
and N. Coineau (Fig. 2) and the three-step model of J.R.
Holsinger. They combine the active migration process of
preadapted animals with a second passive process dur¬
ing marine regressions, and they include a third phase
of ecological isolation in the new inland subterranean
habitats. It is assumed that the three-step evolutionary
model is able to explain also the invasion of both
marine and nonmarine benthic animals into the sub¬
surface environment. Despite the opinion that the
passive regression models have a wide generality,
they were never carefully tested for their basic eco¬
logical and biogeographical assumptions that the
animals cannot follow the marine water during the
regressive periods and do not further disperse inland
far away from the original site (see the next section
for additional arguments).
Rafting and Resettlement of the Fauna
in New Subterranean Habitats
Interstitial marine animals can be transported pas¬
sively on various objects with the whole sediment over
various distances. Harpacticoids and ostracods could
spread in this way.
Erosion of Sediment and the Drift of Subsurface
Fauna with the Surface Water Flow
Surficial interstitial marine and even freshwater
fauna are especially exposed during storms and floods
to the resuspension of the sediment and the passive
transport through the surface water column to other
ENCYCLOPEDIA OF CAVES
INVASION, ACTIVE VERSUS PASSIVE
407
FIGURE 2 Schematic presentation of the two-step evolutionary model, which explains the invasion of inland subterranean habitats by
marine fauna; g.w. = groundwater. Adapted from J. Notenboom (1991). Used with permission.
places. After the resettlement they are able to recolo¬
nize shallow subsurface habitats.
Unique versus Multiple Invasions
Various biospeleologists assumed that important
invasions of a subterranean habitat, such as a cave,
followed by the adaptation of the originally epigean
animals to the hypogean life and the expansion of their
geographic range, occurred during unique historical
events such as the strong climate deterioration on the
Northern Hemisphere during the Pleistocene era.
Today this view is challenged by the evidence of mul¬
tiple invasions of the subterranean environment by
epigean animals preadapted over a wide geographic
range. For instance, R. Rouch and D.L. Danielopol
(1987) consider this a basic aspect of their active colo¬
nization model. Multiple colonization occurs either
simultaneously or independent in time. The amphipod
Gammarus minus independently invaded various caves
of West Virginia at different times, taking advantage of
the suitable subterranean habitats.
Ward and Palmer coined the term interstitial highway to
describe the long-term dispersal routes below the soil sur¬
face. It applies especially to interstitial habitats along river
channels and alluvial plains. Hypogean dwelling animals,
such as the amphipods Bogidiella albertimagni and
Niphargopsis casparyi, known from Western to Eastern
Europe, expanded their geographic range along large river
systems such as the Rhone, the Rhine, and the Danube.
THE ECOLOGICAL EVIDENCE
Evolutionary and/or historical scenarios can be
checked for their plausibility by examining recent
field ecological situations and/or by laboratory experi¬
mental observations. Active migration of animals is
commonly observed for thigmotactic animals living on
the bottom of running waters from where they pene¬
trate the hyporheic interstitial systems without being
constrained by some environmental stress. There are
hundreds of species, insect larvae, crustaceans, water
mites, and various worms observed to invade shallow
subsurface habitats and adopt the hyporheal as their
ENCYCLOPEDIA OF CAVES
408
i
common life place. The colonization and further adap¬
tation to groundwater habitats by crustaceans, for
example, representatives of the harpacticoid genus
Elaphoidella, with wide ecological valence in tropical,
subtropical, and temperate zones, is in accordance
with the active colonization scenario of R. Rouch and
D.L. Danielopol (1987).
Laboratory experiments with the amphipod Gammarus
roeseli demonstrated that this species actively migrates
through the porous space of gravel sediments when
exposed to oxygen gradients; respectively, they avoid
hypoxic zones and look for normoxic water. New
volcanic islands that allow the development of sandy
beaches are rapidly colonized by marine epigean animals
that settle into the interstitial habitats, for instance, the
interstitial habitats of a sandy beach of Long Island near
New Guinea. This is an island that emerged through a
volcanic eruption and was colonized by two species
of crustacean harpacticoids within the first 20 months of
its existence.
Observation in the field and/or laboratory of the
penetration of epigean animals in subsurface habitats
because of predator pressure pointed to a more com¬
plex behavior than previously known. This behavior
was used in the colonization under the constraint
scenario mentioned above. For instance, observation of
Gammarus roeseli exposed to fish predation shows that a
fraction of the surface-dwelling animals migrate into
the sediment, while others reduce their movement
at the sediment surface. However, because animals are
very thigmotactic, they penetrate into the sediment any¬
way, provided there is the necessary vital porous space.
Passive invasion of karstic systems by diverse
epigean animals, both insect larvae and crustaceans,
was thoroughly studied during long-term filtration of
exurgencies by the ecologists working in France.
Studies done on the Baget, Moulis, and Dorvan karstic
drainage basins documented that a surprisingly high
number of animals penetrate during the rainy period
into the subsurface. Much of this fauna is further
released outside at the exurgencies following the water
that flows through the subsurface voids. In the experi¬
mental area Rhitrodat at the Biological Station in Lunz
am See, Austria, it was documented that the hyporheic
zone of a gravel stream can be recolonized within hours
by various meio- and macrofauna such as rotifers,
cyclopoids, and chironomids. A significant relationship
between the volume of the entrapment of sediment and
the number of animals caught pointed to the passive
invasion of this environment by various animals.
The view that aquatic animals take refuge in subsur¬
face habitats, mainly in the riverbed sediments, under
the pressure of climatic drought was checked in the field
and laboratory. Observation of two Northern California
streams, intermittent and perennial, did not show an
increase of the hyporheic fauna in either type of stream
during the dry period. The reaction is more complex in a
stream flowing through the Sonoran Desert in Arizona,
where during the dry period the shallow interstitial habi¬
tats accumulate epigean benthic invertebrates which
later disappear with the increasing sediment dryness.
Also, observation in the laboratory using oligochaetes in
sediment columns pointed out that the progressive dry¬
ness of the sediment determines a low fraction of the
worms to follow the water into the depth of the column.
Most of the animals remain entrapped in the unsatu¬
rated zone of the sediment. Marine interstitial fauna,
especially in tidal littoral zones, are able to migrate
tenths of centimeters during the seasons.
BRIDGING THE GAP—EVOLUTIONARY
ECOLOGY AT WORK
Evolutionary biologists, such as R. Lewontin, point
out that organisms are active subjects which are able
to choose (or select) their habitat; in this way organ¬
isms define their niche. Of course organisms are
also objects of the selective pressures imposed by the
environment that they experience. Hence, the fact that
organisms represent the subject and the object of
selection forms an important biological principle
which gives support to the active invasion scenarios.
The success of invasibility as the penetration and the
settlement of an organism in the explored environ¬
ment depends on the level of resources available for
its survival and reproduction. Organisms with low
energetic requirements and unspecialized for their
resource acquisition (the so-called generalists) have
an advantage over more specialized organisms or
those needing high energetic resources. At a premium
are also those surface organisms already predisposed
to colonize subsurface habitats. There are general
ecological principles that apply to the colonization of
the subterranean environment too.
Both population geneticists and ecologists dealing
with metapopulation studies of plants and insects
recognized that the success of invading new environ¬
ments depends on the selective capacity of organisms
for progressive adaptation during their dispersion paths.
The process, called adaptive infiltration, supports some of
the models that explain the active subterranean invasion,
for example, the zonation model of T.M. Iliffe (1986).
However, in environments that deteriorate rapidly,
animals can become extinct before significant evolution¬
ary changes have time to occur. This latter argument
weakens the credibility of scenarios based on the princi¬
ple of refugium under constraints, for example, the
climatic-relict hypothesis. The generality of invasion by
surface-dwelling organisms of subterranean habitats is a
ENCYCLOPEDIA OF CAVES
INVASION, ACTIVE VERSUS PASSIVE
409
well-documented phenomenon that cannot be explained
by local processes such as the predation and/or compe¬
tition pressure of epigean animals (remember the sce¬
nario of the refugium of animals underground under
the pressure of voracious surface predators).
The passive invasion scenario of J. Stock and its
modern variants of bi- and triphase evolutionary mod¬
els have a weakness; they do not explain why marine
animals that colonize marine interstitial habitats
should remain in place or migrate minimally during
the regression phase of the sea. As mentioned above,
marine interstitial animals display the capacity to
migrate vertically or horizontally depending on the
environmental conditions. Marine regressions proceed
at a low rate, theoretically allowing animals to track
the receding sea and to keep their marine ecological
requirements. Obviously we need more ecophysiologi-
cal information about the way marine littoral fauna
adapt to inland subterranean habitats.
CONCLUSIONS
1. The examination of alternative hypotheses which
could illuminate the problem of subterranean
invasion, as advocated by L. Botosaneanu and
J.R. Holsinger (1991), proved to be a successful
scientific strategy during the past 10—20 years.
It stimulated active research where the quality of
information for various models could be compared.
2. The scenarios regarding which animals repeatedly
colonized the subsurface realm during geologic time
could be tested with field and laboratory data.
3. The evolutionary models following which epigean
animals migrated into the subterranean realm under
external environmental pressures such as climatic
deterioration, predation, and competition pressure
seems to have limited generality, applying (if real)
more to local situations.
4. The passive invasion of marine animals as
described by the scenario that considers migration
in inland groundwater occurring during marine
regressive phases (widely favored by biospeleologists)
needs additional ecological research to keep it as a
robust explanatory model.
5. The scenario describing which deep-sea animals
were able to actively colonize anchihaline caves was
seldom corroborated by recent research and various
students who supported this hypothesis were
converted to an alternative explanation—the
migration of shallow marine organisms into
submarine caves.
Note: Additional information is available in:
Bodergat and Marmonier (1997); Botosaneanu and
Holsinger (1991); Culver and Holsinger (1994); Culver
et al. (1994); Danielopol et al. (1999); Gibert et al. (1994);
Racovita (1907); Uiblein et al. (1996); Wilkens et al.
( 2000 ).
Acknowledgments
Our research was financially supported over the years mainly by the F.
W.F. Austria and the Austrian Academy of Sciences (grants attributed
to D.L. Danielopol) and by the C.N.R.S., France (for R. Rouch). We thank
D.C. Culver (American University, Washington, D.C.) for critical review
of the manuscript.
See Also the Following Articles
Evolution of Lineages
Vicariance and Dispersalist Biogeography
Adaptive Shifts
Bibliography
Bodergat, A.-M., & Marmonier, P. (Eds.) (1997). Contraintes et insta¬
bility de l'envirornement: Strategies adaptatives des organismes
recents et fossiles. In: Geobios-Memoire Special (Vol. 21). Lyon
(in French): Editions University Claude-Bernard.
Botosaneanu, L., & Holsinger, J. R. (1991). Some aspects concerning
colonization of the subterranean realm—Especially of subterra¬
nean waters: A response to Rouch & Danielopol, 1987. Stygologia,
6, 11-39.
Culver, D. C., & Holsinger, J. R. (Eds.) (1994). Biogeography of subter¬
ranean crustaceans: The effects of different scales. Hydrobiologia,
287, 95-104.
Culver, D. C., Kane, T. C., & Fong, D. W. (1995). Adaptation and natu¬
ral selection in caves. The evolution of Gammarus minus. Cambridge,
MA: Harvard University Press.
Danielopol, D. L., Martens, K., & von Vaupel Klein, J. C. (Eds.)
(1999). Crustacean biodiversity in subterranean, ancient lake and
deep-sea habitats. Crustaceana, 72(8), 721—722.
Gibert, J., Danielopol, D. L., & Stanford, J. (Eds.). (1994). Groundwater
ecology. San Diego, CA: Academic Press.
Giere, O. (1993). Meiobenthology. The microscopic fauna in aquatic sedi¬
ments. Berlin: Springer-Verlag.
Howarth, F. G. (1983). Ecology of cave arthropods. Annual Review of
Entomology, 28, 365—389.
Iliffe, T. M. (1986). The zonation model for the evolution of aquatic
faunas in anchihaline caves. Stygologia, 2, 2—9.
Notenboom, J. (1991). Marine regressions and the evolution
of groundwater dwelling amphipods (Crustacea). Journal of
Biogeography, 18, 437—454.
Racovifa, E. G. (1907). Spelaeoniscus debrugei n.g., n.sp., isopode
terrestre cavernicole d'Algerie. Archivs Zooiogie Experimental et
Generate, 7, 69—77 (in French).
Rouch, R., & Danielopol, D. L. (1987). L'origine de la faune aquatique
souterraine. Entre le paradigme du refuge et le modele de la
colonisation active. Stygologia, 3, 345—372 (in French).
Uiblein, F., Ott, J., & Stachowitsch, M. (Eds.), (1996). Deep sea
and extreme shallow-water habitats: Affinities and adaptations
( Biosystematics and ecology series IT). Vienna: Austrian Academy of
Sciences Press.
Vandel, A. (1965). Biospeleology. The biology of cavernicolous animals.
Oxford, England: Pergamon Press.
Wilkens, H., Culver, D. C., & Humphreys, W. F. (Eds.). (2000).
Subterranean ecosystems Amsterdam: Elsevier.
ENCYCLOPEDIA OF CAVES
JEWEL CAVE, SOUTH DAKOTA
Mike Wiles
Jewel Cave National Monument
HISTORY
Early History
The first written record of Jewel Cave is a 1900 mining
claim, the Jewel Tunnel Load, filed by Frank and
Albert Michaud and Charles Bush. The "jewels" of
Jewel Cave are the calcite crystals (Fig. 1) that line
most of the cave's walls, ceilings, and floors. The
cave's entrance was originally a blowing hole that was
too small for human entry. It had to be enlarged to
facilitate mining and to provide easy entrance for tour¬
ists. Little mining was actually done, however, and the
cave failed as a major tourist attraction in the early
years because it was too far off the beaten track. In
1908 Jewel Cave became one of America's first national
monuments—the first one established for protection of
a cave—and in 1933 it became part of the National
Park Service.
Early Exploration
Until the late 1950s, little was known of the cave
and only about 3 km of passages had been discovered.
The situation changed dramatically when Dwight Deal,
a young college student who was writing a master's
thesis on the geology of Jewel Cave, invited Herb and
Jan Conn—climbing friends in their late thirties—to
accompany him in surveying and mapping the cave as
an aid to his research. In the process they discovered
many more miles of cave passages. They quickly real¬
ized that Jewel Cave was not the small cave it had
been assumed to be for so many years.
Later Exploration
After Deal (1962) completed his work and moved
on in pursuit of his career, the Conns continued
exploring the cave for over 20 years. Through the care¬
ful and systematic documentation of a survey, they
discovered over 100 km of cave passages before turn¬
ing the effort over to younger explorers (Conn and
Conn, 1981). They demonstrated that Jewel Cave was
not a small cave; rather, it was one of the world's lon¬
gest cave systems.
BAROMETRIC AIRFLOW
The Conns also recognized that the cave's pro¬
nounced airflow responded very closely to changes in
the outside air pressure. In 1965, Herb Conn measured
the airflow resulting from known barometric pressure
changes, and developed a mathematical model to pre¬
dict the total volume necessary to account for the
behavior of the airflow at the entrance (Conn, 1966).
Although no one can know for sure just how big the
cave may ultimately be, 100 million cubic meters is an
astonishingly large number and is strong evidence that
the vast majority of the cave system is yet to be found.
The wind has been known to blow barometrically at
speeds over 56 km per hour through crawlways that
are thousands of meters from the entrance. There are
strong breezes even at the farthest known extents of
the cave—an irresistible lure for any cave explorer.
PHYSICAL CHARACTERISTICS
Three-Dimensional Cave System
So far, more than 249.6 km of cave passages have
been discovered beneath about 8 km 2 of surface area.
Exploration has shown the cave to be a complex three-
dimensional maze with passages occurring in at least
Encyclopedia of Caves.
411
©2012 Published by Elsevier Inc.
412
J
FIGURE 1 Calcite crystals. The largest crystal is about 5 cm long.
five different levels (Fig. 2). However, the passages are
not distributed homogeneously throughout the subsur¬
face. The cave is typically a maze of large passages
where it is beneath a hill, but it bottles down to just a
few crawlways where it crosses beneath the canyons—
and it again becomes a maze of large passages beneath
the next hill. The entire cave appears to have formed
phreatically along joints and bedding planes, but devel¬
opment was most likely controlled by the same forces
that produced the modern-day landscape. In order for
the cave to avoid the canyons the way it does, it
must have formed concurrently with the downcutting
streams that created nearby Hell Canyon and Lithograph
Canyon.
Loft Level
The uppermost loft level typically has smooth,
domed ceilings and powdery weathered limestone
covering flat floors. Passage development appears to
have been controlled more by horizontal bedding
planes than by joints, and is less likely to follow the
parallel trends commonly exhibited by most of the
other cave levels. The calcite-crystal coating so abun¬
dant in most of the cave is generally absent, but there
are remnants suggesting it once coated the surfaces of
the loft level passages and was later redissolved. These
remnants often contain some of the largest calcite crys¬
tals found in the cave.
Chert Level
Chert level passages are found below the loft level, pri¬
marily in a 3- to 5-meter zone of alternating beds of chert
and limestone. These passages often have flat ceilings
and angular pieces of broken chert littering the floor.
Subchert Level
These passages are found immediately below the chert
zone and usually have rounded surfaces and domed ceil¬
ings. The bedrock usually consists of punky, often brecci-
ated limestone. A smooth, thin crystal coating typically
thins to nothing as it nears the chert level.
Main Level
Main level passages he beneath the chert level. They
are often large, strongly joint-oriented, and usually break¬
down modified. The ubiquitous deposits of manganese
minerals are often found beneath the breakdown collapse.
These manganese deposits are composed of a variety
of manganese oxides and hydroxides; most are mineral-
ogically unidentified and are simply referred to as
"manganese."
Lower Level
The lower level occurs in bedded dolomitic limestone,
and often exhibits some development along bedding
planes. It is common to find the crystal coating separated
from the underlying bedrock, presumably because it had
been deposited phreatically onto a surface that had previ¬
ously weathered to a punky consistency, and thus was
not as firmly attached to the walls and ceilings as else¬
where in the cave. Manganese deposits tend to be thick¬
est in the lower level. One deposit near the scenic tour
route is more than a meter thick.
Basement Level
Only a few hundred feet of passages have been dis¬
covered in the basement level. It is more joint con¬
trolled than the lower level and usually has boxwork,
but no crystal coating or manganese.
Breakdown Modification
Breakdown collapse is common throughout the cave
system, and many of the larger rooms have resulted
from the merging of two or more levels by collapse of
the intervening rock. What would have been flat floors
are now covered in piles of breakdown. This makes
moving through the cave a constant up-and-down
ENCYCLOPEDIA OF CAVES
JEWEL CAVE, SOUTH DAKOTA
413
Jewel Cave Levels
FIGURE 2 Cave levels.
Minnelusa Fm.
Madison Fm. ^ I
! Thunderhead
I Domes 30
Massive limestone
Weathered iron-oxide
breccia
Punky bedrock -rock
flour
Crystal usually absent
Interbedded chert and
limestone
10' thick chert zone
Crystal is sparse
Punky limestone
Rounded, domed
Crystal thin & smooth
Dolomite with
limestone and chert
Thick crystal lining
Dolomite
Joint-controlled
Thick crystal,
commonly peeled
away from bedrock
More joint-controlled
than Lower
Narrower than Lower
Poor connectivity with
other levels
experience. Even without the breakdown, no one level
of cave passages is very continuous. Often the caver
must cross between large passages via small holes and
crawlways, and must constantly change levels to go
from one end of the cave to the other.
Passage Sizes
Jewel Cave passages average 3 m wide and 5 m
high, but range from 20 cm to 33 m in height, and
0.3—55 m in width. Most of the crawls are relatively
short, but one area, named the Miseries, consists of
nearly 550 m of hands-and-knees to belly-crawls. The
Volksmarch is about 800 m of uninterrupted walking.
GEOLOGY
Jewel Cave was formed in the Mississippian Pahasapa
Limestone. It is approximately 130 m thick in the Jewel
Cave area, and all known cave passages are found in the
upper 75 m of the unit. This relationship holds true even
ENCYCLOPEDIA OF CAVES
414
J
S & G Cave
o Blowholes
Exposed Madison Limestone
Minnelusa Formation
Southern Black Hills
Caves and Blowholes
Reed's Cave
□ Miles
10
Coyote Cave
Wind Cave
(^4
Onyx Cave
Relationship between caves and the Minnelusa cap.
FIGURE 3
at Wind Cave, where the limestone is only 107 m thick.
Nearly the entire cave is in limestone that is overlain by
the Pennsylvanian Minnelusa Formation.
Recent Geologic Mapping
Detailed geological mapping of the Jewel Cave
Quadrangle (Wiles, in Fagnan, 2009) has identified six
subunits within the Minnelusa: (1) 12 m of basal sand¬
stone with thickness-compensatory red siltstone; (2) 15 m
of thin-bedded limestone; (3) 37 m of sandstones; (4)
37 m of dolomite with interbedded sandstones; (5) 37 m
of medium-to-coarse-grained sandstones; and (6) over
30 m of brecciated sandstone layers. Unit thicknesses
were constant throughout the quadrangle, allowing the
mapping of subtle faults and folds. The folds express
themselves in extensive dip-slope topography and
exhibit a strong correlation between synclines and sur¬
face valleys. Cross-cutting relationships indicate that the
passages of Jewel Cave postdate all major folding and
faulting events.
Relationship between Caves and the
Minnelusa Cap
One of the most striking observations is the fact
that Jewel Cave exists almost exclusively where the
limestone is capped with the Minnelusa Formation.
This relationship holds throughout the southern Black
Hills to the point that, without exception, no cave over
60 m in length is known to exist within uncapped lime¬
stone (Fig. 3). Furthermore, there is no mappable paleo-
topographical relief within the quadrangle. Reported
examples of paleokarst topography have turned out to
be misinterpretations of previously unrecognized struc¬
tures, often because the Pahasapa Limestone is so easily
confused with the limestone subunits of the Minnelusa.
Cave Fill
Red fill material found within Jewel Cave has tradition¬
ally been called paleofill, based on the belief that a paleo¬
karst topography had developed on top of the Pahasapa
prior to Pennsylvanian time, and that subsequent deposi¬
tion by the Minnelusa sea filled the associated "paleocave"
passages; these were later intersected by the development
of cave passages as they exist today. However, the absence
of paleokarst topography casts doubt on the validity of this
scenario. While the fill material is certainly from the overly¬
ing Minnelusa, the physical evidence suggests a more
recent timeframe, when dissolution within the Pahasapa
reached up into the Minnelusa, and Pennsylvanian mate¬
rial collapsed into the actively forming passage. Some of
Jewel Cave's passages even contain blocks of sandstone
ENCYCLOPEDIA OF CAVES
JEWEL CAVE, SOUTH DAKOTA
415
that are identical to the basal Minnelusa sandstone exposed
in nearby road cuts. If this scenario is correct, it would be
more accurate to think of the "paleofill" as a "neofill,"
because it formed contemporaneously with the develop¬
ment of the cave—long after deposition and lithification of
the Minnelusa.
HYDROLOGY
No part of the cave is known to intercept the water
table, as is the case with the famous Lakes region of
nearby Wind Cave. Water arrives only in the form of
vadose dripping in areas that are near surface drai¬
nages where erosion has cut through the overlying
Minnelusa Formation to within 30 m of the upper sur¬
face of the Pahasapa Limestone (Wiles, 1992); the
lower subunits of the Minnelusa are quite permeable.
Most of the cave is capped with stratigraphically high¬
er subunits that restrict water flow. Areas of the cave
below these subunits show virtually no evidence of
dripping water, present or past. Only about 0.25% of
the cave shows signs of dripping, and only half of that
is presently active. Dripstone and occasional small
pools can be found in the hydrologically active areas,
but Jewel Cave shows no evidence of streams or rivers.
Apparently, a higher water table had at one time sup¬
ported surface flow, and then both the groundwater
level and the surface flow had decreased significantly
soon after the cave had formed. There was little oppor¬
tunity for mechanical enlargement due to vadose flow.
ORIGIN OF JEWEL CAVE
Because recent observations have demonstrated a
strong correlation between the cave passages and mod¬
ern geological features, a new model is needed to explain
how the cave could have formed as a result of the most
recent processes that shaped the present-day structure
and topography. More work is needed to address impor¬
tant features such as calcite spar, boxwork, and manga¬
nese deposits, but this hypothesis does provide a broad
framework to account for the documented correspon¬
dence with modern geologic and topographic features,
as well as provide a plausible source of recharge.
Following is a description of proposed core events. Its
purpose is not so much to say, "Here is the answer," but
to spur the imagination to find better answers for the
new, unexpectedly puzzling observations.
Working Hypothesis
In the southern Black Hills, caves longer than 60 m
are only found in limestone that is presently capped
with Minnelusa. This means that the Minnelusa must
have played an essential role in the development of
large caves, and that this development must have
occurred in a geologically recent time. The base of the
Minnelusa consists of a medium- to coarse-grained
sandstone that readily allows vadose infiltration of
rainwater into the cave today. In the past, this sand¬
stone could have functioned as a local aquifer blan¬
keting the Pahasapa Limestone, which itself has a low
primary permeability. This aquifer would provide
nonpoint-source recharge to the maze of fractures
within the limestone, but there would be little circula¬
tion until down-cutting streams of nearby canyons
intercepted the top of the aquifer, establishing a local
gradient and causing groundwater to move toward
the canyons and discharge into gaining streams.
Some of the water would circulate phreatically as
deep as 75 m below the top of the Pahasapa before its
dissolutional capacity was depleted and it discharged
into the canyons via fractures, rather than conduits.
With groundwater approaching the canyons from
both sides, there would be little cross-flow beneath
the canyons, and consequently little cave develop¬
ment would occur in those areas. Water moving verti¬
cally across the Pahasapa/Minnelusa interface would
enlarge fractures within the limestone, but the overly¬
ing Minnelusa would collapse and fill these conduits
nearly as fast as they formed, resulting in deposits of
"neofill." As the volume of the cave increased and
the climate became drier, discharge would exceed
recharge: the Minnelusa aquifer would dry up and the
water in the cave would drain away. This must have
happened relatively quickly, because there is virtually
no evidence of vadose activity in Jewel Cave.
BIOLOGY
Cave Life
Jewel Cave is largely devoid of life except near the
historic entrance. This is because of the relative lack of
moisture, or a mechanism to carry organic material
deep into the cave where it could serve as a food
source for subterranean creatures. Animals that feed
outdoors do use the entrance area for shelter, however.
These are mostly bats and packrats and the parasites
and other organisms that feed on these mammals and
their excretions.
Bats
Though they almost certainly did not inhabit the
cave prior to the enlargement of the long windy
ENCYCLOPEDIA OF CAVES
416
J
constriction that was the natural entrance of the cave,
several species of bats now make the cave their home.
Many Myotis use the cave as a day roost during the
spring, summer, and fall, and can be found as far as
300 m from the entrance. They are joined in the win¬
ter by several hundred male Corynorhinus townsendii
(Townsend's big-eared bats). Jewel Cave is presently
the world's largest known hibernaculum for this
species. Interestingly, the C. townsendii prefer to
hibernate closer to the entrance, in the path of airflow,
which varies widely in temperature and humidity
when the cave is inhaling. When the temperature
is coldest, they will hang in clusters of a hundred
or more. When the air is warmer, many individuals
will break out of the cluster and form smaller clusters
nearby. The C. townsendii are frequently active through¬
out the winter. On the other hand, the Myotis move very
little, and seem to prefer the warmer and more constant
temperatures of rooms deeper in the cave and away
from the airflow. They commonly roost singly or in
pairs, and are occasionally found in clusters of up to
a dozen individuals. From 2000 to 2009 the total
wintering population averaged over 1300—a modest
increase from the previous decade's average of around
1200 .
Microbiology
Microbes are known to inhabit caves, even when
there is little or no organic food source. A single pre¬
liminary study at Jewel Cave has identified such
microbial life, but more research is needed to build a
base of knowledge from this meager starting point.
SPELEOTHEMS
Some of the speleothems found in Jewel Cave
include dripstone (stalactites, stalagmites, flowstone,
draperies); helictites; gypsum formations (flowers, nee¬
dles, and beards); aragonite; calcite crystals, coatings,
and rafts; boxwork; helictite branches; pool fingers;
popcorn; popcorn stalagmites (many of which are hol¬
low); and hydromagnesite balloons.
Popcorn Stalagmites
The popcorn stalagmites are particularly curious.
They range in size from 5 cm high and 5 cm in diame¬
ter, to 5 m high and 1.2 m in diameter, and many
have holes down their center axis. Sometimes the
hole is off-center, forming a slot along the side of
the stalagmite. In some cases the hole is even deeper
FIGURE 4 Hydromagnesite balloon. The balloon is about 2.5 cm
high.
than the stalagmite is high. Though almost certainly
formed by the action of dripping water, none of these
formations is found in areas where water is dripping
today, nor do they occur where there is any sign
of dripstone from past vadose activity. Clearly, they
formed under significantly different circumstances
than are observed today, and the nature of their ori¬
gin remains a mystery.
Hydromagnesite Balloons
Hydromagnesite balloons (Fig. 4) were first dis¬
covered in Jewel Cave, and have since been found in
half a dozen other caves throughout the world.
These delicate bubbles of hydromagnesite typically
have a pearly white luster and come in a variety of
irregular shapes. They are no more than 4 cm across,
and the shell is made up of layers of hexagonal plates
with a total thickness of only 25 pm. The balloons
appear to have been inflated. Both biological and
chemical mechanisms have been proposed, but no
firm conclusions have yet been reached. Several
ENCYCLOPEDIA OF CAVES
JEWEL CAVE, SOUTH DAKOTA
417
hundred balloons occur in two widely separated loca¬
tions in the cave—both in passages that experience
significant airflow.
OTHER BLACK HILLS CAVES
Jewel Cave is one of over 200 caves known in the
Black Hills. Most are dry caves with joint-oriented
passages, and many have a calcite crystal coating—
though most often thinner and less extensive than
that found at Jewel. Thick manganese deposits are
only found in one other local cave. The majority of
these caves are less than 150 m long. Five of them are
1.5—10 km long, and only a few have a barometric
wind. One notable exception is Wind Cave, with a
prominent barometric wind and over 219.76 km of
surveyed passages. It has the same "feel" as most
Black Hills caves, but has extensive boxwork devel¬
opment, and little in the way of a calcite crystal
coating.
CONCLUSION
Jewel Cave is a unique cave system of immense pro¬
portions. Thus far, only a small part of it has been
found, and there is much to be learned about its extent
and speleogenesis. It has a great potential for future
discovery, particularly in the areas of exploration, geol¬
ogy, and biological science.
Bibliography
Conn, H. (1966). Barometric wind in Wind and Jewel Caves, South
Dakota. National Speleological Society Bulletin, 28(2), 55—69.
Conn, H., & Conn, J. (1981). The jewel cave adventure. St. Louis, MO:
Cave Books.
Deal, D. E. (1962). Geology of Jewel Cave National Monument, Custer
County, South Dakota, with special reference to cavern formation in the
Black Hills. University of Wyoming.
Fagnan, B. A. (2009). Geologic map of the Jewel Cave Quadrangle,
South Dakota. 7.5 Minute Series Geological Quadrangle Map 9.
South Dakota Geological Survey.
Wiles, M. E. (1992). Infiltration at Wind and Jewel Caves, Black Hills,
South Dakota. South Dakota School of Mines & Technology.
ENCYCLOPEDIA OF CAVES
KARREN, CAVE
Joyce Lundberg
Carleton University, Ottawa
The term karren indicates small-scale dissolutional fea¬
tures of rock surfaces of the order of millimeter to meter
scale with an upper size limit of perhaps 5 m. Cave kar¬
ren (sometimes called s peleogens) are usually minor
modifications to the main passage form and smaller
than the passage, although some may dominate the pas¬
sage form or even comprise the whole passage. While
some are inherent to the initial speleogenetic process,
most are subsequently superimposed on passage walls.
General discussions on cave karren can be found in the
following texts: Ford and Williams (2007); Klimchouk
et al., (2000); Palmer (2007); Slabe (1995); White (1988).
INTRODUCTION
Cave karren develop where soluble rocks interact
with solvent fluid (typically water, but it can be water
vapor or aerosol) and typically where that fluid is mov¬
ing. Flow may be in phreatic (pressure flow) or vadose
(free flow) or alternating (floodwater) conditions.
Aggressivity may be enhanced by organic activity or
local acidification. Karren usually develop on bare rock
but occasionally they can form under sediments or
guano. Classification based on genesis is difficult
because many similar forms are produced by dissimilar
processes, for example, circular wall/ceiling cavities
may be produced by phreatic eddies, or in mixing-
water caves, or in air-filled caves from trickling vadose
water, or where water vapor condenses on surfaces
(perhaps in conditions as diverse as deep inside thermal
caves or the entrance zones of tropical caves). Another
example is the notch: some notches are simply dis¬
solved out bedding planes while others mark water
level; some form above sediment level (where sediment
armors the rock against dissolution) while others form
below sediment level (where sediment pore water is
aggressive).
CONTROLS ON KARREN FORM
In most cases the shape of the karren feature relates to
the properties of the flow rather than properties of the
rock, for example, free-fall drips produce pits; film flow
produces longitudinal rills/flntes or sometimes just etched
surfaces; channel flow with turbulent eddies and rollers
produces dissolutional potholes, scallops (probably the most
ubiquitous cave karren feature), transverse flutes, anasto¬
moses, pendants, half-tubes, and floor/wall/ceiling pockets.
If flow is slow enough that turbulence does not dominate
flow, then density and convection currents may govern
the form, producing facets, notches, bevels, and convec¬
tion cupolas. Condensation of vapor on surfaces pro¬
duces corrosion cusps/hollows and, probably, hellholes.
In some cases the dissolutional features are modified
strongly by rock properties. Karren are best developed
(i.e., have the smoothest, most regular shape) in homo¬
geneous, fine-grained, pure rock. Rock heterogeneity
results in differential dissolution where features may
be preferentially etched out (Fig. 1) or the less soluble
material may stand proud of the cave wall {eg., impure
beds, chert nodules, veins, or fossils). Boxzvork is the
result of preferential dissolution of the more-soluble
rock between intersecting sets of less-soluble veins.
There is no standard term, other than etch karren, for
the forms produced from this type of differential disso¬
lution because there is no standard morphology; each
shape is unique to the rock properties that govern it.
The production of cave karren may be complicated
by local, in-cave biological activity and in some cases a
unique suite of forms is produced. An example is photo-
karren (Fig. 2): suites of small tubes, rods, and cones
etched into the rock surface in the dimly lit entrance
zones of some caves (especially tropical caves); the diag¬
nostic feature is that they are all oriented toward the
Encyclopedia of Caves.
419
© 2012 Elsevier Inc. All rights reserved.
420
K
FIGURE 1 The influence of rock properties on cave karren: a
complex stylolite is etched out of the wall, Clearwater Cave, Mulu,
Sarawak. Photo courtesy of Keith Christenson. Used with permission.
FIGURE 2 The influence of local biological activity: photokarren
near to the entrance of Clearwater Cave, Mulu, Sarawak, are angled
toward the incident (low) light.
light. The control mechanism is not simple, but it
appears that photosensitive biota colonize in patches,
stimulating very local rock dissolution such that the
surface immediately under the biota retreats at the
angle of incidence of the light. Organic activity
associated with guano produces its own suite of
features, few of which have yet been documented.
Finally, karren form may be affected by clastic
material carried in the water. Abrasional features
(scour marks) are at the edge of the karren category,
since they are not purely dissolutional. Similarly
many cave potholes are largely abrasional in origin.
DESCRIPTIONS OF FEATURES
Drip pits occur where water drips freely onto bare
rock ( e.g ., into cave enhances, or from small vadose
inlets), producing a suite of simple circular cavities, of
semicircular to parabolic cross section, typically 1—5 cm
in diameter and depth (Fig. 3A). Simple pits are common
in wet caves, but in dry caves they usually get filled
with sediment and modified into round-bottomed disso¬
lution basins. These can be of any size, but are typically
<1 m; they form in dry caves where a thin, patchy sedi¬
mentary or organic cover inhibits drainage, and aggres-
sivity is often enhanced by organic activity (Fig. 3B).
Longitudinal rills/flutes are suites of vertical grooves
that develop where a water film flows down a high
angle rock face (Fig. 4A), for example, on vertical shafts
from descending vadose water. They vary in size from
~1 cm (rills) up to ~30 cm (flutes) in width and per¬
haps many meters in length, often extending the whole
length of the shaft. Although the bigger flutes may be
somewhat interrupted or influenced by bedding planes
and joints, the basic form is not governed by rock prop¬
erties. Sometimes the flute wall may itself be made up
of smaller-scale rills. Decantation flutes form where
stored water drains over a rock face, for example, out
of bedding plane partings or sediment or organic mate¬
rial (Fig. 4B).
Dissolutional potholes are sharp-edged, rounded-inte¬
rior basins in stream beds (Fig. 5). The scale is typically
20—200 cm in diameter and depth. They usually have
a close-to-perfectly-round circumference and are often
deeper than they are wide. They develop in channels
of steep gradient where flow is strongly turbulent and
velocity is high. They may be entirely dissolutional or
modified by abrasion when clastic particles are intro¬
duced. Some become abandoned when a drainage
channel develops in the pothole wall; then a new pot¬
hole develops below, leading to a suite of potholes lin¬
ing channel walls and floor.
Scallops are spoon-shaped hollows (Fig. 6) that
develop under deep flowing water with subcritical tur¬
bulent flow (at velocities generally between 1 cm sec -1
and 3 m sec -1 ). They develop in suites, overlapping
and intersecting each other (Fig. 6B), often covering the
whole of the wetted surface. They are markedly asym¬
metrical, the steep side facing downstream (Fig. 6C).
ENCYCLOPEDIA OF CAVES
KARREN, CAVE
421
FIGURE 3 (A) Drip pits that develop in caves under free-falling drops are similar to subaerial rainpits (although usually larger because of
the larger drop size). (B) Small dissolution basins developed in association with a thin sediment cover in the floor of Stonehorse Cave, Mulu,
Sarawak.
FIGURE 4 (A) Flutes develop on steep walls from film flow. (B) Extraordinarily large decantation wall flutes carved by aggressive waters
draining from the high-level shelf of guano. Deer Cave, Mulu, Sarawak. Photo by Donald McFarlane. Used with permission.
Their wavelength decreases with increasing velocity (a
log-normal length distribution), ranging from ~ 1.5 cm
in flood flows of ~3msec -1 to ~ 100 cm in gentle
flows of ~ 1 cm sec 1 . Thus, dominant water velocity
and direction can be estimated from scallop wavelength.
The scallop shape is cut where flow has separated from
the rock surface, then become reattached (Fig. 6C), allow¬
ing aggressive water direct access to the rock face rather
than the more usual indirect diffusion through the satu¬
rated boundary layer. The initial eddy may be caused
by an initial inhomogeneity; one eddy then triggers
others, so that the process becomes self-propagating.
ENCYCLOPEDIA OF CAVES
422
K
FIGURE 5 Dissolutional potholes in fine-grained limestones,
Lagarta Cave, Ecuador.
(B) (C)
FIGURE 6 (A) Scallops in Cave of the Winds, Mulu, Sarawak
(photograph by Keith Christensen, used with permission). (B) Block dia¬
gram shows typical packing pattern of scallops. (C) Longitudinal
profile of scallop or flute showing separation of flow at 1, reattach¬
ment of flow at 2, and eddy at 3.
The frequency of detachment increases as velocity
increases, reducing wavelength. For any one section of
passage, the scallops of characteristic wavelength are
formed by the dominant discharge. In vadose passages
the upper level scallops, reached only by flood flows, are
smaller.
Transverse flutes are essentially scallops of infinite
width (normal scallops are typically 1 —20 cm wide but
in very stable flow conditions can reach ~ 2 m); they
have the same cross section as scallops, but are lat¬
erally more extensive. They develop the same way as
scallops, but require a laterally continuous separation
of flow, that is, a transverse, roller eddy. Thus, they
are rare.
Scour marks develop where clastic particles are car¬
ried at high velocities; the scallop form becomes elon¬
gated and polished. These are standard fluvial forms,
but relatively rare in caves since they need both very
high velocities (>3 m sec ') and silica sand.
Anastomoses, pendants, and half-tubes are commonly
associated with confined flow. Paragenesis, or anti-
gravitational erosion, occurs where sediments filling
the passage confine flow against the ceiling, thus expos¬
ing the ceiling to pressurized flow and enhanced disso¬
lution. Paragenetic features therefore develop upward.
The term anastamose refers to a reticulate pattern
(such as in the veins of a leaf); in caves it applies to a
pattern of interweaving channels (Fig. 7) cut into a fis¬
sure of low dip that is penetrated by water under pres¬
sure. The fissure may exist between two rock surfaces
(bedding plane or joint), or between sediment fill and
cave roof (i.e., paragenesis). Anastomosing channels
(typically a few centimeters up to a meter in width)
are usually exposed to view only when the lower con¬
fining bed or fill is removed. If the relict channels are
not truly anastomosing, then they may simply be
called drainage grooves.
Floodwater can sometimes be forced into a complex
three-dimensional, reticulate route through the rock,
with interconnecting pockets and tubes at various
scales; this creates a form of s pongework. The scale can
range from a few tens of centimeters up to tens of
meters. Spongework can also be produced from locally
acidified waters and in mixing water caves. In some
cases spongework is produced where phreatic water is
very slow moving: the very slow dissolution kinetics
close to saturation allow differential dissolution to
enhance subtle difference in rock properties.
Pendants are positive forms hanging from passage
ceilings, typically ~10—100 cm in dimension. If they
are remnant from removal of intervening rock through
eddy dissolution they will have sinusoidal cross sec¬
tions (like egg boxes). If they are remnant from anasto¬
moses or paragenesis the cross section resembles an
inverted flat-topped mesa (Fi g. 7A). Paragenetic pen¬
dants can be complex and very attractive (Fig. 8).
Positive forms can also be left on walls and floors, but,
since they do not hang down, are not termed pen¬
dants; for example, echinoliths are remnant forms that
protrude upward from the floor. These are often of
complex shape, multifaceted with concave surfaces.
ENCYCLOPEDIA OF CAVES
KARREN, CAVE
423
FIGURE 7 (A) block diagram of bedding plane anastomosing channels: when the lower block is removed the upper bedding planes' half¬
tubes with intervening remnant pendants are exposed. (B) Paragenetic anastomoses formed at junction of rock and former sedimentary fill in
Cave of the Winds, Mulu, Sarawak. Photo courtesy of Keith Christensen. Used with permission.
FIGURE 8 Paragenetic pendants exposed by removal of sedimen¬
tary fill, Lagang Cave, Mulu, Sarawak. Photo courtesy of Keith
Christensen. Used with permission.
A half-tube (Fig. 9) is a channel of semicircular cross
section usually cut upward into a ceiling but sometimes
laterally into walls, and usually meandering more sinu¬
ously than the channel it occupies. It is almost always
paragenetic, the location, especially of the wall chan¬
nels, explicable only by perching of the stream on a
former fill or confinement against the roof. The parage¬
netic ceiling tube will become more entrenched as sedi¬
ment continues to fill the passage, the meanders
migrating downstream (Fig. 9A).
In some simple phreatic passages nonmeandering,
ogive-shaped (gothic arch) ceiling channels follow the
guiding joint: these have been explained as the result of
mixing corrosion where water trickling through the joint
intersects the main phreatic passage. A third suggested
mode of origin is that CO 2 is trapped against the ceiling
during flooding, enhancing aggressiveness.
If ceiling tube entrenchment cannot keep pace with
sediment filling, then the blockage causes locally very
high hydraulic potentials, and water is forced to adopt
new routes above the main passage, exploiting any
available fracture. The resultant bypass tubes and cork¬
screws often follow very tortuous routes.
Solution pockets are negative forms usually in ceil¬
ings but also developed in walls and floors; they are
blind pockets of rounded cross section, often circular
to elliptical in plan, sometimes multicuspate, and
developed on a guiding fracture (Fig. 10A). They range
in size from ~10cm to many meters. If they are
markedly taller than wide, they are called chimneys
and sometimes avens. They may form through mixing
corrosion where a joint introduces foreign water into a
phreatic cave passage, or in vadose caves where water
seeping from a joint renews its aggressivity by absorb¬
ing CO 2 from the cave atmosphere. Narrow, tapered
solution pockets form in epiphreatic conditions where
aggressive floodwater is forced under pressure into
fractures. Where several of these form side by side the
residual rock may be sharply bladed.
Bell holes are also blind pockets developed in ceil¬
ings; however, they are never associated with guiding
joints or lithological control (Fig. 10B,C). The form has
been called "negative stalagmite"-shaped; they are par¬
abolic at top, flared at base, usually much deeper than
wide, with perfectly vertical long axes. Typically they
are ~ 1 m deep and 20—30 cm wide, and are usually
arranged in suites. They are most often reported from
the humid tropics. Theories of formation have ranged
from phreatic eddying to bat-urine corrosion, but, since
ENCYCLOPEDIA OF CAVES
424
FIGURE 9 (A) Paragenetic ceiling half tube develops by upward entrenchment, meandering more sinuously than the passage it occupies.
(B) Paragenetic ceiling half tube, Niah Caves, Miri, Sarawak (photograph by Keith Christensen, used with permission). The main passage, still
largely sediment-filled, is about 15 m wide.
(A) Solution
pocket:
fracture
control
(B) Bell hole:
no fracture
control
FIGURE 10 (A) A solution pocket has a guiding fracture and often lithological control. (B) A bell hole has no guiding fracture and no
apparent lithological control. (C) A suite of bell holes in the roof of Fruit Bat Cave, Mulu, Sarawak. The photograph was taken after exodus of
the bats roosting inside the bell holes.
many are currently active without phreatic conditions,
recent research suggests that they may form by conden¬
sation corrosion associated with bat roosting.
Another blind pocket feature is the cupola. These are
complex multicuspate alveolar pockets, often in hierar¬
chical suites (Fig. 11 A). They can range from small to
passage-sized scale (at which size they would not be
classified as karren). Cupolas are characteristic features
of thermal caves. Thermal waters rising upward
through rock are often very aggressive (from dissolved
CO 2 or H 2 S). Below water level, thermal convectional
cells may be established in quasi-static water and carve
out rounded, spherical cupolas. Alternatively, the
aggressive medium may be steam from the warm
waters condensing on cooler passage roofs, again cir¬
culating in a thermal convection cell and thus produc¬
ing cupola forms. Cupolas may form part of a complex
three-dimensional network similar to spongework, or
they may be part of a branching dendritic network.
Most cupolas have no guiding fractures.
ENCYCLOPEDIA OF CAVES
KARREN, SURFACE
425
(A) Cupolas (B) Condensation
corrosion hollows
FIGURE 11 (A) Cupolas. (B) Condensation corrosion hollows
and cusps, cutting equally through rock and speleothem (stippled).
Condensation from vapor also produces condensation
corrosion hollows and cusps. These forms can be quite
complex, but are typically large, shallow, scallop-like
hollows, with somewhat sharp cusps where the hol¬
lows intersect. These are usually cut smoothly through
all material regardless of geological structure and tex¬
ture (Fig. 1 IB), for example, a single hollow may be
cut through bedrock and speleothem without any
obvious impact of the change in properties.
Reprecipitation of calcite sometimes occurs downwind
of the dissolution, on the lee side of the hollows.
Sometimes condensation rills/flutes are produced.
Condensation corrosion has long been known from
H 2 S-rich vapor in hydrothermal caves. While conden¬
sation rates are very much greater in hydrothermal
caves, recent research suggests that condensation may
be important in many nonthermal caves, for example,
during summer afternoons in tropical cave entrances.
Condensation requires significant changes in relative
humidity, and corrosion requires dissolved gases such
as CO 2 and H 2 S. Condensation may be caused just by
climatic effects, or by establishment of bat colonies.
Corrosion is most obvious in entrance areas of caves,
in narrow passages where an increase in airflow veloc¬
ity causes a decrease in pressure and temperature, and
in thermal caves directly above steam vents.
The final examples of cave karren occur where
water (either meteoric or thermal) is almost static.
Where open, vadose standing water becomes aggres¬
sive by absorbing C0 2 from the atmosphere, a sharp
horizontal waterline notch develops, exactly at the water
level and exactly horizontally. The notch often cuts
straight across geological structure. The scale is typi¬
cally tens of centimeters, but, where water level is
stable over long periods, it may penetrate the rock face
quite deeply, for example, up to a meter in normal mete¬
oric water caves. Shallow notches often occur in suites
corresponding to changing water levels. Horizontal cor¬
rosion notches can also form where density gradients are
set up by the production of heavy ion pairs and solute
ions; these sink, setting up cellular flows that carry fresh
solute or H + to the water surface and to the rock wall.
In this case notches are formed that taper off steeply
below the water line. Maximum dissolution is focused
on the point of first contact but the cellular flow carries
some H + ions downward, resulting in declining dissolu¬
tion with depth (with a linear or exponential decay).
The form may then be an inclined face at about 45° (a
facet) that meets the flat ceiling of the notch to produce
an inverted-triangular cross section (the laughohle). The
best-studied examples are developed in gypsum, where
1—3-mm-thick density currents have been measured
against the rock surface. Continued retreat of the
inclined face under stable water conditions leaves a hor¬
izontal ceiling, termed a laugdecke or corrosion bevel.
These are relatively rare because conditions are not
often stable enough for long enough.
See Also the Following Article
Scallops
Bibliography
Ford, D. C., & Williams, P. (2007). Karst hydrogeology and geomor¬
phology (especially pp. 248—265). London: Unwin Hyman.
Klimchouk, A. B., Ford, D. C., Palmer, A. N., & Dreybrodt, A. (2000).
Speleogenesis evolution of karst aquifers (especially pp. 100—111,
407—426). Huntsville, AL: National Speleological Society.
Palmer, A. N. (2007). Cave geology (especially pp. 147—155). Dayton,
OH: Cave Books.
Slabe, T. (1995). Cave rocky relief and its speleological significance.
Ljubljana, Slovenia: Zalozba ZRC/ZRC Publishing.
White, W. B. (1988). Geomorphology and hydrology of karst terrains
(especially pp. 91—102). New York: Oxford University Press.
KARREN, SURFACE
Joyce Lundberg
Carleton University, Ottawa
INTRODUCTION
Small-scale dissolutional features of rock surfaces or kar¬
ren are usually of the order of millimeter to meter scale,
but some extreme examples extend to almost 30 m.
Dissolution of the surface of soluble rock produces vari¬
ous rills, channels, pits, and basins. Individually, these
are called karren, each type with its own name; collec¬
tively, a mass of dissolution features is called a
Encyclopedia of Caves.
© 2012 Elsevier Inc. All rights reserved.
426
K
FIGURE 1 The Pinnacles of Mulu, Sarawak; a karrenfield made
up of ~ 30-m-tall pinnacles separated by deep narrow grikes.
karrenfield. General discussions on surface karren can be
found in the following texts: Bogli (1981); Ford and
Lundberg (1987); Ford and Williams (2007); Gines
(2004); Gines et al., (2009); Lundberg (2012); Trudgill
(1985); White (1988). These are important hydrologically
because much of the recharge to karst areas occurs here.
The most widespread of karst features, karren can form
in a great variety of situations and present myriad
forms that have spawned a plethora of jargon in several
languages! Fresh rainwater is the most common agent
of erosion, but snow melt is important in alpine areas,
and coastal areas (lacustrine or marine) have their own
suite of processes and karren forms. Most karren are the
result of simple dissolution but many are mediated by
biological activity, sometimes apparent only at the
microscopic scale.
Most commonly the feature of interest is the negative
form—the shape that has been removed—such as the
various types of channels and basins; but some of the
most dramatic karren landscapes are made of the posi¬
tive forms—the emergent remnant features. Classic
among these are the pinnacle karsts of the world, such
as the Pinnacles of Mulu (Fig. 1), the Tsingy of
Madagascar, and the Stone Forest of China, mostly
impenetrable terrain of deep clefts and extremely sharp
vertical blades. Although karren are so widespread,
several important karst areas show little visible karren
on the surface ( e.g ., the Sinkhole Plain of Kentucky).
BASIC CONTROLS
Karren develop where a solvent comes into contact
with a soluble rock. The most common solvent is fresh
meteoric water. The most common rocks are carbo¬
nates (limestone, dolostone, marble), but karren are
very nicely displayed on evaporite rocks (gypsum,
halite).
FIGURE 2 Karren of San Salvador Island, Bahamas. The smooth,
rounded, white surfaces are only recently emergent from under the
soil cover. The longer-exposed surfaces, detectable by their covering
of gray epilithic cyanobacteria, are considerably sharpened.
Obviously, dissolution requires that the water be
aggressive but even the very gentle acidification caused
by absorption of atmospheric CO 2 in rainwater is suffi¬
cient for shallow karren to form on bare rock. Dissolution
is much faster where water has picked up additional soil
CO 2 or biogenic acids. Snow melt is more aggressive than
simple rainwater because of the higher solubility of CO 2
in cold water. Seawater is generally not aggressive to car¬
bonates but, in areas of restricted circulation with intense
biological activity, can become aggressive.
One of the more important controls on karren mor¬
phology is the hydrodynamic control (i.e., how the fluid
impinges on the rock surface, the amount of discharge,
and the frequency of wetting). Bare rock can have direct
raindrop impact, or, with greater discharges, sheet flow
or channel flow; additionally, bare rock can quickly dry
out as soon as rainfall has stopped. However, where
rock has a cover (of snow, of vegetation, or of sediment)
the fluid has to impinge on the surface by percolation, it
is generally spread out more evenly, and it generally
remains damp for much longer after cessation of rainfall.
Covered-rock forms are therefore distinctly smoother
than bare-rock forms (Fig. 2).
ENCYCLOPEDIA OF CAVES
KARREN, SURFACE
427
FIGURE 3 Dissolutional pavements in dolostones of Dodo
Mountain, Mackenzie Mountains, northwest Canada. Dissolution,
although slow in this permafrost region, is not overwhelmed by frost
action so the joints have been opened by about 7 cm since retreat of
Wisconsinon ice. Scale can be estimated from the 4—6-m-tall cliffs.
Biological activity is important indirectly in that it
governs the amount of biogenic CO 2 and/or organic
acids available. However, sometimes biological action
is more direct, in the form of rock boring. Borings can
be quite large ( e.g ., up to several centimeters for gastro¬
pods and echinoderms) to microscopic (e.g., many bor¬
ing fungi, algae, and sponges may penetrate the rock
surface by only a few millimeters).
The environmental setting is important since climate
dictates precipitation (amount, intensity, and physical
form—snow, dew), and temperature (which largely
controls biological activity). Karren of the wet tropics
are generally the largest, sharpest, and most dramatic,
while karren of the arctic are generally subdued.
However, if frost action is not dominant, karren can be
well displayed even in the arctic (e.g., the Mackenzie
Mountains, Canada, is a periglacial area, but shows
nice dissolutional pavements; Fig. 3).
Finally, lithological factors also control the devel¬
opment of karren. Rock purity obviously governs sol¬
ubility, but beyond this, lithological factors may
dominate the morphology such that karren forms are
not obvious. For example, they are not usually well
FIGURE 4 The impact of rock texture on small pinnacles in for¬
est, Sarawak, Borneo. (A) Massively bedded and homogeneous lime¬
stone in Mulu. (B) Massively bedded but highly stylolitic limestone
in Niah. Both regions have similar climate and vegetation.
displayed on highly fossiliferous rocks or very thinly
bedded rocks. Figure 4 shows the effect of rock tex¬
ture on the morphology of emergent subsoil pinnacles
from Sarawak, Borneo.
DESCRIPTION OF SOME
KARREN FEATURES
In this account non-English terms are included where
they are regularly used, even by English speakers.
Negative forms of circular plan are called pits if
they are deep and narrow, basins if they are round-
bottomed and shallow, and pans if they are flat-bottomed
and shallow (Fig. 5). There is no limitation on size but
most are < ~ 1 m in diameter and depth. Most have
some evidence of biological colonization of the sur¬
faces and most retain water for some time after
ENCYCLOPEDIA OF CAVES
428
K
FIGURE 5 (A) Flat-bottomed solution pan (or kamenitza) with ver¬
tical to slightly overhanging walls, in San Salvador Island, Bahamas.
The floor retains some dried-out cyanobacteria and/or algae that
rehydrate on wetting. The interpan remnant rock surface shows the
small pits and spikes that are ubiquitous on bare karst rock.
rainfall. They are best displayed on bare rock, the sub¬
soil surfaces generally simply showing rounded
depressions rather than distinct basins or pans.
Negative features of linear plan that are governed
strictly by hydrodynamic controls (i.e., water flow)
include rills, flutes, and channels. Most channels form
when discharge becomes focused into a stream.
However, rills ( rillenkarren , sometimes called solution
flutes) are the only type of channel that forms at the crest
of a block, from the top down (Fig. 6), and from direct
impact of raindrops. They give way downslope to a
smooth nonchanneled surface. They are parabolic in
cross section, just under 2 cm in width on limestone,
always packed together in parallel suites, and separated
by knife edges. Where rills develop to either side of a
ridged crest, they meet at the top in a herringbone
pattern.
Drainage channels from focused flow under gravity
control, solution runnels, are sharp-edged and narrow on
bare rock ( rinnenkarren ) and smoother and wider under
soil cover ( nmdkarren ). They develop toward the foot of
slopes and extend upslope over time. These are the rock
equivalent of a normal fluvial gully, and so, depending
FIGURE 6 Rillenkarren on salt; Cardona, Spain. These are deeper
and wider than most rillenkarren on limestone.
on slope, may have a dendritic plan like a river system,
and will get deeper downslope as they collect more dis¬
charge. On bare rock they often meander. They are typi¬
cally several decimeters wide and deep. The smooth
runnels that form by focused drainage under soil are
exposed to view by soil erosion (Fig. 7).
Another type of channel is the decantation runnel or
decantation flute. If water is stored and its aggressivity
is enhanced (e.g., by biological activity) then the over¬
flow will form a runnel that is deepest at the point of
emergence and peters out downslope. The storage may
be in bedding planes, or clumps of soil or vegetation.
If the drainage emerges as a sheet then a suite of
decantation flutes will form.
Although water is required for karren formation,
many features show the influence of rock properties
more than flow properties. A very common type of kar¬
ren is the dissolutionally opened joint. The first stage of
opening is the splitkarren (like a small tear). The more
fully opened joint is called a grike (Fig. 8). Suites of
grikes develop along intersecting sets of joints leaving
the remnant emergent, flat-topped dint, or more
pointed, round-topped pinnacle, whose shape and size
ENCYCLOPEDIA OF CAVES
KARREN, SURFACE
429
FIGURE 7 The rounded shape of this rundkarren from
Vancouver Island, Canada, is indicative of its subsoil origin. Since
soil erosion, the bare surface has become etched by tiny meandering
channels (rillenstein).
FIGURE 8 Clints and grikes of northwest Burren, County Clare,
Ireland, making up a dissolutional pavement. The flat-topped dints
are remnant from the glacially scoured bedding surface. The grikes
are partly full of soil and vegetation.
vary with joint angles and frequency. The joints
exposed since ice retreat in glacially scoured limestone
landscapes are often now some 10—30 cm wide and up
to a few meters deep. However, joints that have been
opening without such a time limitation may be many
meters wide and deep ( e.g ., the grikes that separate the
FIGURE 9 Emergent rounded subsoil pinnacles in Wee Jasper,
Australia, becoming modified by rillenkarren.
huge pinnacles in Fig. 1). The larger ones are called
bogaz or corridors and make up labyrinth karst systems.
Dissolution under a cover of soil, vegetation, or even
sometimes snow, can be many times faster than on bare
rock, and hydrodynamic controls are quite different.
Capillary forces prolong contact between solvent and
solute. Where hydraulic gradient is low, this allows
exploitation of subtle differences in rock properties
resulting in cavernous karren, with curved intersecting
holes rather like a sponge. Subcutaneous karren ( bodenkar-
ren) are always smoothly rounded. Steeply sloping rund¬
karren develop where hydraulic gradient is high. Under
more acidic conditions features become more exagger¬
ated. A cutter is a deep grike that has formed under an
acidic soil cover, often with a complex curved cross sec¬
tion, some cavernous karren, and remnant rounded pin¬
nacles. Where rounded subsoil pinnacles are exposed by
soil erosion they become sharpened by development of
rillenkarren and runnels (Fig. 9). Small sharp pinnacles
are called s pitzkarren, examples of which are well
displayed in coastal karren (see below).
The junction between standing water or soil and rock
is often marked by the development of a corrosion notch,
whose form is not related to lithological controls. This
may simply be a shallow horizontal indentation but
ENCYCLOPEDIA OF CAVES
430
K
FIGURE 10 Cliff-foot swamp notch developed at Niah Caves,
Sarawak, Borneo.
swamp notches are often quite narrow vertically while
extending into the rock by several meters (Fig. 10).
KARREN ASSEMBLAGES
Sometimes suites of karren are developed over
large areas, creating landscape-scale karren assem¬
blages or karrenfields. Examples are alpine karren-
fields of bare rock above the treeline, showing
features associated with snow melt (such as heelprint
karren— trittkarren).
Pinnacle karst (Fig. 1) is another example of a karren-
field, made up of pinnacles separated by deep, narrow
grikes. The sharpest and tallest pinnacle karst develops
only on the purest and most mechanically strong
rock. Most are inherited from subsoil pinnacles after
regional uplift or base-level lowering.
Dissolutional pavements (Figs. 3, 8) are developed
on the bedding planes of flat or gently dipping strata
where joints have been opened up by dissolution.
Typically the bedding plane was exposed by glacial
scour (but any lateral planation process will do), so
pavements are characteristic of the formerly glaciated
regions. A pavement often degrades into rubble or veg¬
etation patches. However, if the tops of the dints remain
largely intact while the grikes enlarge, the end product
is a ruiniform karrenfield of tall, narrow emergent dints.
A unique suite of karren forms develops in marine
coastal regions where direct bioerosion is common.
Coastal karren can be divided into three basic land-
forms: the intertidal platform; the intertidal notch (plus
associated overhanging visor); and the ramp of basins
and pinnacles. The first two are well displayed in trop¬
ical regions. The third is typical of tops of the visors,
the more exposed parts of tropical coasts, and most of
the temperate coasts (Fig. 11).
FIGURE 11 Coastal karren of Curacao, Netherlands Antilles. The
basins and spitzkarren of the supra-littoral can be seen in the fore¬
ground, and the intertidal notch at the breaking wave in the
distance.
Bibliography
Bogli, A. (1981). Solution of limestone and karren formation. In
M. M. Sweeting (Ed.), Karst geomorphology (pp. 64—89),
Benchmark Papers in Geology 59. Stroudsburg, PA: Hutchinson
Ross Publishing Company.
Ford, D. C., & Lundberg, J. (1987). A review of dissolutional rills in
limestone and other soluble rocks. Catena Supplement, 8, 119—140.
Ford, D. C., & Williams, P. (2007). Karst hydrogeology and geomorphol¬
ogy. London: Unwin Hyman.
Gines, A. (2004). Karren. In J. Gunn (Ed.), Encyclopedia of caves and
karst science (pp. 64—89). New York, London: Fitzroy Dearborn.
Gines, A., Knez, M., Slabe, T., & Dreybrodt, W. (Eds.), (2009). Karst
rock features—Karren sculpturing (Carsologica, 9). Ljubljana,
Slovenia: Zalozba ZRC/ZRC Publishing.
Lundberg, J. (2012). Micro-sculpturing of solutional rocky landforms.
In Frumkin A. (Ed.), Karst geomorphology, 6, Treatise on geomor¬
phology, Elsevier.
Trudgill, S. T. (1985). Limestone geomorphology. London, New York:
Longman.
White, W. B. (1988). Geomorphology and hydrology of karst terrains.
New York: Oxford University Press.
KARST
William K. Jones and William B. White'
Karst Waters Institute,
^T he Pennsylvania State University
INTRODUCTION
The wonderfully diverse landscapes that make up the
surface of planet Earth are the result of a complex
interplay between tectonic forces, erosive forces, and
the mineralogical composition of the underlying
Encyclopedia of Caves.
© 2012 Elsevier Inc. All rights reserved.
KARST
431
bedrock. Tectonic forces elevate the land, thus provid¬
ing the gravitational gradients needed to drive the ero¬
sive forces. The erosional carving of the landscape is
accomplished by a combination of chemical reactions
and mechanical transport. Chemical weathering breaks
down the mineral grains and their cements, taking
portions into solution and leaving the insoluble por¬
tions as regolith or soil. Mechanical transport involves
gravity-driven downslope movement such as soil
creep and landslides and transport by flowing water in
storm runoff and flowing streams and sometimes by
wind scour and by creeping glaciers. In rocks such as
limestone, dolomite, and gypsum, chemical weathering
dominates over mechanical transport, resulting in a
unique set of landscapes known as karst. Because of
the dominant removal of rock masses in solution, karst
regions are those with internal drainage.
The distribution of karst landscapes over the Earth's
surface to a large extent follows the distribution of car¬
bonate (limestone and dolomite) and gypsum rocks
(Fig. 1). As a percentage of total land area, a compila¬
tion by Paul Williams gives values for carbonate rock
karst (Table 1).
The data in Table 1 are only karst areas on carbon¬
ate rocks. If areas of gypsum karst and a few rare
occurrences of karst on salt and other rocks are
included, the total comes into the 15—20% range
widely cited as the fraction of the Earth's land area
that is karst.
The objective of this article is to provide a broad
overview of karst. There is a large literature describing
individual karst regions. Regional descriptions include
Europe and Russia (Herek and Stringfield, 1972);
China (Sweeting, 1995); and North America (Palmer
and Palmer, 2009).
ORIGIN OF THE WORD KARST
Location
The word karst is a Germanized form of the name of
a carbonate plateau that is situated above the Adriatic
Sea immediately to the east of Trieste, Italy, and covers
about 750 km 2 (Fig. 2). The regional name for this area
is Kras from the pre-Indo-European root Ka(r) meaning
stone or rock (Gams, 2003). Variations on the name
include Carusadus, Karnsad, Grast, Krs, and Carso. Some
researchers argue that the "classical'' karst should be
considered just the Kras plateau, with about one third
of the area in Italy and two thirds in Slovenia. In as
much as the Kras is the type locality for karst there is
some justification for this rather narrow view. The idea
of a "type locality" is that the location gives its name
FIGURE 1 World map showing the distribution of karst regions. Map courtesy of Professor Paul W. Williams. See also Williams and Ford, 2006.
ENCYCLOPEDIA OF CAVES
432
K
TABLE 1 World Distribution of Karst
Percentage of
Land Area
Karst area in
million km 2
Russian Federation
16.1
3.3
South America
2.1
0.37
Africa
9.2
2.8
North America (including Central
America and the Caribbean)
18.3
4.1
Southeast Asia (including China,
Malaysia, and Indonesia)
10.8
1.7
Middle East and Central Asia
23.0
2.6
Europe (excluding Iceland and
Russia)
21.8
1.3
Australasia (including Pacific islands)
6.2
0.59
World total
12.5
16.7
FIGURE 2 Map showing the Kras Plateau, the railway line,
Skocjanske Cave, and the resurgence of water from the cave at
Timava (Timavo) Springs. This is the region that gives karst its name
and is sometimes referred to as the "karst of Trieste." Map from R.A.
Baroody after Petrie, 2005.
to the geologic formation or feature being described,
presumably for the first time in the scientific literature.
The type locality is not necessarily the best example of
a particular feature; it is just the location where that
type of feature was first described. Note that "Kras"
refers to a specific region and "karst" refers to a land-
form. To add to the possible confusion, in the Slovene
language Kras is the region and kras is the landform.
A more realistic definition of the classical karst is an
expanded area that extends eastward from Trieste at
least to Postojna and south to Rijeka (Croatia). This is
sometimes called the karst triangle and this area picks
FIGURE 3 Map showing the expanded view of the classical karst
that includes the caves and karst poljes in the vicinity of Postojna.
Map from R.A. Baroody after Gams, 2003.
up some important features such as karst poljes that
are not included in the landforms of the Kras plateau.
Geologists from Vienna who gave us the first modern
descriptions of karst landforms in the late 1800s proba¬
bly considered the karst to include the karst of the
Dinaric Mountains of Slovenia and much of the exten¬
sive karst landforms extending south through the for¬
mer Yugoslavia (Fig. 3). Kranjc (1994) suggests that all
of the Dinaric karst in present-day Slovenia be consid¬
ered the classical karst. Many of the technical names
for karst features also come from the Dinaric karst
region. Terms in international usage include doline,
polje, ponor, and hum. Of course many spectacular karst
landforms such as the tower karst of south Asia were
essentially unknown to Western geologists at the time,
so regardless of how we define the "classical karst" it
does not include all karst features and landscapes.
History
Slavic tribes followed the Roman occupation of the
Kras and were established in the area by 800 a.d. The
Romans used Trieste as a port and called it Tergeste.
The present territory of Slovenia and Trieste, Italy,
were part of the Austrian-Hungarian (Habsburg)
empire from 1382 to 1919. Trieste became the major
imperial port of the Habsburgs when it was declared a
"free port" in 1719. Trieste and part of Slovenia as far
east as Ljubljana were under Italian control from 1919
until 1945 when Yugoslavia emerged as an indepen¬
dent nation. The population of Trieste was almost
ENCYCLOPEDIA OF CAVES
KARST
433
evenly split between Italians and Slavs, but in 1954 the
UN handed Trieste back to Italy and it is today a
"border" city between Italy and Slovenia. Slovenia,
with most of the classical karst, became an indepen¬
dent country in 1991 and is presently in the European
Union. Many of the names and spellings have changed
from the Slavic roots to German to Italian to Slovene,
so it is sometimes confusing to read the earlier litera¬
ture and keep track of the changing place names. For
example, Postojna with its famous cave was known as
Adelsberg in German and Postumia in Italian.
When the Archduke Ferdinand Maximilian crossed
the Kras on a "highway" in 1850 he called the Kras a
"cursed desert" and was apparently very happy to
reach the Obelisk in Opicina above Trieste that marks
the end of the Kras plateau above the city (Morris,
2001). The Kras landscape appeared sparsely popu¬
lated, deforested, rocky, and waterless. The Kras
"threshold" was a region dreaded by many travelers
and traversed as quickly as possible. The Kras is now
easily crossed on modern highways and has been
extensively reforested, but it still strikes the traveler as
an alien landscape in many respects.
Karst phenomena were well known to the ancient
civilizations of the Mediterranean. Descriptions of
caves and karst features are found throughout Greek
and Roman mythology (Clendenon, 2009). Many of the
Greek settlements along the Adriatic coast appear to be
preferentially located on areas characterized by lime¬
stone rocks and karst springs (Crouch, 1993). Ideas and
descriptions of the Dinaric karst had been written long
before the science of geomorphology came into being.
Posidonios of Apameia (Roman, 135—50 b.c.) studied
the Skocjan Caves and the Timava (Timavo) Springs
(Kranjc, 1994). In 1599 Father F. Imperato tried (unsuc¬
cessfully) to conduct a tracer test using "floaters" to
establish the connection between Skocjan Caves (Fig. 4)
and the springs at Timava some 35 km northwest of
the cave (Shaw, 1992). An early "scientific" description
of the Kras was a monograph by Janez Valvasor (1686)
that describes in detail the polje of Cerknica and its
periodic lake (Fig. 5) and Skocjanske Jame (cave).
Hacquet published several papers around 1780 that
describe karst landforms and hydrology in Slovenia
and viewed karst as a special landform and a geologi¬
cal and hydrological specialty (Kranjc, 1994). The term
karst had certainly appeared in papers before the mid-
1850s, but it did not become established in the interna¬
tional literature until the late 1800s.
Karst
Caves and karst had been studied in many locations
around the world for several thousand years, but the
"modern" science of geomorphology dates to about
1858 (Roglic, 1972). The term karst became established
in the international scientific literature as a result of an
outstanding type locality in the Kras and the timing of
events. Vienna became a center of geomorphological
studies, and the Kras plateau with its many caves and
dolines was part of the empire. Travel between the capi¬
tal of the Habsburg Empire in Vienna and its port on
the Adriatic was slow and tedious until the Austrian
Southern Railway line was opened in 1857. After that,
geomorphologists could easily visit Postojna and the
Kras Plateau. Some of the early descriptions of dolines
were from road cuts right along the railway tracks.
Professor Albrecht Penck from the Vienna School of
Physical Geography published the first textbook on
geomorphology in 1894, but it was his student Jovan
Cvijic who really established the term karst in the geo¬
logical literature when he published Das Karstphanomen
in 1893. Cvijic was from Serbia and he described karst
features throughout the Dinaric Mountains of
Yugoslavia with many examples from the area of
Postojna and Trieste. Many of the most influential geo¬
morphologists of the late 1800s and early 1900s visited
the classical karst and the name became firmly estab¬
lished, at least in Western geology circles.
Caves represent the subsurface aspect of a karst
landscape and some researchers focused primarily on
the subterranean features observed in caves rather than
on the surface landscape. Understandably, speleologists
sometimes formed different theories about the process
that created the unique landforms on the surface. The
French cave explorer E.A. Martel used a decidedly
"speleological" approach to his studies of karst areas
and he preferred the term Phenomenes du calcaire
(Martel, 1894), which he probably felt put less emphasis
on surface features. Examples of other terms for karst
include limestone geomorphology, limestone landscapes, and
carbonate aquifers. Nevertheless, the term karst (noun or
adjective) and its many forms such as karstic (adjective)
is firmly established in the international literature and is
found in the title of many books and articles.
KARST LANDFORMS
Karst regions are characterized by a unique set of
landforms as described below. Karst regions also differ
from other geomorphic regions by the presence of
subsurface landforms—caves—which are also a distin¬
guishing characteristic. In general, surface karst fea¬
tures are more pronounced in regions where the caves
have direct hydrologic connections to the land imme¬
diately overlying the caves and the surface and subsur¬
face features develop more or less simultaneously.
Some refer to the surface expression as exokarst and the
subsurface expression as endokarst.
ENCYCLOPEDIA OF CAVES
FIGURE 4 Entrance to Skocjanske Jame, Kras region, Slovenia. Photo by W.K. Jones.
Pavements and the Epikarst
Carbonate rocks react with rainfall and infiltrating
water either at the exposed rock surface or at the soil-
bedrock contact. Chemical attack on the rock tends to
be concentrated on structural weaknesses such as
joints and bedding plane partings. If the limestone is
very pure or if the regolith has been removed, the
exposed limestone surface becomes a pavement karst.
Dissolution along joints produces deep crevices known
as grikes (in the U.K.) or cutters (in the U.S.). The
exposed bedrock is often sculptured into complex
small solutional forms known as karren. A karst surface
covered with regolith may appear featureless from
above but the regolith/bedrock contact, which is
where the active dissolution takes place, is likely to be
highly irregular with deep regolith-filled crevices sepa¬
rated by high pinnacles. The regolith/bedrock inter¬
face is usually sharp with carbonate-free regolith on
one side and unweathered limestone on the other side.
The crevices and other solution features often disap¬
pear at depths of a few to a few tens of meters as the
infiltrating waters consume their acidity and the disso¬
lution process slows. The zone that includes the rego¬
lith and the solutionally sculptured bedrock is known
as the epikarst.
Closed Depressions (Dolines, Sinkholes)
Karst drains internally through fractures in the bed¬
rock. This internal runoff tends to concentrate along
more transmissive zones in the fractures or at fracture
ENCYCLOPEDIA OF CAVES
KARST
435
FIGURE 5 Cerknisko jezero, the intermittent lake in the Cerknisko Polje east of Postojna, was described in detail by Valvasor in 1689. Karst
poljes are considered a typical karst feature of the Dinaric karst but are not represented in the Kras plateau of Trieste. Photo by W.K. Jones.
intersections and as a result dissolves a depression in
the bedrock. Closed depressions, variously known as
sinkholes or dolines, are one of the most characteristic
features of karst landscapes. Bedrock dissolution is
only one process contributing to the formation of
closed depressions. Cave roof collapse can migrate
upward to the surface, producing a collapse sinkhole
(or doline). In karst regions with thick soils, piping of
soil into the subsurface can produce a cover-collapse
sinkhole with little modification of the bedrock.
Sinkholes are found in a very wide range of sizes
from meter-size depressions to multikilometer diam¬
eters and in depths from fractions of a meter to hun¬
dreds of meters. Regions where closed depressions are
the dominant landform are not typically dissected by
surface streams and are known as sinkhole plains or
doline karst. These landscapes tend to have a similar
appearance in many parts of the world (Fig. 6).
Poljes
The largest known closed depressions are the poljes
found in the Dinaric karst and identified in many parts
of the world. The typical polje (a Slavic word meaning
"field") has a flat floor which may represent a deep
alluvial fill or a planated bedrock surface. The polje
floors flood seasonally and are frequently the best agri¬
cultural land. Streams typically enter the poljes from
springs, flow across the floor as surface streams, and
sink on the opposite side in swallets known as ponors.
Cones and Towers
In nonkarstic landscapes, flowing streams are inte¬
grated from small creeks to large creeks to small rivers
(A)
FIGURE 6 (A) Doline karst in Slovenia. (B) Doline karst in West
Virginia, USA. Photos by W.K. Jones.
to large rivers and on to the sea. The result is a land¬
scape of mountains, hills, and valleys all with open
outlets for drainage. Internal drainage in karst sculp¬
tures the landscape into closed depressions with
ENCYCLOPEDIA OF CAVES
436
K
FIGURE 7 (A) Cone and tower karst in Puerto Rico. (B) Cone and
tower karst (fengcong) along the Lijiang, Guangxi Province, China.
Photos by W.B. White.
individual drainage outlets on the bottom and the
residual rock mass in-between. If the residual rock
mass is dominant, the landscape is viewed as a surface
pocked with closed depressions. If the closed depres¬
sions are dominant, the residual rock masses stand out
in relief as what is known as cone and tower karst
(Fig. 7).
Cone and tower karst takes many forms. There may
be closed-spaced, steep-sided hills separated by chaotic
gorges. There may be isolated towers separated by pla-
nated surfaces. There may be large closed depressions
with roughly pyramidal hills at the intersections; this
form is sometimes known as cockpit karst. All of these
possibilities can be incorporated into the concept of
polygonal karst (Williams, 1972) where the cones,
towers, and closed depressions are analyzed in terms
of the pattern of their drainage divides.
Caves
Although the chemistry of the dissolution process is
similar, the degree of landform development on the
surface and in the subsurface can be different. There
are karst regions with some development of surface
landforms but very few enterable caves. There are
large and extensive cave systems with little expression
of karst landforms on the surface above. The internal
drainage in karst regions follows localized conduits
from recharge points to discharge at karst springs.
Explored caves are fragments of conduits or fragments
of abandoned conduits that are sufficiently large for
human exploration and which happen to have a con¬
nection with the land surface. The population of
known caves is only a small sampling of the remaining
unknown and inaccessible part of the conduit system.
LANDSCAPES OF MIXED ORIGINS
The original description of karst from the Dinaric
mountains was of a region entirely underlain by car¬
bonate rocks. Such regions were termed holokarst,
meaning the region was entirely karst with little or no
surface drainage. The Dinaric mountains are excep¬
tional in that they contain thousands of meters of car¬
bonate rock. In many karst regions the stratigraphic
thickness of carbonate rocks is much less. Other pro¬
cesses will be operating, and the resulting landscape is
of mixed origin.
Fluviokarst
Fluviokarst is the name applied to many landscapes
where exposed karstic rocks make up part but not all of
the drainage basins. In fluviokarst there are stream
tributaries and headwater drainage basins on insoluble
rocks. These surface streams typically sink at karstic rock
contacts, leaving only dry channels or dry valleys until,
downstream, the underground drainage reappears from
large karst springs. The size of fluviokarst basins varies
from less than 1 km 2 to hundreds of km 2 . Details depend
on the thickness and lithology of the karstic rocks, on the
relief of the drainage basin and the placement of karstic
and nonkarstic rocks within the basin, and on the struc¬
tural characteristics of the basin.
Alpine Karst
The term alpine karst refers to landscapes developed
in high mountain regions such as the Alps and the
Pyrenees in Europe or the Rocky Mountains in the
United States and Canada. Alpine karst is the result of
both the setting and the processes. The alpine setting
implies high relief, providing high hydraulic gradients
for the movement of water, and usually structural com¬
plexity in the rocks with extreme folding and faulting.
The dissolutional processes are also influenced by the
ENCYCLOPEDIA OF CAVES
KARST
437
extremes in climate, cold winters with thick snow pack
and hot summers often with intense storms.
In many alpine regions, soils are thin or nonexistent
with the only soil accumulations in the bottoms of
closed depressions. The exposed bedrock is sculptured
into a complex of karren forms. Deep shafts and deep
widened crevices develop along open fracture systems.
Caves usually have an extensive vertical component
with many pits and waterfalls.
Glaciokarst
The time scale for the evolution of most karst areas
is several million years. Those karst areas at high alti¬
tudes or moderate to high latitudes have, therefore,
been exposed to the comings and goings of the
Pleistocene glaciations. In some regions such as west¬
ern Ireland and northern England, the glaciations have
scraped off the soil, exposing dramatic pavement karst.
In other regions such as the American Midwest, the
carbonate rock surface has been covered with thick
layers of glacial till. Much of alpine karst also bears the
imprint of mountain glaciers.
COASTAL KARST, EOGENETIC KARST,
AND MIXING-ZONE KARST
Several factors set coastal karst and island karst
apart from karst developed in continental interiors. For
karst development directly on the coast, continuous
wetting by waves, tides, and storm surges makes the
karstic rocks a suitable habitat for a variety of organ¬
isms. Algal growths provide organic acids to supple¬
ment the effects of carbonic acid. Boring organisms
provide a mechanism for sculpturing the rock surface.
Near the coast is a zone where salt water mixes
with freshwater from the interior. The mixing-zone
waters become undersaturated even if both salt water
and freshwater are themselves at saturation. Mixing-
zone water dissolves out very characteristic caves
known as halocline caves or flank margin caves. Mixing-
zone karst is mainly a subsurface phenomenon except
for cave entrances exposed by receding coastlines.
A third factor, not limited to coastal areas but most
frequently found there, is the presence of highly
porous limestone. The carbonate rocks that support
most karst areas in the continental interiors have been
strongly compacted by deep burial and tectonic defor¬
mation. As a result, the primary permeability is
extremely low and groundwater movement is through
fractures and bedding plane partings. Caves develop
along these pathways and surface karst landforms are
strongly controlled by them. Collectively, landforms
developed in dense compacted rocks is known as telo-
genetic karst. In contrast, young limestones, typically
found in coastal regions, are still undergoing diagene¬
sis, have never suffered compaction or tectonic defor¬
mation, and as a result have a high primary
permeability. Karstification and diagenetic processes
are both at work, resulting in what is called eogenetic
karst.
HYPOGENETIC KARST
Traditional karst, both surface and subsurface,
forms by a top-down process with C02-containing
meteoric water infiltrating through the epikarst and
downward through the vadose and phreatic zones of
the aquifer to eventually reappear at karstic springs. In
recent decades, it has been realized that there are also
bottom-up processes by which karst features are
formed by up-welling deep waters. These features,
mostly caves, are known as hypogenetic karst.
Hydrothermal Karst
Some deep-seated solutions are also primarily car¬
bonate waters although often at high CO 2 pressures.
These solutions are also at temperatures well above
surface ambients and so the resulting dissolution fea¬
tures have been called hydrothermal karst.
Sulfuric Acid Karst
A completely different chemistry is responsible for
some large cave systems such as Carlsbad Caverns
and Lechuguilla Cave in New Mexico. The up-welling
solutions contain H 2 S which oxidizes to sulfuric acid
when the solutions reach the oxygen-bearing shallow
groundwater. Action of the sulfuric acid on the lime¬
stone in the reaction zone can produce very large cave
chambers although there may be little expression of
surface karst. Sulfuric acid caves often have character¬
istic three-dimensional patterns with large intercon¬
nected chambers. Residual masses of gypsum are also
an indicator.
See Also the Following Articles
Closed Depressions in Karst Areas
Karren, Surface
Soil Piping and Sinkhole Failures
Tiankeng
ENCYCLOPEDIA OF CAVES
438
K
Bibliography
Clendenon, C. (2009). Hydromythology. Lansing: Fineline Science
Press.
Crouch, D. P. (1993). Water management in ancient Greek cities. New
York: Oxford University Press.
Cvijic, J. (1893). Das Karstphanomen. Versuch einer morphologischen
Monographie. Geographische Abhandlungen, 5(3), 218—329.
Gams, I. (2003). Krasv Sloveniji. Ljubljana: Zalozba ZRC, ZRC SAZU.
Herek, M., & Stringfield, V. T. (1972). Karst: Important karst regions of
the Northern Hemisphere. Amsterdam: Elsevier.
Kranjc, A. (1994). About the name and the history of the region Kras.
Acta Carsologica, 23, 83—89.
Martel, E. A. (1894). Les abimes. Paris: Delagrave.
Morris, J. (2001). Trieste and the meaning of nowhere. New York: Simon
and Schuster.
Palmer, A. N., & Palmer, M. V. (2009). Caves and karst of the USA.
Huntsville, AL: National Speleological Society.
Petrie, M. (2005). Hydrogeological characteristics of Kras. In A.
Mihevc (Ed.), Kras: Water and life in a rocky landscape (p. 21).
Ljubljana: Zalozba ZRC.
Roglic, J. (1972). Historical review of morphologic concepts. In M.
Herak & V. T. Stringfield (Eds.), Karst: Important karst regions of
the Northern Hemisphere (pp. 1—18). Amsterdam: Elsevier.
Shaw, T. R. (1992). History of cave science. Broadway, Australia:
Sydney Speleological Society.
Sweeting, M. M. (1995). Karst in China. Berlin: Springer-Verlag.
Williams, P. W. (1972). Morphometric analysis of polygonal karst in
New Guinea. Geological Society of America Bulletin, 83, 761—796.
Williams, P. W., & Ford, D. C. (2006). Global distribution of carbon¬
ate rocks. Zeitschrift fur Geomorphologie, Supplement, 147, 1—2.
KAZUMURA CAVE, HAWAII
Kevin Allred
Hawaii Speleological Survey
INTRODUCTION
Among the many known lava tube caves presently
explored throughout the world, Kazumura Cave, on
the big island of Hawaii, stands foremost in length and
vertical extent. It is also the longest linear cave in the
world, extending farther over a straight-line distance
than any other. This cave contains a wide range of spe-
leothems and morphology with populations of cave-
adapted invertebrates. Because much of the cave is
located under rural subdivisions, it is a popular place
to visit for locals and out-of-state cavers.
PHYSICAL SETTING
Location and Surface Relationships
Kazumura Cave was formed, along with many
nearby caves, approximately 300—500 years BP in
a large tholeiitic lava flow complex known as the
Ai-laauShield Flow on the eastern side of the Island of
Hawaii (Holcomb, 1987). It is thought that the
Kazumura Cave flow was one of the last in the Ai-laau
flow complex; however, black glassy-skinned lava has
intruded into the cave in several places, showing that
later flows occurred after the cave had cooled signifi¬
cantly. The flow originated from the summit of
Kilauea Volcano at about 1189 m. The upper end of the
cave was collapsed and filled from highway construc¬
tion at the 1130-m elevation, but continuous passage
can be traversed downslope for 41.8 km to the lower
end at an elevation of 28 m (Fig. 1). It is, by far, the
longest linear cave in the world, covering a straight-
line distance of 32.2 km. The cave has an average slope
of less than 2°. It has a temperature range from 15°C
near the Kilauea caldera to 22°C at the lower end.
The surface slopes of the Ai-laau flows are presently
undergoing dramatic transformation from a barren lava
wasteland to lush tropical old growth jungle atop ever-
thickening soils. In many cave passages, delicate tree
roots penetrate the thin soils and then cracks of the roof
supply an energy source for cave-adapted crickets, fan-
tails (Fig. 2), and spiders (Howarth, 1973). Lava tube
slime, a little-known-about organic growth, sometimes
coats ceilings and walls with gold, white, or red.
Human Impacts
Because much of the cave is beneath developing rural
subdivisions, it is being negatively impacted. Sometimes
with a roof less than 3 m thick, it is prone to collapse from
bulldozing, and is sometimes used for domestic sewage,
graywater, and trash disposal. Recently, its increased
notoriety has led to some industrial tourism resulting in
many thousands of feet of passage modification. In these
areas, it no longer retains its virgin natural appeal.
FIGURE 1 The location of Kazumura Cave on the Island of Hawaii.
Encyclopedia of Caves.
© 2012 Elsevier Inc. All rights reserved.
KAZUMURA CAVE, HAWAII
439
HISTORY OF EXPLORATION Survey and Systematic Exploration
Local Visitation
Not much is known of the earliest exploration of
Kazumura Cave. There is some evidence of use by pre¬
historic Hawaiians. Many parts of the cave also contain
modern paraphernalia such as rotten rope, string,
trash, and bits of worn-out shoes. In one section, some¬
one used an innovative approach by placing pages of a
Reader's Digest magazine along the floor to mark his or
her way back to the nearest entrance.
FIGURE 2 A fantail clings to a tree root which has penetrated the
roof of the cave. Photo by Mike Shambaugh. Used with permission.
An entrance of Kazumura Cave was a designated civil
defense shelter in the early 1960s, and its name origi¬
nated from this. In August 1979, acting on reports of a
major lava tube, a U.K. speleological expedition explored
and surveyed 11.7 km of the cave with a vertical range
of 261 m, making Kazumura Cave one of the three lon¬
gest lava tube caves in the world. It was also visited and
partially surveyed by Japanese cavers. Finally, 14 years
later, members of the Hawaii Speleological Survey of the
National Speleological Society began a more determined
survey using profile views. The original British portion
was extended to 12.5 km and various plugs and col¬
lapses were dug through to eventually extend the cave
to 65.4 km with a vertical extent of 1101.8 m. The cave
presently has 101 known entrances (Fig. 3).
DESCRIPTION OF KAZUMURA CAVE
Speleogenesis
Kazumura Cave is a well-developed "master tube,"
formed from the crusting over of a highly liquid pahoe-
hoe lava flow issuing steadily for a long period from
Kilauea Volcano. Because of the insulative nature of the
lava tube, the Kazumura flow extended 39 km to the
Pacific Ocean and significantly extended the coastline
there. But the cave itself becomes low and pinches off in
pooled lava 3.5 km from the beach. Although some por¬
tions of the cave developed by the crusting over of the
main open lava stream, evidence indicates that
FIGURE 3 One of many entrances to Kazumura Cave. Photo by Mike Shambaugh. Used with permission.
ENCYCLOPEDIA OF CAVES
440
K
FIGURE 4 Over time, the flowing lava cut down the floor of Kazumura Cave, creating deep canyon-like passages. Here, near an entrance
(upper center) an intermediate roof formed, reinsulating the lava. Photo by Mike Shambaugh. Used with permission.
FIGURE 5 During the thermal erosion of Kazumura Cave, increased turbulence at the outside of bends has formed cutbanks (right).
Slipbanks result on the inside of bends (left) where there is less thermal erosion. Photo by Mike Shambaugh. Used with permission.
generally, the cave passage first formed as a braided,
interweaving delta-like pattern near the growing snout
of the flow. The cave roof was thickened through accre¬
tion against the cool ceilings. Overflow or breakout
further built up the roof. Kazumura Cave passages vary
widely in shape from wide and low typical of the
embryonic braided passages to deep and canyon-like,
which indicates significant downcutting (Fig. 4).
Thermal Erosion
As the flow extended, one or two braids pirated
the flow and downcutting occurred through a process
known as thermal erosion (Hulme, 1973). This is a
partial melting of the underlying substrate with an
accompanying mechanical plucking away of unmelted
crystals. Deepened canyon-like passages, lavafalls, and
ENCYCLOPEDIA OF CAVES
KAZUMURA CAVE, HAWAII
441
FIGURE 6 Early-formed braided tubes are in black. The developed master tube is in white. One or two braids pirated the Kazumura lava
flow and were deepened by thermal erosion (note cross sections).
exposed beds of old pahoehoe and aa flows exposed
behind ruptured wall linings are all evidence of this
thermal erosion. More turbulent flow eroding aggres¬
sively into the substrate on the outside of bends cre¬
ated cutbanks (Fig. 5). The early formed braided
passages that are not lava filled are commonly found
at the ceiling level off the main passage (Fig. 6). As
thermal erosion progressed, portions of the now
unsupported ceiling collapsed into the free-flowing
lava and this breakdown was carried away or wedged
into upper areas during temporary flooding.
Extraneous Tubes
The most outstanding proof of thermal erosion is
found where the downcutting lava stream encountered
older lava tubes or large bubbles. Here, melting of the
substrate was not possible due to the air-cooled voids.
The stream was not able to cut into and invade these
spaces unless the thin separating rind eventually col¬
lapsed. These rinds are as little as 10 cm thick (Allred
and Allred, 1997). Several extraneous tubes have been
identified in the cave. They are recognized as large,
egg-shaped bulges protruding from walls. Most tubes
ruptured from the contraction of hot walls during the
cooling of the cave, but some of them remain unbroken;
thus the extent of the voids within is unknown. The
longest extraneous tube found in Kazumura Cave is
nearly 1 kilometer long under an extraordinarily large
hump in the floor which diverted the stream to the side.
The thin separation finally collapsed during the cooling
of Kazumura Cave, allowing access into a master tube
which crosses nearly perpendicular to the Kazumura
passage. This passage contains lava formations unlike
any found in Kazumura Cave proper. Heligmites and
miniature volcanoes erupted out of the floors.
FIGURE 7 Near entrances or in spacious passages, later-formed
intermediate roofs insulated the lava stream.
Multilevel and Lavafall Development
Intermediate roofs are formed in some spacious pas¬
sages and below entrances that were open during
the eruption (Fig. 7). Cooling formed crusts which
reinsulated the lava stream. Because of these
ENCYCLOPEDIA OF CAVES
442
FIGURE 8 With increased turbulence in steeper slopes, the lava flow in Kazumura Cave cut back into the substrate, creating lavafalls and
plunge pools. This lavafall is 9 m high. Its molten 14-m-wide plunge pool is estimated to have been 27 m deep. Photo by Mike Shambaugh and
Kevin Allred.
secondary roofs, entrances in the cave system have not
completely bisected the underground continuity of the
cave. Up to four stacked levels are found in some
places. Breakout overflows through skylights probably
also occurred, but they probably did not leave much
connected passage unless they happened during the
early braiding stage. Lavafalls are well developed in
the cave closer to the caldera where steeper slopes con¬
tributed to greater turbulence. It is thought that some
headward backcutting occurred on many of the lava¬
falls until they reached a level, nonturbulent upstream
floor (Allred and Allred, 1997). Lavafalls can be up to
12.9 m high and their solidified plunge pools up to
15.5 m in diameter (Fig. 8). By calculating the amount
of crystallization contraction in the sunken plunge
pools, some of these are estimated to have eroded
20—30 m in depth.
LAVA SPELEOTHEMS
Primary Speleothems
In many parts of the cave, fluctuations in the flow¬
ing stream against a ceiling have formed sharktooth
stalactites composed of many layers of linings. Other
primary speleothems are splashed stalactites, stretched
lava, Pele's Hair, squeeze-ups, and lava blades which
sometimes resemble rillenkarren on a bare carbonate
karst. Boiling within partially solidified lava linings
has resulted in extruded forms such as tubular lava
stalactites above globular stalagmites, helictites, and
blisters (Allred and Allred, 1998).
Secondary Speleothems
Secondary deposits of white crusts and crystals appear
to be gypsum. They occur in dry places in the cave away
from rain seepage through the porous ceilings.
See Also the Following Article
Volcanic Rock Caves
Bibliography
Allred, K., & Allred, C. (1997). Development and morphology of
Kazumura Cave, Hawaii. Journal of Cave Karst Studies, 59(2),
67-80.
Allred, K., & Allred, C. (1998). Tubular lava stalactites and other
related segregations. Journal of Cave and Karst Studies, 60(3),
131-140.
Holcomb, R. T. (1987). Eruptive history and long-term behavior of
Kilauea Volcano. U.S. Geological Survey Professional Paper, 1350,
261-350.
Howarth, R. T. (1973). The cavernicolous fauna of Hawaiian lava
tubes. 1. Introduction. Pacific Insects, 15(1), 139—151.
Hulme, G. (1973). Turbulent lava flows and the formation of lunar
sinuous rilles. Modern Geology, 4, 107—117.
ENCYCLOPEDIA OF CAVES
KRUBERA (VORONJA) CAVE
443
KRUBERA (VORONJA) CAVE
Alexander Klimchouk
Ukrainian Institute of Speleology and Karstology, Simferopol,
Ukraine
INTRODUCTION
At the dawn of the new millennium, in January 2001,
Krubera (Voronja) Cave in the Arabika Massif,
Western Caucasus, became the deepest known cave in
the world, with a depth of —1710 m (Klimchouk and
Kasjan, 2001). For the first time, the deepest known
cave was found to be outside Western Europe. In the
article on Krubera Cave in the previous edition of this
encyclopedia (written in 2003), when the explored
depth of the cave was still at —1710m, the present
author wrote: "The future possibility of locating a
2000 m+ system in the area is exceptionally good."
Discovering the first cave on the planet deeper than
2000 m had been a long-standing dream of cavers
around the world, and this was set in 2001 as an official
goal of the Call of the Abyss project of the Ukrainian
Speleological Association (Ukr.S.A.), one of the most
ambitious and successful exploration projects in the his¬
tory of speleology. In October 2004 this goal was
reached, when Krubera Cave was pushed to depth of
2080 m. In subsequent years, the Ukr.S.A. expeditions
have explored the cave to depth of 2191 m in the main
branch, and also explored the second branch in this
cave, called Nekujbyshevskaja, to a depth of 1697 m.
THE ARABIKA MASSIF: LOCATION
AND PHYSIOGRAPHY
The Arabika Massif is one of the largest limestone
massifs of the Western Caucasus (Fig. 1A,B). It is
located in Abkhazia, a republic that officially belongs
to Georgia but since 1992 holds claim to being an
independent state.
To the northwest, north, northeast, and east,
Arabika is bordered by the deeply incised canyons of
Sandripsh, Kuturusha, Gega, and Bzyb rivers (Fig. 1C).
The Bzyb River separates Arabika from the adjacent
Bzybsky Massif, another outstanding karst area
with many deep caves, including the Snezhnaja-
Mezhonogo-Iljuzia system (—1,753 m) and Pantjukhina
Cave (—1508 m). To the southwest, Arabika is bor¬
dered by the Black Sea, with limestones dipping con¬
tinuously below the sea level.
The Arabika Massif has a prominent high central
sector with elevations above the tree line at
~1800—1900 m. This is an arena of classical Alpine-
type karst, a glaciokarstic landscape with numerous
glacial trough valleys and cirques, with ridges and
peaks between them. The bottoms of the trough val¬
leys and karst fields lie at elevations of 2000—2350 m,
and ridges and peaks rise to 2500—2700 m. The high¬
est peak is the Peak of Speleologists (2705 m) but
the dominant summit is a typical pyramidal horn of
the Arabika Mount (2695 m). Some middle- to low-
altitude ridges covered with forest lie between the
central sector and the Black Sea. A plateau-like
middle-altitude outlier of the massif in its south sec¬
tor is Mamzdyshkha, with part of the plateau slightly
emerging above the tree line.
KRUBERA CAVE AND OTHER DEEP
CAVES IN THE ARABIKA MASSIF
Among several hundred caves known in the Arabika
Massif, 15 have been explored below 400 m and five
below 1000 m (shown in Fig. 1C). Several are located
within the Ortobalagan Valley, a perfectly shaped, rela¬
tively shallow, perched glacial trough of the sub-
Caucasian stretch, which holds the advanced position in
the central sector toward the seashore (Fig. 2). Since 1980,
Ukrainian cavers have been undertaking systematic
efforts in exploring deep caves in the Ortobalagan Valley
including Krubera (Voronja) Cave (number 1 on Fig. 1C;
—2191 m), and the Arabikskaja System (number 4 on
Fig. 1C), which consists of Kujbyshevskaja Cave
(—1110 m) and Genrikhova Bezdna Cave (—965 m to the
junction with Kujbyshevskaja). Another deep cave in
the valley, located in its very upper part, is Berchilskaja
Cave (—500 m; number 11 on Fig. 1C) explored by
Moldavian and Ukrainian cavers. The Ortobalagan
Valley extends along the crest of the Berchilsky anticline,
which dips gently northwest.
An open-mouthed 60-m shaft, the Krubera entrance,
was first documented in the early 1960s by Georgian
researchers, who named it after Alexander Kruber, a
founder of karst science in Russia. The early exploration
was stalled by an impassable squeeze in a meandering
passage at — 95 m which led off from the foot of the
entrance shaft. During the 1980s, the main focus of
the Ukrainian expeditions to the Ortobalagan Valley
were Kujbyshevskaja Cave (—1110 m) and Genrikhova
Bezdna Cave (—956 m) connected into a single system in
1987 (the Arabikskaya System). At the same time, these
expeditions had pushed Krubera Cave from —95 to
—340 m by breaking through a series of critically nar¬
row meanders between successive vertical shafts.
During this period the cave received its secondary name
Voronja (Crow's Cave), owing to the number of crows
nesting in the entrance shaft.
Encyclopedia of Caves.
© 2012 Elsevier Inc. All rights reserved.
444
K
ichmarda
* f/amzdyshkha
'T-V- ■ • At
Black Sea
Bzybsky
Kaldakharsky Fault
Caves and their numbers
Submarine
springs
Sea-floor
topography
O Springs
db Wells
GV- Gegsky Vodopad
KR - Kholodnaja Rechka
BH - borehole
RE - Reproa
GO - Goluboje Ozero
A
r M#"
. 3 : 4 ?' <-
Romania
Bulgaria ,rf r
•Ukraine
FIGURE 1 (A) and (B): Location and shaded maps of the Arablka and Bzubsky massifs. The sea-floor topography is based on the
SRTM30plus data (NASA). (C) Deep caves are indicated by dots and numbers (explained in the text) and major springs are specified by two-
character indexes (explained in the legend). Solid black lines show the crests of major anticline folds. Red arrows reflect previous ideas on
groundwater basins and flow directions controlled by major fold structures. White and yellow arrows indicate the actual hydrologic connec¬
tions established by dye-tracing experiments in 1984—1985.
From 1992—1999 the explorations in Arabika were
suspended due to the Georgian-Abkhazian ethnic con¬
flict and subsequent turmoil. In 1999, the Ukr.S.A. expe¬
dition recommenced work in Krubera Cave and made a
major breakthrough by discovering and exploring two
branches that stretched from the old series in different
directions: the Main Branch to —740 m and the
Nekujbyshevskaja Branch to —500 m. The Main Branch
ENCYCLOPEDIA OF CAVES
KRUBERA (VORONJA) CAVE
445
Martel's
-956m
■1110m
Old Series
entrance
Krubera
-2191m
Berchil'skgja
Berc hHR id(
-500m
Berchil 2495nr
■2080m
A.KIimchouk, Ukr.S.A. - UISK, 2006
FIGURE 2 Major caves in the Ortobalagan Valley. Red dots indicate dolines.
was quickly pushed farther in 2000: in August to
— 1200 m and September to —1410 m. In January 2001,
the Ukr.S.A. expedition explored the cave to —1710 m,
establishing it as the new deepest cave in the world.
In 2004, the cave was explored to depth of 1840 m in
August, and to —2080 m in October. Further efforts in
the Main Branch led to reaching the terminal sump
at —2145 m in 2006. The current deepest point at
—2191 m was reached by the Ukr.S.A. expedition in
2007 through diving to —46 m in the terminal sump. In
the Nekujbyshevskaja Branch, systematic digging efforts
in boulder chokes since 2004 resulted in a series of
breakthroughs and the eventual exploration to the depth
of 1697 m in 2010. Most of the Ukr.S.A. explorations in
Krubera Cave since 1999 have been led by Yury Kasyan.
All the large caves of the Ortobalagan Valley belong
to a single hydrologic system, developed in and near the
crest zone of the Berchil'sky anticline. The direct connec¬
tion of Krubera Cave with the Arabikskaja System,
although not established yet, is a sound speleological
possibility. The Main Branch and the Nekujbyshevskaja
Branch in Krubera are largely independent, predomi¬
nantly vertical, parts of the system which deviate from
the Krubera old series at —220 to —240 m (Figs. 2 and 3).
Their upper sections (to altitudes of about 900—800 m)
cut through the thickly bedded and massive, often sandy
Upper Jurassic limestones. The caves are predominantly
combinations of vadose shafts and steep meandering
passages, although in places they cut apparently old fos¬
sil passages at different levels ( e.g ., at altitudes of
2070—2040 m in Kujbyshevskaja Cave, 1200—1240 m in
the Main Branch of Krubera, and 980—1150 m in the
Nekujbyshevskaja Branch of Krubera Cave, etc.). The
antiquity of these passages is supported by the ages of
speleothems falling beyond the 230 Th dating limit
(>500 ka). Both branches of Krubera Cave are extremely
vertical up to the depths of about 1600 to 1700 m (alti¬
tudes of about 750—650 m). The Main Branch below this
level becomes more inclined, largely following the strata
dip, but then it goes steeply down again to the depths of
about 2050 to 2150 m (altitudes of about 200—100 m).
Cave development is strongly controlled by the block-
fault structure. The cave entrances are aligned along the
anticlinal crest (Figs. 2 and 3) but the cave passages and
shafts are controlled by diagonal and orthogonal frac¬
tures and faults, and comprise complex winding pat¬
terns in the plan view, remaining largely within and
near the anticlinal crest zone. Two dominant diagonal,
SE—NW-stretching lines are recognizable in the plan pat¬
tern of Krubera Cave. The other diagonal direction, of
SW—NE orientation, is also used by many smaller frag¬
ments of Krubera Cave and by the most part of
Genrikhova Bezdna Cave. The south — north orthogonal
direction is expressed in a large fossil passage in the
ENCYCLOPEDIA OF CAVES
446
Ortobalagan glacial trough valley
Berchil’skaja
Arahikskaja System
Kujhyshevskaja
Krubera
2256m
Genrikhova Bezdna A
2110m
2180m,
Depth Alt.
{ Choke -100
Ugrjum-Zaval
l # ^p^-1363
3 All depth marks are given in meters,
~~P relative to the respective cave entrances.
_ ' Depth on the scale bar is given
\ relative to the Krubera entrance
Depth Alt.
2000 m
FIGURE 3 Combined cave profile projected on the axis of the Ortobalagan Valley. Krubera Cave is shown in white, other caves in light
brown.
ENCYCLOPEDIA OF CAVES
KRUBERA (VORONJA) CAVE
447
Nekujbyshevskaja Branch at altitudes of 980—1150 m, in
the terminal sump passage in the Main Branch, and also
in some parts of the nearby Kujbushevskaja Cave. The
deep portion of Krubera display a more pervasive con¬
duit pattern with a mixture of phreatic morphology,
characteristic of the zone of high-gradient floods, and
vadose downcutting elements that are observed even
below the water table (in the terminal sump).
The cave hydrology is very variable, depending on
season. During winter, flow through the cave is at a
minimum as recharge from the surface is virtually
absent. Late May and July are marked by the maximum
flux due to massive snow melting. During this period
of the highest flow, conduits in many parts of the sys¬
tem at altitudes below 1100—1000 m get filled with
water above local obstructions up to local levels of
10—20 m (the "bottleneck" effect). In the summer and
fall seasons, flow varies with the precipitation regime.
During low-flow periods, a small permanent stream (up
to 1 L s -1 ) first appears in the Krubera Main Branch at a
depth of about 340 m. The flow disappears and reap¬
pears at various levels, but never increases significantly,
even in the deepest sections. There are only small per¬
manent water flows in the Nekujbyshevskaya Branch,
but at the depth of 1660 m a large influent stream with
discharge of about 10 L s -1 comes from a tight side pas¬
sage. This water most likely arrives from the nearby
Kujbyshevskaya Cave, where the only stream of com¬
patible discharge is known, disappearing in a boulder
choke at the bottom at —1110 m.
The cave is very cold, with water temperatures of
1.0°C at —100 m depth, rising slowly up to 7.2°C at
2000 m depth. Air temperature is only a few tenths of
a degree higher. The phreatic zone effectively inter¬
cepts the geothermal heat flux and drains it through
springs. This leaves the vadose zone above under
strong influence of cold percolation that mainly comes
from snow melt at altitudes of about 2000—2300 m.
This results in an expressed geothermal depression
characteristic for alpine-type karst massifs.
Other deep caves in Arabika include the Iljukhina
System (—1273 m; number 3 in Fig. 1C) located in the
center of the massif, Dzou Cave (—1090 m; number 5
in Fig. 1C) and Moskovskaja Cave ( —1125 m; number 6
in Fig. 1C) in the northeastern part, and Sarma Cave
(—1760 m; number 13 in Fig. 1C) in the southeastern
part of the high sector of Arabika. Sarma Cave, now
the second deepest cave in the world, is located along
the same anticline as Krubera but on its southern
slope, at the similar advanced position relative to the
seashore.
THE ARABIKA MASSIF:_GEOLOGY
The Arabika Massif is composed of Lower Cretaceous
and Upper Jurassic limestones that dip continuously
southwest to the Black Sea and extend below the
modern sea level (Fig. 4). In the central part of Arabika
the Cretaceous cover is retained only in a few ridges and
FIGURE 4 Schematic geological and speleohydrological section of the Arabika Massif. See text for explanations.
ENCYCLOPEDIA OF CAVES
448
K
peaks, as well as in small patches within the trough val¬
leys (Valanginian and Hauterivan limestones, marls, and
sandstones). In the south and southeast, on the coastwise
ridges, the Cretaceous succession is more continuous and
includes Barremian and Aptian-Senomanian limestones
and marly limestones with abundant concretions of black
chert. The core part of the massif is composed of the
Upper Jurassic succession resting on the Bajocian
Porphyritic Series, which includes sandstones, clays, and
conglomerates at the top, as well as tuff, tuff sandstones,
conglomerates and breccia, porphyry, and lava. The
Porphyritic Series forms the nonkarstic basement of
Arabika, which is exposed only on the northern and east¬
ern outskirts, locally in the bottoms of the Kutushara and
Gega River valleys.
The Upper Jurassic succession begins from the bot¬
tom with thin-bedded Kimmeridgian-Oxfordian cherty
limestones, marls, sandstones, and clays, which are iden¬
tified in the lowermost part of Krubera Cave. Above lies
the thick Titonian succession of thick-bedded limestones
with marly and sandy varieties. Sandy limestones are
particularly abundant through the upper 1000-m sec¬
tions of deep caves of the Ortobalagan Valley.
The tectonic structure of Arabika is dominated by the
large sub-Caucasian (NW—SE oriented) anticline, with
the gently dipping southwestern megaflank, complicated
by several low-order folds, and steeply dipping north¬
eastern flank. The axis of the anticline roughly coincides
with the ridge bordering the Gelgeluk Valley to the north.
Located on the southwestern flank of the major anticline
is another large one (Berchil), in which the crest is brea¬
ched by the Ortobalagan Valley. There are several smaller
subparallel anticlines and synclines farther southwest,
between the Berchil anticline and the coast (Fig. 1C).
The plicative dislocation structure of the massif is
severely complicated by faults, with the fault-block
structure strongly controlling both cave development
and groundwater flow. Major faults of the sub-
Caucasian orientation delineate several large elongated
blocks that experienced uplift with different rates during
Pliocene and Pleistocene. This had a pronounced effect
on the development of deep groundwater circulation
and the karst system. Both longitudinal and transverse
faults and related fracture zones play a role in guiding
groundwater flow; the fracture zones guide flow across
the strike of major plicative dislocations, from the central
sector toward the Black Sea.
THE ARABIKA MASSIF:
HYDROGEOLOGY
Major onshore karst springs with individual average
discharges of 1— 2.5 m 3 s -1 are located at altitudes
ranging from 1 m (Reproa Spring) to 540 m (Gegsky
Vodopad). Two of them are located in the shore area:
these are Reproa (average discharge 2.5 m 3 s _1 , alt.
1 m asl; RE in Fig. 1C) and Kholodnaja Rechka
(1.2 m 3 s _1 , 50 m asl; KR in Fig. 1C). Other two major
springs are located in the river canyons bordering
Arabika to the east: Goluboe Ozero in the Bzyb canyon
(2.5 m 3 s ', 90 m asl; GO in Fig. 1C) and Gegsky
Vodopad in the Gega canyon (1 m 3 s -1 , 540 m asl; GV in
Fig. 1C). There are also several smaller springs in the
town of Gagra.
Some boreholes located along the shore of the
Black Sea yield karstic groundwater from depths of
40—280 m below sea level. Georgian hydrogeologists
reported that other much deeper boreholes tapped
karstic waters low in total dissolved solids at depths of
500 and 1750 m in the Khashupse Valley near
Tsandripsh and 2250 m near Gagra. This suggests the
existence of a deep karst system and vigorous karst
groundwater circulation at depth. Submarine springs
are known in the Arabika area, emerging from the
floor of the Black Sea in front of the massif. Shallow
springs at depths of 5—7 m below sea level can be
reached by free dive near Tsandripsh. Other subma¬
rine springs are known near the eastern part of Gagra
at depths of 25—30 m. Offshore hydrochemical profil¬
ing revealed submarine discharge at depths up to
400 m. Recently, an outstanding feature of the sea-floor
topography near Arabika has been recognized using
the DEM based on the SRTM30Plus dataset (NASA).
This is a huge submarine depression on the shelf in
front of the Zhoekvara River mouth, which has dimen¬
sions of about 5x9 km and a maximum depth of
about 380 m bsl. The Arabika Submarine Depression
(ASD; shown on Figs. 1C and 4) is a closed feature
with internal vertical relief of about 120 m (measured
from its lowest rim) separated from the abyssal slope
by the bar at a depth of about 260 m. It has steep
northern and northeastern slopes (on the side of the
massif) and gentle south and southwestern slopes. Its
formation is apparently karstic. Presently ASD seems
to be a focus of submarine discharge of the karst sys¬
tems of Arabika. The existence of ASD, along with
other lines of evidences, points to the possibility of
much lower sea level positions in the past than is sug¬
gested by Pleistocene glacioeustatic oscillations.
The hydrogeologic model for the Arabika Massif
that dominated before the 1980s did not allow a pos¬
sibility of hydrologic connection between the central
high sector of the massif and the coastal springs. The
model implied the existence of elongated groundwa¬
ter basins corresponding to synclines and separated
by anticlines of the sub-Caucasian trend, with sev¬
eral superimposed aquifers in each basin vertically
ENCYCLOPEDIA OF CAVES
KRUBERA (VORONJA) CAVE
449
separated by minor "nonkarstifiable" beds within
the carbonate succession (Kiknadze, 1972; 1979).
According to this notion, groundwater recharged
within the central part of the massif would flow
northeast, beneath the nonkarstic cover, and south¬
west to the Goluboje Ozero spring, as shown by red
arrows on Figure 1C. The recharge areas for the
coastal and submarine springs were assumed to be
only the proximal low-altitude ridges.
This model was disproved by speleological explora¬
tions and dye-tracing studies during the 1980s under the
coordination of the Institute of Geological Sciences of the
Ukrainian Academy of Science. A series of large-scale
dye-tracing experiments was conducted in Arabika in
1984 and 1985. Tracers injected in the Kujbyshevskaja
and Iljukhina caves were detected in the Kholodnaja
Rechka and Reproa springs, proving groundwater flow
to the south-southwest across major tectonic structures
over a distance of 13—16 km as the crow flies (shown by
white and yellow arrows on Fig. 1C). The tracer from
Kujbyshevskaja was also detected in a borehole located
between these two springs (BH on Fig. 1C), which yields
groundwater from a depth of 200 m below sea level. This
has been interpreted as an indication of the connection of
the cave with the submarine discharge. The large Central
Karst Ffydrologic System, which encompasses most of
the southeastern flank of the Arabika anticline, had been
identified in this way (Klimchouk, 1990). These experi¬
ments have revealed the deepest karst hydrologic system
in the world with its overall vertical range of about
2500 m (measuring to the borehole water-bearing hori¬
zon) or even 2700 m (measuring to the deepest reported
submarine discharge points). Another tracer was injected
in the Moskovskaja Cave (—970 m) and detected at the
Gegsky Vodopad spring, indicating the presence of a
karst hydrologic system comprising the northeastern
flank of the Arabika anticline (the Northern System). No
connections have been proven of any cave with yet
another major spring, Goluboje Ozero in the Bzyb River
canyon, although it apparently drains a large area of the
eastern sector of the massif (the hypothetical Eastern
Karst Hydrological System). It is not clear where Sarma
Cave (—1760 m) drains to, Goluboje Ozero to the south¬
east or Reproa to the southwest, at the shore.
The results of the dye-tracing tests have radically
changed notions of the hydrogeology of Arabika and
revealed its outstanding speleological perspectives,
strongly stimulating further efforts for exploration of
deep caves. They demonstrated that groundwater flow
is not subordinate to the fold structure but is largely
controlled by faults that cut across the strike of major
folds, and that the large part of the central sector of
Arabika is hydraulically connected to the springs along
the seashore and with submarine discharge points.
EVOLUTION OF THE DEEP CAVE
SYSTEM IN ARABIKA
Krubera Cave has an extremely steep profile and
reveals a huge thickness of the vadose zone. The lower
boundary of the vadose zone (the top of the phreatic
zone) is at an elevation of about 110 m at low flow,
which suggests a low overall hydraulic gradient of
0.007—0.008. Groundwater low in dissolved solids is
tapped by boreholes in the shore area at depths of
40—280, 500, 1750, and 2250 m below sea level, which
suggests the existence of a deep flow system with vigor¬
ous flow. Submarine discharge along the Arabika coast
is reported at depths up to ~ 400 m bsl. A huge closed
submarine depression is revealed at the sea floor next to
Arabika, with the deepest point of ~ 400 m bsl. It is dif¬
ficult to interpret these facts in terms of the development
of karst systems controlled by contemporary sea level,
or even within the range of its Pleistocene fluctuations
(up to -150 m).
A hypothesis is suggested that early karst systems in
Arabika could have originated in response to the
Messinian salinity crisis (5.96—5.33 Ma BP), when the
Black Sea (Eastern Paratethys) could have almost dried
up. The dramatic sea level drop of ~1500 m is well
established for the Messinian time in the adjacent
Mediterranean. French karstologists have recently dem¬
onstrated (e.g., Mocochain et ah, 2006) that the Messinian
crisis played a great role in karst development in the
Mediterranean region, where deep conduits had formed
in response to the Messinian lowering of the base level
and imposed a strong influence on subsequent karst
evolution.
The hypothesis that the dramatic sea level drop
could have taken place in the Black Sea basin during
the Messinian time had been put forward long ago
(Hsu and Giovanoli, 1979) mainly based on deep-sea
drilling data (DSDP Sites 379, 380, 381). It has been
strongly corroborated by recent studies of regional
geology, including data from bio- and magnetostrati¬
graphy of the key sedimentary sequences, seismic pro¬
filing (establishing the Messinian erosional surface in
the Eastern Paratethys), studies of deep-water delta
complexes, and so on.
Before the Late Miocene, the present coastal Western
Caucasus was a low- to middle-altitude mountain ter¬
rain. Temporary desiccation of the Black Sea in the
Messinian time established the base level at many hun¬
dreds of meters below the present level and caused con¬
duit development deep in the Arabika Massif. These
early systems were flooded after the Pliocene transgres¬
sion. Uplifts of the Arabika area during Pliocene and
especially Pleistocene were highly differentiated by elon¬
gate zones (blocks) of the sub-Caucasian stretch (parallel
ENCYCLOPEDIA OF CAVES
450
K
to the coast). The total uplift amounted to 2—2.5 km in
the central sector of Arabika, whereas it was minimal
(0.1—0.2 km) in the zone proximal to the coast.
Hydraulic continuity was always maintained across the
zones, between the main recharge area in the rising cen¬
tral sector of Arabika and the coastal zone and subma¬
rine springs. The presence of high conduit porosity of
the Messinian origin in the coastal/submarine sector
allowed the zone of high hydraulic gradient during
uplift to be pushed far inland, beneath the rising central
sector (Fig. 4). This created favorable conditions for the
enhanced conduit development at depth in the central
sector, quick adjustment of the water table to new uplift
pulses, and eventual development of a huge vadose
zone and extremely steep and deep cave systems such
as Krubera Cave. This was further favored by recurring
sea level drops up to —150 m during the Pleistocene,
which caused the steepening of hydraulic gradients
beneath the central sector of Arabika and enhancement
of conduit porosity in the upper part of the present
phreatic zone.
This evolution scenario is indirectly supported by
230 Th dating of speleothems from the deep parts of
Krubera Cave. Stalagmites from depths of 1640 m and
1820 m (elevations 640 m and 436 m asl) have yielded
ages older than 200 ka (max. 276 ka), which suggests
that the deep vadose zone already existed in the
Middle Pleistocene at the latest.
The above discussion demonstrates that it is not by
chance that the deepest cave in the world with the
exceptionally high vertical range of almost 2200 m has
been discovered in the Arabika Massif. There were
unique geological and paleogeographic preconditions
for that.
Acknowledgment
This article is an outcome of enormous efforts of several generations
of cavers who have explored the deep caves in Arabika over the past
30 years.
Bibliography
Hsu, K. J., & Giovanoli, F. (1979). Messinian event in the Black Sea.
Palaeogeography, Palaeoclimatology, Palaeoecology, 29(1—2), 75—94.
Kiknadze, T. Z. (1972). Karst of the Arabika Massif. Tbilisi, Russia:
Metzniereba.
Kiknadze, T. Z. (1979). Geology, hydrogeology and activity of limestone
karst. Tbilisi, Russia: Metzniereba.
Klimchouk, A. B. (1990). Karst circulation systems of the Arabika Massif.
Peschery (Caves). Perm: Perm University.
Klimchouk, A., & Kasjan, J. (2001). Krubera (Voronja): In a search
for the route to 2000 meters depth: The deepest cave in the world
in the Arabika Massif, Western Caucasus. NSS News, 59(9),
252-257.
Mocochain, L., Clauzon, G., Bigot, J. Y., & Brunet, P. (2006).
Geodynamic evolution of the perimediterranean karst during the
Messinian and the Pliocene: Evidence from the Ardeche and the
Rhone Valley systems canyons, southern France. Sedimentary
Geology, 188-189, 219-233.
ENCYCLOPEDIA OF CAVES
LAMPENFLORA
Janez M ulec
Karst Research Institute, Scientific Research Centre of the
Slovenian Academy of Sciences and Arts
DEFINITIONS
Lampenflora—originally a German term adopted in
the English vocabulary, sometimes called lamp flora, or
in French la maladie verte —designates the phenome¬
non—proliferation of phototrophic organisms near
artificial light sources. Lampenflora grows at sites
where, under natural circumstances, it would not
appear. This alien flora can be found in the proximity
of light sources in caves, mines, and other artificial
and light-deprived environments, such as cellars and
stores (Fig. 1). Besides artificial light, high humidity,
typically around 90%, which sometimes approaches
saturation, is also a characteristic for these environ¬
ments. Humidity and, occasionally, seeping water,
together with light, support the growth of photo-
trophic organisms. Various types of aerophytic cyano¬
bacteria and algae, as well as some mosses and ferns,
dominate. The term aerophytic designates living on air
in terrestrial habitats, on rocks, stones, sediments, bark
of trees, needing only the atmospheric moisture
(humidity) to sustain life cycle or part of it. Several
other terms are also used to define the ecology of
lampenflora phototrophs, such as subaerophytic, indi¬
cating organisms living in aerophytic localities with an
intermittent water supply. Sometimes in this context
the term subaerial is used, describing the location of the
habitat which is near the Earth's surface, that is, under¬
ground. To describe interactions between an organism
and a solid substratum, the term epilithic is used for
lampenflora as well. It is generally used for organisms
attached to stones or rocks or their growth on stony
substrata, aerophytic, or submerged under (the film of)
water. The term endolithic denotes growth of biota
under the surface of stones and rocks, including small
microcavities and fissures. Organisms are usually
strongly adhered to the substratum with their thalus.
In the early phase of colonization and succession, cya¬
nobacteria and eukaryotic algae usually play the most
important role, while mosses and ferns appear later in
succession. Vascular plants are sometimes found
around lamps, but almost always only as germinating
shoots. Using the plant's reserves some growth can
occur, even in the absence of light (mixotrophy). Only
lampenflora in show caves and tourist mines as a
result of artificial lighting is discussed here.
Many caves are recognized because of natural
beauty, amazing cave formations, and places of inspi¬
ration. Humans developed intimate bonds with caves
early in history, as they offered temporary or perma¬
nent shelter for our ancestors and, later, caves became
attractive for visiting and recreation. Especially inter¬
esting are those caves with prehistoric paintings,
such as Lascaux Cave in France. A few dozen caves of
natural and cultural importance are listed on the
United Nations Educational, Scientific, and Cultural
Organization (UNESCO) World Heritage List (World
Heritage Convention). Cave tourism is considered to
be one of the oldest forms of tourism. Nowadays in
many of these places, artificial illumination has been
installed to attract visitors. Illuminated areas such as
rocky surfaces, sediments, and artificial materials
around lamps quickly become colonized by photo-
trophic organisms. The sites of tourist interest, such as
prehistoric paintings or extraordinary cave formations,
are frequently well lit to present them to the public,
and consequently they become suitable substrata for
lampenflora. A clear distinction between naturally
present aerophytic microscopic phototrophs and high¬
er plants in cave entrances and lampenflora is that
lampenflora organisms colonize natural and artificial
substrata deep in the underground and are completely
dependent on artificial lighting. Lampenflora in caves
Encyclopedia of Caves.
451
© 2012 Elsevier Inc. All rights reserved.
452
L
FIGURE 1 Lampenflora in Postojnska Jama, Slovenia, 2008.
Photograph taken by Svatava Kubesovd.
can also be found attached to solid surfaces in illumi¬
nated submerged cave formations, such as cave ponds
and riverbeds. Although lampenflora is a result of
light eutrophication underground and a cause of
biodeterioration or biofouling of various types of sub¬
strata in the underground, it has been preserved in
some caves as a tourist attraction.
COMPOSITION AND ECOLOGY
OF LAMPENFLORA
Every exposed solid surface is subject to coloniza¬
tion by organisms, and in this aspect the underground
is no exception. With the introduction of light on a
solid surface in light-deprived underground habitats,
phototrophic organisms become competitive for this
ecological niche. The problem with the growth of lam¬
penflora appears with installation of permanent elec¬
tric lighting in show caves and tourist mines. The use
of carbide lamps or portable electric lights, such as
LEDs (light-emitting diodes), during cave exploration
does not sustain the growth of phototrophs. Generally,
various abiotic parameters play important roles for
development and community structure of phototrophs:
light, temperature, characteristics of the substratum,
carbon dioxide, oxygen, availability of nutrients, and,
if present, the characteristics of surrounding water
(streaming, pH, alkalinity, acidity, hardness, salinity,
pollutants). Among biotic factors, the following are
important: competition, parasitism, predation, and
organic exudates.
Diversity of lampenflora depends mainly on lighting
regime, quantity, and quality—light spectrum of avail¬
able energy. Compared to sunlight, the artificial light
source shows no oscillations in intensity. An incandes¬
cent lamp or a fluorescent lamp does not notably influ¬
ence the community composition. When a cave is lit
for a longer period, flora becomes more diverse and
succession is faster. Diversity is further dependent on
temperature, and quantity and routes of introduced
cells, spores, or other viable propagules underground.
At low photosynthetic photon flux densities (PPFDs),
a small increase in temperature can result in a large
increase of biomass yield. For example, the green
alga Chlorella sp., frequently identified in lampenflora
communities, increased biomass 30-fold at 2.5 |imol
photons m -2 s~ 1 when the temperature increased from
9 to 11°C. At cave temperatures, the light saturation
point is fast reached at low PPFDs (<10 |imol photons
m -2 s -1 ), and biosynthesis of accessory photosynthetic
pigments below the light saturation point is consider¬
ably elevated. Lampenflora can be found at PPFDs
lower than 0.2 p mol photons m 2 s -1 up to several sev¬
eral hundred |imol photons m 2 s -1 .
In a contrast to cave fauna that can move around,
lampenflora uses different modes of dispersion in the
underground. Three main modes of transport of viable
propagules in the underground can be distinguished:
air currents, water flow, and introduction by migratory
animals and humans. Caves with mass tourism are
subjected to high input of lint and other detritus from
visitors which serve as a nutrient source. Air currents
in caves appear due to meteorological changes of
external climate (oscillations of temperature and air
pressure, and seasonal dynamics) and consequently
these changes are reflected in underground spaces. An
important factor in spreading lampenflora inside show
caves is a local air current, which is formed because of
heating of cave air with strong, mainly halogen lamps.
In such places lampenflora growth was observed in a
distance more than 10 m from a lamp. An interesting
observation came from Aggtelek Cave, Hungary,
where green patches appeared after snow melt in
spring, but when water evaporated, the green coating
in the immediate vicinity of lamps disappeared.
Extremely strong lighting dried the surface and
restricted lampenflora growth. But in the zone
more distant from the lamps, flora continued to grow.
ENCYCLOPEDIA OF CAVES
LAMPENFLORA
453
Seeping water passing through crevices and cracks
brings viable propagules from the epikarst zone above
the caves. Vegetation in caves thrives better in places
with seeping water. Abundance of algae is probably
also dependent on air humidity. As mentioned above,
in the immediate vicinity of incandescent or halogen
lamps humidity drastically drops. An important source
of allochthonous organisms entering underground is
the formation of bioaerosols from fine mist or spray
at a ponor due to splashing of a river. Microscopic
allochthonous organisms, such as algae, are then fur¬
ther transmitted in a cave with air currents.
Solid surfaces subjected to colonization, which are
covered with sediments, enable faster growth of 1am-
penflora; in addition, rough surfaces with minor irregu¬
larities represent a better substratum for its attachment.
Substratum characteristics at the micro level notably
influence lampenflora growth and its distribution
around lamps. This may be why there is often an
absence of correlation between PPFDs and the concen¬
trations of main photosynthetic pigment—chlorophyll a
at individual lamps.
In comparison to naturally growing prototrophs in
cave entrances, diversity of lampenflora is poor.
Generally, in comparison with cave entrances, areas
around lamps have more stable conditions with no
extreme environmental oscillations, for example, free¬
zing-thawing. Typical representatives are soil aero-
phytic cyanobacteria and algae. Organisms successful
in colonizing a new ecological niche are cosmopoli¬
tan—ubiquitous in nature—having simple nutrition
requirements and wide ecologic tolerance; they repro¬
duce fast and are easily adaptable to new conditions.
Taking only microscopic lampenflora into consider¬
ation, that is, cyanobacteria and algae, it is not possible
simply to define their general distribution pattern. In
the species succession, the first colonists are usually
fast growing and are slowly ousted by the more persis¬
tent organisms. Although at first sight the community
of lampenflora seems static, it is not so as the species
succession process is rather dynamic. In the develop¬
ment of a phototrophic community, a random combi¬
nation of species first occurs; vegetation zones are
formed later, usually after many years. However, in
this phase an important factor to initiate growth is the
presence of appropriate inoculum.
On occasion, cyanobacteria are pioneering organ¬
isms in many harsh environments (Whitton and Potts,
2000), but the first colonizers are frequently fast-grow¬
ing green algae which are gradually overgrown by
cyanobacteria. On electrically lit sites with high
humidity, or with condensed water on solid surfaces,
and especially at sites with seeping water, diatoms
slowly become more abundant in the community.
Diatom adhesion can occur in the light or dark, but for
successful adhesion energy is required. One of impor¬
tant factors influencing diatom adherence is the pres¬
ence of Ca 2+ ions. Greenish patches of microscopic
cyanobacteria and algae are progressively becoming
covered by islands of mosses and finally ferns.
Cyanobacteria are usually the most abundant (~50%),
followed by Chrysophyta or Chlorophyta. Among cya¬
nobacteria, the following genera are frequently encoun¬
tered: Aphanocapsa, Aphanothece, Chroococcus, Gleocapsa,
Leptolyngbya , Myxosarcina, and Synechocystis. Among
green algae are Apatococcus lobatus, Chlorella spp.,
Chlorococcum spp., Coccomyxa spp., Klebshormidium flacci-
dum, and Stichococcus spp. Among diatoms in lampen¬
flora communities, researchers report on frequent
presence of the genera Achnanthes, Aulacoseira, Cocconeis,
Diadesmis, Fragilaria, Hnntzschia, Navicida, and Nitzschia.
Most mosses identified in caves are not specific for
this habitat and can exist also in other places with
very low light irradiances. Low PPFDs permit bryo-
phytes to live in places inhospitable to other plants.
Bryophytes in caves sometimes express etiolation—
adaptation to low quantity of available light by using
their physiological capacity to expose more surface
area to capture the few available photons. For exam¬
ple, in some species the number of chloroplasts and
size of grana can increase in response to low light.
Some bryophytes are able to grow over a relatively
wide range of PPFDs (Glime, 2007). In caves, moss
protonema frequently grow. Tufa-forming moss,
Eucladium vertidllatum, is a frequent dweller in poorly
and powerfully lit places in caves; for example, in
Slovenian show caves E. vertidllatum had the highest
ecological tolerance width based on the span of
PPFDs, from 1.4 to 530.0 pmol photons m -2 s -1 .
Sometimes in lampenflora communities even at low
PPFDs, bryophytes can complete their life cycle by
longer exposure to light irradiance. Such an example
is Cratoneuron filidnum which lives in zones where
lights are on 24 h d" 1 with PPFD between 2.1 and
2.4 11 mol photons m 2 s (Mulec and Kubesova,
2010). Besides species succession of algae and cyano¬
bacteria, succession in the bryophyte community
should be taken into consideration as well. Ferns are
also found in the lampenflora community, but their
diversity is low, and frequently they are found only
as prothalli.
ALTERATION OF UNDERGROUND
HABITAT
The cave ecosystem is in dynamic equilibrium,
which is disturbed when high amounts of nutrients
and/or energy—light have been introduced under¬
ground. Newly formed lampenflora biomass is
ENCYCLOPEDIA OF CAVES
454
L
available for cave-adapted animals and other occa¬
sional dwellers of caves. Such new conditions indi¬
rectly influence cave fauna as they are adapted to
(nearly) oligotrophic conditions. Along with easier
available nutrients, newcomers become competitive for
a new ecological niche and a population of troglobiotic
animals can be affected. Terrestrial troglobiotic species
are frequently observed grazing on lampenflora.
Vegetation and other heterotrophic microbial com¬
munities do not change just the aesthetic appearance
of a cave, but also mechanically and biochemically
cause the deterioration of cave paintings and spe-
leothems. Organisms interact with rock, mobilize ions,
and release compounds. Electron transfer, proton uptake,
removal of cations from the substrate due to the forma¬
tion of chelating complexes, and phototrophic processes
with transfer of carbon dioxide, oxygen production, and
alkaline reaction, all contribute to biodeterioration of the
stony substrata (Warscheid and Braams, 2000). At the
micro level, changes of pH are based on the lighting
regime, increase of pH during photosynthesis, and low¬
ering of pH during respiration or fermentation
(Albertano et ah, 2000). The mucilage surrounding cells
acts as a slimy adhesive for airborne particles. In addi¬
tion, these polysaccharides help to retain humidity,
increase adhesion to the substratum, and stabilize dust
particles.
A very important problem occurs when live or dead
lampenflora becomes encrusted with calcium carbonate,
irrespective of whether this carbonate is a result of abi¬
otic or biotic precipitation. Limestone and the other cal¬
careous substrata, which often occur in subterreanean
environments, can promote the growth of calcifying spe¬
cies. Such an amorphous mix of dead phototrophs and
carbonate irreversibly destroys speleothems or other
objects of cultural value.
CONTROL AND RESTRICTION
OF GROWTH
Cave management faces two urgent problems
regarding lampenflora: (1) how to remove lampenflora
without damaging the substratum, once it is estab¬
lished; and (2) how to prevent its growth and biodete¬
rioration in the future (Fig. 2). Many soil algae, which
are common in lampenflora, are known to have also
heterotrophic metabolism, that is, organisms which
use organic compounds for most or all of their carbon
requirements. In addition to these, some cyanobacteria
can even use fermentative biochemical pathways to
obtain energy (Whitton and Potts, 2000). These physio¬
logical adaptations allow these organisms to survive
periods with no available light, which is why the
decreasing PPFDs or shortening of time of exposition
to light irradiance is not a satisfactory and long-lasting
solution. Even short but continuous pulses of light can
drive the light phase of photosynthesis.
Problems connected with lighting and phototrophic
biofilms are usually not properly solved. The main
obstacle is that the cause, that is, light, remains on the
site and new viable propagules are constantly arriving.
The proper way to prevent lampenflora growth is still
unknown. The simplest solution would be complete
removal of existing phototrophic communities, cessa¬
tion of lighting, and abolition of tourist visits, which in
many cases is not acceptable for the cave management.
FIGURE 2 Lampenflora encrusted in flowstone.
Contact between flowstone and algae coating. Image
was taken with low vacuum scanning electron
microscope 5500 LV JEOL (back scattered electron
image). Photograph taken by Alenka Mauko.
ENCYCLOPEDIA OF CAVES
LAMPENFLORA
455
Many different procedures have already been
applied to control lampenflora, including physical and
chemical approaches (Mulec and Kosi, 2009). To elimi¬
nate lampenflora from cave formations, mechanical
removal with water and brushes is usually applied.
This is not a sustainable approach because many frag¬
ile formations are destroyed. Introduction of ozone-
producing lamps or UV, when tourists are not visiting,
also may negatively affect cave biota.
In the past, in order to kill lampenflora several very
toxic compounds were applied, such as DCMU (diuron,
N-(3,4-dichlorophenyl)-N'-di methyl urea), bromine
compounds, formalin, cupric ammoniac solution,
which are nowadays totally unacceptable to introduce
into the cave environment. To kill lampenflora nowa¬
days, such places are frequently sprayed with 5%
bleach solution (NaCIO) that effectively kills lampen¬
flora. From the NaCIO solution, gaseous chlorine can be
released which gives its characteristic smell in a cave.
In the reaction of hypochlorite and nitrogenous com¬
pounds, toxic chloramines and even carcinogenic triha-
lomethanes are released. Low chlorine concentration in
the cave environment can kill other cave biota. Less
toxic, odor-free, and more environmentally friendly is
hydrogen peroxide (H 2 O 2 ) which is already applied in
some caves; however, the pH of unbuffered 15% H 2 O 2
solution is 4 and thus corrosive for limestone and other
formations. Oxidized plant material later becomes a
source of nutrients for bacteria and fungi, and for this
reason dead material should be removed from a cave.
Unfortunately, all the above-mentioned compounds are
not specifically targeted against vegetation and have a
biocidal effect for the cave fauna as well. In the future,
more studies should go in this direction. There are
reports on the potential use of selective herbicides such
as Atrazine (6-chloro-/V-ethyl-/V'-(1 -methylethyl)-l ,3,5-
triazine-2,4-diamine) and Simazine (6-chloro-N,AT-
diethyl-l,3,5-triazine-2,4-diamine), but their wide¬
spread use underground is not suitable because the
green coloring on the karst formations persists
(Grobbelaar, 2000). The use of bleach or other nonspe¬
cific oxidizing compounds to control lampenflora in
places with art is likely to damage features of signifi¬
cance, for example, inscriptions, prehistoric paintings,
frescoes, stuccoes, and mosaics, and is thus strongly
dissuaded for managing such places. The crucial step in
limiting the adverse impacts of lampenflora is to start
with treatment as soon as the lampenflora is visible.
Future perspectives regarding lampenflora manage¬
ment should undoubtedly include development of a
procedure which would kill lampenflora with no or
minimum impact for caves. Still, before applying any
procedure to remove lampenflora it seems reasonable
to check for their species composition, but such data
can be of inventory value. Even when a new taxon is
FIGURE 3 Barbarossahohle, anhydrite show cave in Germany.
Note color of water, lightened with cool white LED and additionally
with red LED light. Photograph taken by Alexander Chrapko.
identified or a peculiar community structure revealed,
one has always to keep in mind that this alien flora
normally does not occur in underground environments,
and its occurrence is directly linked to human interven¬
tion in the underground.
An important step in controlling lampenflora
growth is an appropriate installation of lamps, hous¬
ings, and modes of illumination: short-time lighting of
individual sectors, avoidance of damp surfaces, places
covered with sediments and easy accessibility for tour¬
ists, places with air currents, dripping, and seeping
water. Lamps installed in hypogea should emit light
with a spectrum out of the absorption maxima of chlo¬
rophyll a. If we strictly follow this direction, green
lighting does not give a natural view of the under¬
ground (Roldan et al. r 2006). Chlorophyll a has two
peaks in the absorption of red and blue parts of a light
spectrum. Another possible alternative is yellow light¬
ing because algae do not absorb strongly in the yellow
part of the light spectrum. The common observation in
hypogea where colored filters are used to restrict lam¬
penflora growth is that there is less growth compared
to white lights. A very promising recent approach is
the installation of LED lighting (Fig. 3) because of low-
energy consumption, long-lasting LEDs, and potential
to tune the desirable emission spectrum (Toomey et ah,
2009). With recent advances in technology, the use of
fiber optics in cave lighting seems another alternative.
With optimization and more studies in cave lighting,
such installations have a potential for enhanced control
to reduce lampenflora growth.
See Also the Following Article
Show Caves
ENCYCLOPEDIA OF CAVES
456
Bibliography
Albertano, P., Bruno, L., D'Ottavi, D., Moscone, D., & Palleschi, G.
(2000). Effect of photosynthesis on pH variation in cyanobacterial
biofilms from Roman catacombs. Journal of Applied Phycology, 12,
279-384.
Glime, J. M. (2007). Light: Cave mosses. In Bryophyte ecology, physio¬
logical ecology (Vol. 1, Chapter 9-5, pp. 43—52) (Ebook sponsored
by Michigan Technological University and the International
Association of Bryologists). Accessed on 3 October 2011 at
http://www.bryoecol.mtu.edu/.
Grobbelaar, J. U. (2000). Lithophytic algae: A major threat to the karst
formation of show caves. Journal of Applied Phycology, 12, 309—315.
Mulec, J., & Kosi, G. (2009). Lampenflora algae and methods of
growth control. Journal of Cave and Karst Studies, 71, 109—115.
Mulec, J., & Kubesova, S. (2010). Diversity of bryophytes in show
caves in Slovenia and relation to light intensities. Acta Carsologica,
39, 587-596.
Roldan, M., Oliva, F., Gonzales del Valle, M. A., Saiz-Jimenez, C., &
Hernandez-Marine, M. (2006). Does green light influence the fluo¬
rescence properties and structure of phototrophic biofilms?
Applied and Environmental Microbiology, 72, 3026—3031.
Toomey, R. S. Ill, Olson, R. A., Kovar, S., Adams, M., & Ward, R. H.
(2009). Relighting Mammoth Cave's new entrance: Improving vis¬
itor experience, reducing exotic plant growth, and easing mainte¬
nance. In W. B. White (Ed.), ICS 2009 15th International Congress of
Speleology Proceedings (Vol. 2, pp. 1223—1228). Kerrville, TX:
Greyhound Press.
Warscheid, T., & Braams, J. (2000). Biodeterioration of stone: A
review. International Biodeterioration and Biodegradation, 46,
343-368.
Whitton, B. A., & Potts, M. (Eds.), (2000). The ecology of cyanobacteria:
Their diversity in time and space Dordrecht: Kluwer Academic
Publishers.
LECHUGUILLA CAVE,
NEW MEXICO, U.S.A.
Patricia Kambesis
Cave Research Foundation
Prior to 1986, Lechuguilla Cave was known as a minor
historic site in the back country of Carlsbad Caverns
National Park. At that time the cave contained 120 meters
of dry, dead-end passages that were accessed via a 27
meter entrance shaft known as Misery Hole. In the 1950s
cave explorers noted a significant amount of wind that
issued from a rubble pile at the bottom of the known
cave, indicating the potential for more cave passages,
and over the years several attempts were made to dig
out the rubble pile. In 1984, on a park-approved permit,
cavers once again worked on finding a way through the
rubble choke. In May of 1986, after two years of digging
efforts, a team of explorers broke through into a major
cave passage in Lechuguilla Cave. This began one of the
most remarkable chapters in American cave exploration
and speleology. The cave quickly grew in length and
depth and became known as the deepest limestone
cavern in the United States and ultimately fourth in
length for U.S. caves. On the world scale it is the fifth
longest.
Lechuguilla Cave has been called the "jewel of the
underground," and rightfully so, with its incredible vari¬
ety of spectacular speleothems that decorate many of
the cave's massive chambers, circuitous tunnels, and
confounding mazes. However, it is not just the cave's
aesthetic beauty and explorer's appeal that define
Lechuguilla as a world-class cave. Due in large part to
the stewardship of the National Park Service and the
foresight of the American caving community, science
has been a part of the exploration efforts in the cave
since the very beginning (Turin and Plumer, 2000).
Consequently, important studies in geology, cave micro¬
biology, microclimatology, mineralogy, geomicrobiology,
and geochemistry have been conducted in the cave.
Scientific discoveries have gone hand-in-hand with cave
exploration and survey.
Geologically, Lechuguilla Cave is unusual in that its
speleogenesis involves sulfuric acid and the degassing
of hydrogen sulfide from a hydrocarbon-rich basin. The
concept of basinal degassing is significant not just for
cave formation in the area but also to the understanding
of migration of hydrocarbons, the formation of petro¬
leum reservoirs, the deposition of uranium ore, and the
origin of Mississippi Valley—type lead-zinc deposits
(DuChene and Hill, 2000). Lechuguilla Cave is one of
the world's best-documented examples of a cave formed
by sulfuric acid in a hypogene setting
Mineral residues and some of the more unusual
speleothems in Lechuguilla Cave show indications of
microbial influence on cave development. The signifi¬
cance of these geomicrobiological interactions is pro¬
viding insight into the basic mechanisms of dissolution
and precipitation by microorganisms and on the origin
of specific types of speleothems (Northup et ah, 2000).
The microbe environment in Lechuguilla Cave is also
being studied by NASA scientists who believe that it
may be an analog to similar environments on other
planets.
The exploration and concurrent survey of Lechuguilla
Cave have influenced the modes and methods of cave
mapping in the United States. The abundance and deli¬
cacy of formations, unusual sediments, and microbiol¬
ogy within the cave have caused both explorers and
resource managers to reevaluate human impact on all
caves in general and has raised the profile of cave con¬
servation and restoration efforts.
PHYSICAL SETTING
Lechuguilla Cave is located within Carlsbad Caverns
National Park, which is part of the Guadalupe Mountains
Encyclopedia of Caves.
© 2012 Elsevier Inc. All rights reserved.
LECHUGUILLA CAVE, NEW MEXICO, U.S.A.
457
FIGURE 1 Location map of Carlsbad Caverns National Park, Eddy County, New Mexico.
of southeastern New Mexico (Fig. 1). These mountains
are situated on the northern margin of the Chihuahan
Desert. The area climate is semi-arid to arid with average
winter temperatures of 7°C and summer temperatures of
27°C. The annual rainfall ranges from 20—50 cm with
more than half of the precipitation coming during the
summer months. Vegetation of the area includes cacti,
succulents, and desert shrubs. The cave's name comes
from one of the local flora. Agave lechuguilla, which grows
in large clumps on limestone ledges and slopes, and is a
common plant in Lechuguilla Canyon and at the entrance
of the cave.
GEOLOGIC SETTING
Lechuguilla Cave is developed within the Capitan
Reef Complex which forms a belt of Permian-aged
carbonates 8 km wide and 650 km long and defines
the northwest perimeter of the Delaware Basin, a
major hydrocarbon basin in the southwest United
States. The Delaware Basin lies between the structur¬
ally active Basin and Range Province and the non-
structurally active Great Plains (Hill, 2000). The
Guadalupe Mountains are a fault block that has been
exhumed by erosion and evaporite solution-subsi¬
dence within the Delaware Basin.
Though the Guadalupe Mountains are considered a
karst area, features typical of classic karst areas are rel¬
atively rare. There are some sinkholes, paleokarst, and
a variety of springs both small and large, but there is
no direct correlation between the surface topography
and geomorphology and the extensive caves that lie
within the mountains. This is a reflection of the cave's
development in a hypogene setting. Cave entrances
GUADALUPE RIDGE
FIGURE 2 Rock units exposed in Lechuguilla Cave. After Jagnow,
1977.
are not past or present points of hydrologic recharge
or discharge. Rather, they are fortuitous collapses that
have exposed the caves to the surface. In all of its great
length, Lechuguilla Cave has only one entrance.
STRATIGRAPHY AND STRUCTURE
Paleontologic and stratigraphic data from rock
exposures within Lechuguilla Cave indicate that most
of the cave is developed either in the back reef facies
of the Permian reef or within the reef and forereef
slope facies. The cave spans four stratigraphic units
including (in descending order) the Yates Formation,
Seven Rivers Formation, Queen Formation, and Capitan
Formation (Fig. 2). Backreef facies include the Yates,
Seven Rivers, and Queen formations. The reef and fore¬
reef units encompass the Capitan Formation. It has
been speculated that the cave could extend into the
ENCYCLOPEDIA OF CAVES
458
L
upper reaches of the Goat Seep Formation, a unit that
consists of massive reef and forereef talus of dolomite
composition, deposited along the shelf margin.
However, to date no Goat Seep exposures have been
documented in the cave.
Rock units exposed in the cave include the Yates
Formation which is composed of dolomites, calcareous
quartz siltstones, and fine-grained sandstones. This unit
is exposed at the entrance of Lechuguilla Cave, in the
upper levels of the Chandelier Graveyard, in Tower
Place, and in many of the other upper level passages.
The Seven Rivers Formation consists of dolomites and,
occasionally, quartz siltstones. This formation is evident
throughout most of the upper parts of the cave from the
entrance through the Rift. The Queen Formation is
made up of dolomites interbedded with siltstones
and fine sandstones. Within the cave this formation
is exposed in Windy City, Sugarlands, and in the north
and south ends of the Rift.
The Capitan Formation is divided into massive and
breccia members. The massive member lacks bedding
and consists of fossiliferous and dolomitic limestones.
The breccia member is made up of reef limestone frag¬
ments mixed with lenses of siliciclastic material. The
massive Capitan outcrops at depths of 175—275 m
within the cave. The breccia member outcrops at
depths below 250 m throughout the cave.
The major folds within the vicinity of Lechuguilla
Cave include the Guadalupe Ridge anticline and the
Walnut Canyon syncline. The cave lies within the
structural limb shared by these two folds. Joint sets
trend either parallel or perpendicular to these major
folds. Cave passage orientations are controlled by
these two major joint sets. Lechuguilla Cave is located
within a structural trap for natural gas located just off
the Guadalupe Ridge anticline. The cave is also located
beneath the Yates Formation which functions as a
stratigraphic trap for H 2 S (Hill, 2000).
The development of Lechuguilla Cave was influenced
by the arrangement of stratigraphic units, differential
dolomitization of these units, joint patterns, and folds
(Hill, 2000). All four controls first appeared in the Late
Permian with some modification by later tectonic and
diagenetic events (Hill, 1996). However, the overall lay¬
out and geometry of the cave is a function of regional
groundwater flow and sulfuric acid production (Palmer
and Palmer, 2000). Because the genesis and development
of the cave is not coupled to surface hydrogeology,
Lechuguilla Cave is classified as a hypogene cave system.
REGIONAL/LOCAL HYDROGEOLOGY
The permeable Capitan Limestone is the aquifer for
the region and the upland surface of the Guadalupe
Mountains is the recharge area. Water moves through
the shelf rocks of the Artesia Group and drains
down-dip into the Capitan Limestone aquifer. Water
moves slowly through the reef and discharges into the
Pecos River at Carlsbad Spring in the city of Carlsbad,
New Mexico. The caves of the Guadalupe Mountains,
including Lechuguilla Cave, show a combination of
deep phreatic and water-table characteristics. Water-
table conditions are responsible for horizontal level
development of cave passages. In these areas, cave pas¬
sages cut across bedding planes. The vertical develop¬
ment of the system is a function of deep phreatic
conditions. Lechuguilla Cave was formed in a diffuse-
flow aquifer regime. The cave may have functioned as a
mixing chamber for hypogene-derived H 2 S and mete¬
oric derived freshwater (Hill, 2000).
The Aretesia Group (Yates, Seven Rivers, Queen, and
Goat Seep formations) was probably the original reser¬
voir where water containing hydrogen sulfide was
stored prior to and during sulfuric acid speleogenesis.
The aquifer also functioned as the avenue for eastward
migration of hydrogen-sulfide water with oxygenated
meteoric water (DuChene, 2000). The aquifer possibly
also served as the site where the mixing of H 2 S water
with oxygenated water occurred and which ultimately
formed the passages of Lechuguilla Cave. Hydrogen
sulfide speleogenesis took place at or near the water
table (Palmer and Palmer, 2000).
Over time, as the water table dropped, cave passage
development occurred where H 2 S was ascending to
the water table in significant quantities (Jagnow, 1989).
As the water table lowered during uplift and deforma¬
tion, H 2 S was episodically released, forming major
rooms and passages (Palmer and Palmer, 2000). When
H 2 S occurred during times when the water table was
stable, horizontal levels of cave passages developed.
The massive rooms and long passage trends within
Lechuguilla Cave reflect these events (Fig. 3).
The relationship of the ages of the mineral alunite
with respect to elevations within the cave suggests
that there was an 1100-meter decline in the water
table from 12 million years ago to the present
(Polyak et al., 1998). This suggests that the Capitan
aquifer water table was relatively flat during the
Late Miocene and Pliocene (Polyak & Provencio,
2000). It is speculated that the paleospring that would
have been associated with Lechuguilla Cave has not
been preserved because it has been truncated or oblit¬
erated by valley downcutting and erosion of the
ancient Pecos River Valley (Hill, 2000).
Flowing water is relatively rare in caves of the
Guadalupe Mountains. There is only one area in
Lechuguilla Cave that has flowing water (Kambesis,
1989). This is the Lost Pecos River in the Far East
Branch of the cave. A small stream issues from a crack
ENCYCLOPEDIA OF CAVES
LECHUGUILLA CAVE, NEW MEXICO, U.S.A.
459
1400
1300
CD 1200
T3
< 1100
1000
Tower Place
, lower Boulder Falls
1 Underground Atlanta
Chandelier Graveyard
| Glacier Bay
Lebarge * Rift top
Prickly Ice Cube Rm.g . Mouse , s De|ight
■ Chandelier Ballroom i
Hoodoo Hall '
main Far East
Western Borehole
" -□
Void Nirvana 1 ° utback
I lower Far East
Rooms or distinct
passage levels
Ascending phreatic
passages
FIGURE 3 Elevation of rooms and passages in
Lechuguilla Cave. The width of each block is propor¬
tional to the width of the room or passage. After Palmer
and Palmer, 2000.
in the wall and feeds a series of pools in the area. This
water probably originates from a small perched aqui¬
fer that the cave passage has intersected.
Drip and pool water is fairly common throughout
the cave and is sourced from vadose meteoric water
that has seeped down from the surface along joints,
bedding planes and interconnected pores in the
rock. The geochemistry of Lechuguilla Cave's pool
waters is quite diverse. The water chemistry is a func¬
tion of precipitation chemistry, bedrock, and the
gypsum deposits that occur in close proximity (Turin
and Plummer, 2000). The water is further modified by
evaporation, speleothem deposition, and microbial
activity. Recent research indicates that the water
levels in cave pools respond to regional climate change
rather than just random drip migration (Despain and
Stock, 2008).
BIOLOGY
The megascopic biology of Lechuguilla Cave is
somewhat limited. Cave crickets and arachnids are
present in the entrance passages and to some extent
past Boulder Falls, the first significant pit in the cave.
Bat bones have been found throughout the system but
no bats currently inhabit the cave, except at the
entrance.
However, the microscopic biology is quite a differ¬
ent story. It is speculated that since the microbial sys¬
tems within Lechuguilla Cave have no multicellular
grazers or predators, and no meteoric influences, and
have been in existence for long spans of time, the evo¬
lution of these microbial communities are indepen¬
dent of surface organisms (Boston, 1999). Kiym
Cunningham was the first to recognize microbial evi¬
dence in corrosion residues and in some speleothems.
Microbial research in the cave indicates that the
microbe communities are chemolithotrophs. These
types of bacteria metabolize sulfur, iron, and manga¬
nese minerals and may assist in enlarging the cave
and determining the shapes of some unusual spe¬
leothems. Other studies indicate that some microbes
may possess medicinal qualities that are beneficial to
humans.
Lechuguilla Cave contains remnants of what could
be past and perhaps present geomicrobiological
interactions. These features include corrosion resi¬
dues, moonmilk, pool fingers, webulites, and u-loops.
Microorganisms of today have been found in asso¬
ciation with carbonate and silicate speleothems, sul¬
fur compounds, and iron and manganese oxides
(Cunningham et al., 1995).
Research on microbial ecology has revealed that the
microorganisms in Lechuguilla Cave, though surviving
in nutrient-poor conditions, are metabolically versatile
and unexpectedly diverse. Preliminary research results
indicate that microbial communities overcome the lim¬
itations of a nutrient-poor environment by cooperative
and mutualistic associations (Barton and Jurado, 2007).
MINERALOGY
Lechuguilla Cave is well known for its unusual min¬
eralogy and for the spectacular nature of its spe¬
leothems. Almost of the caves of the Guadalupe
Mountains display many of the same features as
Lechuguilla. However, the Lech displays almost fea¬
tures as large-scale, text book examples. Lechuguilla
Cave contains some of the most spectacular examples
and varieties of gypsum speleothems including massive
selenite chandeliers, delicate clear gypsum crystals, gyp¬
sum needles, gypsum cotton, and crusts. Calcite spe¬
leothems, though not in as great abundance, are also
spectacular in their shape and configuration and occur
in places throughout the cave where meteoric waters
seep into the underground. Large stalagmites have
formed in Tower Place. Stunning arrays of soda straws,
stalactites, and draperies festoon such places as Deep
Secrets, the Oasis Pool, and the Pearlsian Gulf.
ENCYCLOPEDIA OF CAVES
460
L
Large amounts of massive or laminated gypsum are
common within the cave. These deposits accumulated
where groundwater flow was not strong enough to
carry away the sulfate produced by the sulfuric acid
reaction with the limestone when the cave was enlarg¬
ing (Davis, 2000). Some of the gypsum beds have been
dissolved away by meteoric water that percolates into
the cave or from freshwater reflooding. Many of these
gypsum remnants occur in dry sections of the cave
and in places resemble small glaciers (Glacier Bay,
Prickly Ice Cube Room, Land of Fire and Ice, Blanca
Navidad).
Elemental sulfur deposits occur as yellow masses
or as stringers within larger masses of gypsum.
The sulfur, which can be massive, granular, or
platy, is exposed by dissolution of the surrounding
gypsum (Davis, 2000). Locations where sulfur depos¬
its have been observed include the Void, Ghosttown,
Ghostbusters Balcony, and near Blanca Navidad.
Caves formed by sulfuric acid-bearing waters con¬
tain direct by-products of their origin including gyp¬
sum, hydrated halloysite, and alunite. Lechuguilla
contains many of these minerals. The alunite is
important because it contains potassium which gives
it the potential to be dated by the K-Ar or 40 Ar/ 39 Ar
dating methods (Polyak and Provencio, 2000). Polyak
et al. (1998) used 40 Ar/ 39 Ar dated alunite to deter¬
mine that Lechuguilla Cave is at least 5 million years
old.
Lechuguilla Cave is rich in corrosion residues that
contain bacterial and fungal communities, indicating
that microbes could be active participants in limestone
dissolution in conjunction with abiological processes
(Cunningham et al., 1995). Gypsum and sulfur in the
Guadalupe Mountains are significantly enriched in the
light isotope of sulfur (Hill, 2000). The enrichment of
light sulfur implies that hydrocarbons and sulfur bac¬
teria were involved in speleogenesis.
Endelleite and montmorillonite fill solution pockets
in many places in the cave. The minerals celestite and
barite have also been found in many of the maze areas
of the cave (Chandelier Maze).
Some speleothems have a biologic origin. Rusticles,
which are observed in only two locations with in the
cave, are composed on iron oxide minerals which con¬
tain fossil microbial casts (Davis, 2000). Pool fingers,
which grew underwater, are calcite in composition but
seem to have a microbial origin.
One of the "signature" speleothems of Lechuguilla
Cave are underwater helictites. These are vermiform
in shape and contain a central canal. These forma¬
tions developed as a result of the common-ion effect
(Davis, 2000). Other unusual mineralogies include
folia, hydromagnesite balloons and fronds, raft accu¬
mulations and cones, splash rings, and silticles.
Though the wide variety of mineralogies and spe¬
leothems seem randomly diverse, they are related to
each other by the unusual origin of the cave and/or
from convective airflow loops that are driven by tem¬
perature gradients that result from the cave's great
vertical relief (Davis, 2000).
PALEONTOLOGY
Vertebrate paleontologic remains are very rare.
Pleistocene fossil remains have been found in fills near
the entrance area. It is speculated that the fill is from
an ancient sinkhole that may have trapped some local
mammals. There are bat remains scattered throughout
the system, some of them of paleontologic significance.
There is only one other locality within Lechuguilla
Cave that contains vertebrate fossil remains. A fully
articulated ringtail cat skeleton was found in the Land
of Awes, an area that is far removed from the current
entrance and the surface. This specimen indicates the
possibility of an ancient entrance that is no longer
accessible.
Lechuguilla Cave has formed within the Capitan
Reef complex. The primary frame-building organisms
of the reef include calcareous sponges, calcareous
algae, and bryozoans. These fossils are very common
within the bedrock of Lechuguilla Cave. Many of the
fossils found within the cave have been broken into
sandstone size fragments that accumulated in voids in
the Capitan Reef and in the backreef lagoon (DuChene,
2000). Fossils are also exposed in the bedrock walls
that are not covered with mineral crusts and coatings.
The most abundant fossil remains are those of inverte¬
brate reef organisms including bivalves, brachiopods,
bryozoans, crinoids, porifera, gastropods, urchin spines,
corals, and cephlapods. These occur throughout all
areas of the cave.
DESCRIPTION OF THE CAVE
The overall pattern of the chambers and passages of
Lechuguilla Cave is characteristic of caves formed
by sulfuric acid speleogenesis. These types of caves
are characterized by ramiform passage morphology,
large passage size, and horizontal passages connected
by deep shafts and fissures. Irregular passage shapes
are common, as are dead-end rooms and galleries
and blind alcoves. The ramified shape and layout
of Lechuguilla Cave (Fig. 4) are formed by the oxidation
of rising H 2 S, where the main passages and large
domed rooms correspond to favored paths of H 2 S
movement (Palmer, 1996). The large passage sizes are a
ENCYCLOPEDIA OF CAVES
LECHUGUILLA CAVE, NEW MEXICO, U.S.A.
461
Lechuguilla Cave
Eddy County, New Mexico
Surveyed Length as of 2002: 172 km
Vertical Extent: 465 m
Entrance Series
Plan View
Windy
Eastern Branch
Southwestern Branch
Entrance, , —•> Wooden Lettuce
Profile View
Far East
FIGURE 4 Map of Lechuguilla Cave.
function either of large H 2 S inputs or H 2 S inputs over
longer periods of time.
Lechuguilla Cave contains a number of fissures (or
rifts) that descend from the floors of large rooms or pas¬
sages and pinch downward with bottoms that are
clogged with calcite crust, carbonate sediment, or that
terminate in water (Fig. 5). These rifts extend below the
level of major cave rooms and may be inflow routes
along which oxygenated and H 2 S-rich water first began
to mix (Palmer and Palmer, 2000). Such areas include the
routes from the Chandelier Graveyard to Sulfur Shores
and from the Outback to Lake of the White Roses.
Different levels within Lechuguilla Cave are con¬
nected via ascending passages such as the Great White
Way, which connects the south Rift to Deep Secrets, and
the Gulf of California fissure, which connects the mazes
of Snow White's Passage to the LeBarge Borehole.
Ascending tubes with large cross-sections and great
vertical extent may also be the result of the mixing of
two major water sources (groundwater and meteoric
water). Such an ascending tube rises 200 meters from
the Rift to the cave's entrance (Palmer and Palmer,
2000). The Wooden Lettuce Passage (whose highest
point is located approximately 6 meters above the cur¬
rent entrance) may have been a paleo-entrance.
Horizontal passages may represent former water
table levels (Fig. 6). The Western Borehole is a 1.6 km-
long linear trend of horizontal passages with areas of
elliptical passage cross-section, which may have served
as an outlet for water from zones of maximum cave ori¬
gin (Palmer and Palmer, 2000). The Western Borehole
ends abruptly without breakdown collapse or major
passage extensions.
The walls of many of the large passages in
Lechuguilla Cave are honeycombed with intercon¬
nected, nontubular solution cavities of various sizes and
geometry forming three-dimensional mazes of sponge-
work-type morphology (Fig. 7). These passages, called
Boneyard, are ubiquitous throughout Lechuguilla Cave
and in most other caves in the Guadalupe Mountains.
They develop where many interconnected openings are
solutionally enlarged at comparable rates. Mazes
ENCYCLOPEDIA OF CAVES
462
FIGURE 5 Fissure descending into the LeBarge Borehole,
Southwestern Branch. Photo courtesy of Dave Bunnel. Used with
permission.
originate in mixing zones where the aggressiveness of
the water is high.
HISTORY OF EXPLORATION
Prior to May 1986, the known extent of Lechuguilla
Cave consisted of a 27 meter entrance pit that dropped
FIGURE 6 Large passage of the Great Western Borehole, Western
Branch. Photo by Dave Bunnel. Used with permission.
FIGURE 7 Boneyard maze. Photo by Dave Bunnel. Used with
permission.
into a large three-level chamber containing several
insignificant side passages. Surveyed passage length
was 120 meters. Covering the floor was a thick veneer
of guano which coated a large rubble pile. Of note was
the fact that a tremendous volume of air issued from
the rubble pile, sometimes strong enough to blow sedi¬
ment up the entrance pit. Later, other visitors to the
cave suggested it be named Misery Hole because of
the great amount of dust that blew out of the entrance
pit.
Lechuguilla Cave had been known at least since
the early part of the twentieth century. The name and
date of the cave's original discoverer is unknown but
it is possible that guano miners may have made the
first entrance into the cave. In 1914, John Ogle, Cad
Ogle, and C. Whitfield filed a placer mining claim to
mine bat guano from the cave. At that time guano
mining was a major industry in the Guadalupe
Mountains. Only a small amount of guano was actu¬
ally removed from the cave and it appears that the
project was abandoned after only preliminary mining
efforts (Frank, 1988). After the initial mining activi¬
ties, the cave saw a minimal amount of activity. Park
ENCYCLOPEDIA OF CAVES
LECHUGUILLA CAVE, NEW MEXICO, U.S.A.
463
rangers visited the cave in the early 1940s, and in the
early 1950s Colorado Grotto members mapped the
cave and published a descriptive report.
In the 1970s Cave Research Foundation (CRF) and
park personnel made several trips to the cave. CRF
remapped the cave and the map was published in
D.J. Jagnow's Cavern Development in the Guadalupe
Mountains, in January of 1979. CRF digging crews
began working the lowest levels of the blowing rub¬
ble pile, having to shore it up as they went. Progress
on the unstable rock pile was minimal and CRF aban¬
doned the dig project.
In 1984 digging activities led by cavers from
Colorado took a new approach on the breakdown pile
using the wind and bedrock walls as their guide. Over
the course of four digging expeditions, significant
progress was made on the rubble pile. On May 25,
1986 the dig broke out on the side of a talus slope into
a low, wide room. Continuing through a squeeze and
following the wind, the diggers/explorers broke into a
large-going passage.
This marked the beginning of one of the greatest
North American cave discoveries of the twentieth
century. Between 1986 and 1998, over 160 km of cave
passages were discovered, explored, and documented.
After the breakthrough, exploration moved at
breakneck speed, with 5—8 km of passages documen¬
ted with each expedition. The first 48 months of
exploration and survey work delineated the main
framework of the cave and established its vertical
extent. Explorers pushed and mapped the east—west
trending entrance passages to a 43-meter pit named
Boulder Falls. The cave continued at the base of the
pit, through a series of large, gypsum-floored corri¬
dors and rooms (Colorado Room, Windy City, Glacier
Bay, Sugarlands). The east—west trend terminated at
the Rift, a large fissure that has a vertical extent of
over 30 meters in places. At the south end of the Rift
the cave splits into three separate east—west trending
branches.
Explorers pushed a small hole in the rift which
leads to a tall, slanting gypsum fissure named the
Great White Way. Deep Secrets was the first signifi¬
cant room at the base of the fissure and it contained
not only an abundance of gypsum coatings and
assorted calcite speleothems but also some of the lon¬
gest soda straws ever observed in the Guadalupe
Mountains. A series of mazey rooms and climb-ups
eventually broke into the Great Western Borehole, a
1.6 km-long, gypsum-coated trunk passage which
branched to two terminations. The southern terminus
ended at the calcite-encrusted Oasis Pool located at
the base of a sheer wall. The northern terminus was
pushed to the Rainbow Room, an ascending bedding
plane passage. In other places in the Western
Borehole, lofty avens led to upper-level segments
including the Chandelier Graveyard and Pelucidar
and Barsoom.
A major junction in the Rift (the EF Junction) led to
more gypsum-coated mazes and yet another descend¬
ing, gypsum-coated fissure. This proved to be the
gateway to the Southwest Branch, one of the most
spectacularly decorated sections of the cave.
Exploration and survey documented lakes and
pools in areas containing flowstone, dripstone, and
gypsum formations (Lake LeBarge, Lake Chandelar,
Yellow Brick Road, Lake Castrovalva). Long, tubular
passages and convoluted boneyard mazes coated
with all manner of gypsum mineralogy were discov¬
ered and mapped (LeBarge Borehole, Land of Awes,
Chandelier Ballroom, Prickley Ice Cube Room,
Chandelier Maze, Hoodoo Hall).
Passages trending to the southwest of the Chandelier
Maze opened into more boneyard mazes (Land of the
Lost), large austere chambers floored with breakdown
(the Voids), high-vaulted rooms floored with pools,
flowstone, and cave pearls (Big Sky and Pearlsian Gulf).
Thick deposits of rock flour coat the floors of many of
the rooms and passages. Another steeply descending
fissure leads to Sulfur Shores and what may be the top
of the water table. High-level fissures trending north
from the Prickly Ice Cube Room lead to large shafts,
upper-level rooms containing some of the largest flow¬
stone formations in the cave (Underground Atlanta and
Tower Place), and more perched lakes and boneyard
mazes (High Hopes).
At the far end of the Rift is a pit leading to a series
of ramps and fissures that are coated with dark corro¬
sion residues that drop into the Eastern Branch of
Lechuguilla Cave. This branch of the cave proved to
be the most technically challenging, requiring the rig¬
ging of multiple pitches and traverses. More perched
lakes, and large-scale mazes and rooms characterized
this branch of the system (Lake of the Blue Giants,
Nirvana, Megamaze, Moby Dick Room). Explorers
were temporarily stopped by a tall, aragonite-coated
fissure named the Aragonitemare. A series of climbing
attempts finally reached the top of the fissure (which
was measured at 60 meters) and led to a continuation
of the Eastern Branch called the Far East Maze. Cavers
documented east-trending maze passages, some with
corrosion residues or massive aragonite trees covering
the floors and other areas with perched pools coated
with calcite rafts (Grand Guadalupe Junction, Silver
Bullet Passage, Land of Enchantment, Boundary
Waters). At the Grand Guadalupe Junction, upward-
trending passages were explored in a series of complex
upper mazes (La Morada Maze, Firefall Hall). Pits
in the floor of Grand Guadalupe Junction dropped
into the Outback, a sparsely decorated section of
ENCYCLOPEDIA OF CAVES
464
L
lower-level cave defined by more slanting fissures and
long, low-pancake crawlways. In the middle of the
Outback, the bottom of the passage literally dropped
out to the deepest point of the cave known as Lake of
the White Roses at 486 meters, which may also be the
water table. By 1990, more than 80 km of passages had
been discovered, explored, and documented.
Between 1991 and 1998 another more than 80 km of
discoveries had been documented. The slowdown in
pace of exploration was a function of the remoteness
of many of the leads, the necessity of technical climb¬
ing in some areas (Davis, 1999), and because the NPS
had, at various times, closed the cave to exploration
trips for management purposes.
Since 1999, exploration and survey work has contin¬
ued to fill in more maze passages that surround all
major branches of the cave. New discoveries identified
significant northern and southern extensions in the
Western Branch of the cave. From the central Western
Borehole, a technical climb led to the discovery of
another one of the cave's major chambers, Hudson Bay.
North west-trending passages led to the northernmost
extent of Lechuguilla Cave. At the known limits of the
Western Borehole, explorers extended the perimeters of
the cave via breakouts to the Wild West, South Winds,
the Sanctuary, and Emerald City. This section of the
cave contains passages that head back toward the High
Hopes block, suggesting a possible connection between
the Southwest and Western branches of the cave. To
date, that connection has not been made.
In 2005, a major extension named the Coral Sea was
discovered off the eastern edge of the Outback in the
Far East Branch of the cave. This section of cave has
seen close to three kilometers of survey and contains a
high density and variety of calcite and aragonite spe-
leothems, numerous occurrences of active flowstone
and dripstone, and over 100 lakes and pools. These
features are in striking contrast to the Outback which,
overall, is dry and speleothem-poor. The Coral Sea is
the easternmost extension of Lechuguilla Cave.
The new explorations in the Southwest Branch
of the cave have concentrated on pursuing climbing
leads and upper-level passages, specifically in the
Chandelier Maze. Though no significant passage
extensions have been discovered as of yet, the various
climbs have led to short sections of upper levels, some
that connect back to known areas of the complex
Chandelier Maze.
In addition to new exploration, significant efforts
have been made in resolving survey issues and raising
the overall quality of the survey work in Lechuguilla
Cave. Loop closures have been improved, survey
errors identified and resolved, and in-cave sketches
upgraded to today's standards. The Cave Resource
Office (CRO) at Carlsbad Caverns National Park
oversees the cartographic effort in Lechuguilla Cave
and a team of volunteer cartographers work on the
series of quadrangle maps for the cave system.
As of 2010, Lechuguilla Cave has a surveyed length
of 209 km. There is still good potential for more exten¬
sions in all sections and levels of the cave. Access to
Lechuguilla Cave is limited to approved scientific
researchers, survey and exploration teams, and National
Park Service management-related trips.
See Also the Following Article
Sulfuric Acid Caves
Bibliography
Barton, H., & Jurado, V. (2007). What's up down there? Microbial
diversity in caves. Microbe, 2(3), 132—138.
Boston, P. (1999). A bit of peace and quiet: The microbes of
Lechuguilla. NSS News, 57(8), 237-238.
Cunningham, K., Northup, D., Pollastro, R. M., Wright, R., &
LaRock, E. (1995). Bacteria, fungi and biokarst in Lechuguilla
Cave, Carlsbad Caverns National Park, New Mexico. Environ¬
mental Geology, 25(1), 2—8.
Davis, D. G. (1999). Lechuguilla Cave: The second 50 miles. NSS
News, 57(8), 228-238.
Davis, D. G. (2000). Extraordinary features of Lechuguilla Cave,
Guadalupe Mountains, New Mexico. Journal of Cave and Karst
Studies, 62(2), 147-157.
Despain, ]., & Stock, G. (2008). Dated cave pool shelfstones as indica¬
tors of climate change. Lechuguilla Cave NewsletterO), 38—39
(Winter 2007-2008).
DuChene, H. R. (2000). Bedrock features of Lechuguilla Cave,
Guadalupe Mountains, New Mexico. Journal of Cave and Karst
Studies, 62(2), 109-119.
DuChene, H. R., & Hill, C. A. (2000). Introduction to the caves of the
Guadalupe Mountains Symposium. Journal of Cave and Karst
Studies, 62(2), 53.
Frank, E. F. (1988). Cave history—Lechuguilla Cave. NSS News,
46(11), 378-379.
Hill, C. A. (1996). Geology of the Delaware Basin, Guadalupe, Apache,
and Glass Mountains, New Mexico and West Texas, Publication
96—39. Midland, TX: Society of Economic Paleontologists and
Mineralogists—Permian Basin Section.
Hill, C. A. (2000). Sulfuric acid, hypogene karst in the Guadalupe
Mountains of New Mexico and West Texas, U.S.A. In A. B.
Klimchouk, D. C. Ford, A. N. Palmer & W. Dreybrodt (Eds.),
Speleogenesis, evolution of karst aquifers (pp. 309—316). Huntsville,
AL: National Speleological Society.
Jagnow, D. H. (1977). Geologic factors influencing speleogenesis in the
Capitan Reef Complex, New Mexico and Texas. Unpublished M.S.
Thesis, University of New Mexico, Albuquerque.
Jagnow, D. H. (1979). Caverns development in the Guadalupe
Mountains. Columbus, OH: Cave Research Foundation.
Jagnow, D. H. (1989). The geology of Lechuguilla Cave, New Mexico.
In P. M. Harris & G. A. Grover (Eds.), Subsurface and outcrop exami¬
nation of the Capitan Shelf Margin, Northern Delaware Basin
(pp. 459—466). San Antonio, TX: Society of Economic Paleontologists
and Mineralogists (Core Workshop 13).
Kambesis, P. (1989). Extending the limits, a major breakthrough in
Lechuguilla Cave. NSS News, 47, 300—305.
ENCYCLOPEDIA OF CAVES
LIFE HISTORY EVOLUTION
465
Northup, D., Dahm, C., Melim, L., Spilde, M., Crossey, L., Lavoi, K.,
et al. (2000). Evidence for geomicrobiological interaction in
Guadalupe Caves. Journal of Cave and Karst Studies, 62(2), 80—89.
Palmer, A. (1996). Cave patterns in the Guadalupe Mountains (pp.
51—52). National Speleological Society Convention, Salida,
Program with abstracts.
Palmer, A., & Palmer, M. (2000). Hydrochemical interpretation of
cave patterns in the Guadalupe Mountains, New Mexico. Journal
of Cave and Karst Studies, 62(2), 91—108.
Polyak, V. }., McIntosh, W. C., Guven, N., & Provencio, P. (1998).
Age and origin of Carlsbad Cavern and related caves from
40 Ar/ 39 Ar of alunite. Science, 279(5358), 1919-1922.
Polyak, V. J., & Provencio, P. (2000). Summary of the timing of sulfu¬
ric-acid speleogenesis for Guadalupe Caves based on the ages of
alunite. Journal of Cave and Karst Studies, 62(2), 72— 74.
Turin, J., & Plummer, M. (2000). Lechuguilla Cave pool chemistry,
1986—1999. Journal of Cave and Karst Studies, 62(2), 135—143.
LIFE HISTORY EVOLUTION
David C. Culver
American University
INTRODUCTION
The concept that life history characteristics, for exam¬
ple, age at first reproduction, number of offspring, and
longevity, were not just a passive reflection of what
the environment would allow, but rather that they
were molded by natural selection gained great cur¬
rency in the 1960s and 1970s. Since then a variety of
models of life history evolution have been put forward
to explain the multifaceted differences in life history
characteristics of species. The subterranean environ¬
ment in general and the cave environment in particu¬
lar have characteristics that act as strong selective
agents in the evolution of life histories. Foremost
among these are extreme food scarcity and the reduc¬
tion in seasonal cues for reproduction. Nearly all cave
animals are characterized by increased longevity,
greatly diminished reproductive rates, and increased
investment in the offspring that are produced, often in
the form of increased egg size (Culver, 1982).
There is perhaps no aspect of the biology of cave ani¬
mals that is both more interesting and more difficult to
study. Many undergraduate and graduate students have
set out to study life history and population growth of
cave animals, but given longevities that often measure
in the decades and reproductive rates that are exception¬
ally low, very little change is often observed in popula¬
tions over the course of a normal research project. This
is further complicated in many cases by the inability to
sample the subterranean habitat in times of high water,
often the critical time in demography of subterranean
populations. What we know about the evolution of life
history of cave animals is based on a few long-term
studies with bits and pieces of information from many
shorter term studies (Trajano, 2001).
LIFE HISTORY EVOLUTION OF
AMBLYOPSID FISH
The fish in the family Amblyopsidae are a small
lineage of fish comprising six species, four of which are
stygobionts in North American caves and one of which
is a stygophile in North American caves (Niemiller and
Poulson, 2010). The extensive work of Thomas Poulson
(1963) on life history characteristics of these species is
both the most extensive comparative study of cave and
non-cave species in the same lineage and the most
influential study of any cave organisms. Poulson was
the first cave biologist to look at the life history of cave
animals as an adaptation to the cave environment.
One species, Chologaster cornuta, lives in freshwater
swamps in the Coastal Plain of the southeastern
United States. The other species associated with caves,
the stygophilic Forbesichthys agassizi, is found in
springs and caves in the central United States, espe¬
cially Kentucky. The other four species are all obligate
cave dwellers. The genus Amblyopsis has two species,
A. rosae in the Ozarks of Arkansas, Kansas, Missouri,
and Oklahoma, and A. spelaea in the Interior Low
Plateaus of Indiana and Kentucky. Typhlichthys subtcr-
mneus is found in a wide area from Missouri to
Tennessee. Speoplntyrhinus poulsoni is one of the rarest
vertebrates in North America, known from only a sin¬
gle cave in Alabama. Little is known about its life his¬
tory and it will not be considered further.
When the stygobiotic species are compared to the
other two species, they differ in a set of characteristics
which represent a reduction in reproductive effort on
the part of the cave species (Table 1). Compared to
Chologaster and Forbesichthys, all the obligate cave¬
dwelling species show at least a 20% reduction in the
number of eggs, at least a 100% increase in the age of
first reproduction, at least a 50% reduction in the pro¬
portion of the population breeding at any given time,
and at least a 40% reduction in growth rate. Poulson
(1963) argues that all of these characteristics are an
adaptation to a food-poor environment, and there
is little doubt that he is correct in this assertion. It is
important to note that all of the characteristics listed in
Table 1 could also be the result of food stress on a pop¬
ulation. The response of almost any population to food
stress is an increase in the age at first reproduction, a
reduction in the number of eggs, a reduction in the
number of breeding females, and a reduction in growth
rate. In a sense, these are imposed by the environment.
However, in the case of cavefish, there is strong evi¬
dence that this is not just an environmental response.
Encyclopedia of Caves.
© 2012 Elsevier Inc. All rights reserved.
466
L
TABLE 1 Life History Characteristics of Amblyopsid Fish That Are Consistent with a
a Food-Poor Environment
Response to Starvation, as
Well as Adaptation to
Species
Age at First Reproduction
(months)
Number of Eggs
Maximum Proportion
of Females with Eggs
Growth Rate
(mm per year)
Cholcigaster cornuta
12
93
1.00
2.4-3.8
Forbesichthys agassizi
12
150
1.00
1.7—2.2
Typln/lichthys subterraneus
24
50
0.50
1.0
Amblyopsis spelaea
40
70
0.10
1.0
Amblyopsis rosae
37
23
0.20
0.9
TABLE 2 Life History Characteristics of Amblyopsid Fish That Are an Adaptation to
a Food-Poor Environment
Species Longevity
(years)
Egg Diameter
(mm)
Average
No. of
Broods
Maximum
No. of Broods
Reproductive Effort
per Brood
(mm 3 gm _1 of female)
Maximum Lifetime
Reproductive Effort
(mm 3 gm _1 of female)
Chologaster cornuta 1.3
0.9—1.2
1.0
1
64
64
Forbesichthys agassizi 2.3
1.5-2.0
2.0
2
148
297
Typhylichthys 4.2
subterraneus
2.0—2.3
1.5
3
452
903
Amblyopsis spelaea 7
2.0—2.3
0.5
5
52
260
Amblyopsis rosae 4.8
1.9—2.2
0.6
3
83
249
but rather it is a genetically programmed response to
the cave environment which is the result of selection
for a set of life history characteristics.
The strongest evidence for the adaptive nature of
cavefish life histories comes from an additional set
of life history characteristics that cannot be the simple
result of the environmental pressure of starvation
(Table 2). These include at least a doubling of life
span, at least a 40% increase in egg size, and at least a
50% increase in the maximum number of broods as
a result of increased longevity. One stygobiotic species,
T. subterraneus, shows an increase in reproductive
effort per brood as measured by total egg volume per
gram of female. All stygobiotic species show an
increase in maximum lifetime reproductive effort com¬
pared to the swamp-dwelling C. cornuta and the same
or greater effort than the stygophilic F. agassizi. The
maintenance and even a possible increase in lifetime
reproductive effort is somewhat surprising, and may
not simply be the result of increased longevity. This
apparent anomaly will be addressed in a later section.
Differences in life history characteristics among the
stygobiotic fish have two general explanations. One is
that the species have had different amounts of evolu¬
tionary time in caves. Thus, the more "extreme" life
histories of A. rosae and A. spelaea compared to T. sub¬
terraneus may be because the Amblyopsis species have
been isolated in caves longer. There is no direct evi¬
dence for this and a second explanation seems more
likely: the different species occupy different subterra¬
nean habitats which result in different selective pres¬
sures. For example, A. rosae occurs in areas near
the surface while T. subterraneus occurs deeper in the
groundwater.
GENERALITY OF LIFE HISTORY
CHARACTERISTICS
While there is no other comparative study within a
lineage of both cave- and surface-dwelling species that
is as complete as that of the amblyopsid fish, available
evidence from other groups indicates that the patterns
observed in amblyopsid fish are widespread. Only a
few cases will be mentioned here.
Longevity in cave animals, at least aquatic species,
seems to be greatly increased relative to surface species.
The French cave-dwelling amphipod NipJwrgus virei has
a maximum age of 35 years and routinely reaches the
age of 10 years, based on annual growth rates (Turquin
and Barthelemy, 1985). There are several reports of
cavefish living for more than ten years in a laboratory
setting, including Pimelodella kronei (Brazil), Phreatichthys
andruzzi (Somalia), and Caecobarbus geertsii (Congo).
ENCYCLOPEDIA OF CAVES
LIFE HISTORY EVOLUTION
467
Cave springtails (Collembola) often have double the life
span of soil- and litter-dwelling species, and harpacti-
coid crustaceans living in the interstices of gravel aqui¬
fers also have double the life span of surface-dwelling
species. In these and many other cases of reported lon¬
gevity, it is impossible to determine if increased longev¬
ity was a direct result of selection in the subterranean
environment or whether only those lineages with great
longevity can successfully invade caves. The reason
these causes cannot be separated is that it is often
impossible to find a closely related surface-dwelling
ancestor to compare to the subterranean species and
thus lineage and selection effects are confounded.
The epitome of longevity in cave animals is from
the study of the crayfish Orconedes australis australis in
Shelta Cave, Alabama, by John Cooper (Cooper, 1975).
Based on growth rates of marked individuals over a
period of several years, the minimum age at sexual
maturity was 15 years, and the best estimate, based on
average growth rates rather than maximum growth
rates, was 35 years. Furthermore, this was the esti¬
mated age of sexual maturity. The estimated age at
first reproduction was over 100 years. Since Cooper's
study was completed, the population of crayfish in
Shelta Cave crashed to less than 10% of its former
size. This was probably the result of an ill-advised
gate to the cave that kept out bats, which probably
were an important food supply to the crayfish in the
form of guano or the organisms that fed on guano.
Nevertheless, it is an impressive example of the ability
of cave animals to survive for extended periods of
time.
Increased egg size and reduced numbers of eggs is
widely reported among subterranean animals, includ¬
ing harpacticoid copepods living in interstitial habitats
and cave amphipods. One especially clear-cut example
without the confounding effects of lineage differences
is the amphipod Gammarus minus. Populations of this
species (actually a species complex) have invaded sub¬
terranean drainages in extensive karst areas in Virginia
and West Virginia. The average egg number in a cave
population was 50% less than in a spring population
less than 5 km away, for females of the same size. In
contrast, the volume of an individual egg is about 40%
larger for the cave population. Many other subterra¬
nean animals have very small numbers of eggs, but
usually comparison with the appropriate surface spe¬
cies is unavailable.
One remarkable life history adaptation to cave life
is shown by some leiodid beetles in the genus Isereus
(Ginet and Decou, 1977). Two species, I. colasi and
I. serullazi, have a nonfeeding larval stage that lasts
nearly 100 days followed by a pupal stage lasting more
than 115 days. The evolutionary advantage of the sup¬
pression of larval feeding is not entirely clear, but it is
probably related to escaping predation, a factor whose
importance is suggested by the length of the pupal
period as well. If prey available to the prey are scarce,
the loss from lack of feeding might be quite small.
One feature of the life history of subterranean
organisms shows a great deal of variability, the season¬
ality or lack of seasonality of reproduction. For preda¬
tors such as the carabid beetle, Neaphaenops tellkampfi,
which feeds on the cricket eggs, reproduction is timed
to the availability of cricket eggs. Because the crickets
forage in surface habitats at night, their reproduction
is seasonal. For most subterranean species, seasonality
is much less clear-cut. It is a common (and frustrating)
observation that a cave population that shows a clear
seasonality one year fails to show the same pattern the
next year. Work by the French biologist Marie-Jose
Turquin and colleagues (Mathieu and Turquin, 1992;
Turquin and Barthelemy, 1985) provides the basis for
this variation. She points out that species living in cave
streams are subject to floods year round. For the popu¬
lations of the amphipod Niphargus that she studied in
southern France, flood frequency reached a maximum
in the summer. The impact of these floods was large: it
caused significant mortality and it brought in signifi¬
cant amounts of food. Both mortality and reproduction
were keyed to these events, which of course do not
occur at the same time or intensity each year. Other
populations that live in cave pools or live in the inter¬
stices of gravels in streams are not faced with the same
hydrological regime. In these environments, floods do
not occur. Instead their habitat can shrink due to dry¬
ing. In these populations, ovigerous females and egg
laying follow several months after the onset of the
wet season. Of course, terrestrial populations would be
subject to different temporal patterns, but Turquin's
point that reproduction is keyed to events of high mor¬
tality and / or food input is likely to be a general one.
THE PARADOX OF HIGH
REPRODUCTIVE EFFORT
IN CAVE ANIMALS
Except for Turquin, few investigators have noted
what seems to be a paradoxical aspect of the life his¬
tory of stygobionts and troglobionts. Even though the
number of eggs produced in any one brood is small
and the possibility of enough resources to allow for
reproduction is small, the potential for reproduction of
cave animals is often quite large (Culver and Pipan,
2009). This is even the case for the amblyopsid cave-
fish. The reproductive effort per brood is highest in the
stygobiont T. subterraneus, and the lifetime maximum
reproductive effort of the other two stygobionts,
A. rosae and A. spelaea, is much greater than that of the
ENCYCLOPEDIA OF CAVES
468
L
swamp-dwelling C. cornuta and nearly that of the
spring-dwelling C. agassizi. How can we account for
this paradox of long life span and a potentially high
reproductive rate even in the amblyopsid cavefish,
which are so exquisitely adapted to subterranean life?
Turquin suggests that this paradox is in fact essential
to life in caves. Although there are situations where
there is a low but constant source of food such as pro¬
duction by chemoautotrophic bacteria, the more usual
situation in caves is that food comes into the subterra¬
nean realm in pulses or spurts, which may be more or
less predictable. Turquin suggests that the combination
of a life span and the ability to expend significant
reproductive effort make life in food-poor subterra¬
nean environments possible. If this idea is indeed cor¬
rect, then it goes toward explaining some otherwise
perplexing features of the life history of subterranean
organisms. The high variability of reproduction
between years makes sense because the animals are
responding to opportunities for reproduction when
there is food input that is not necessarily
predictable from year to year. The increased longevity
of cave animals allows them to "wait" for opportu¬
nities. It also fits in nicely with the idea that for most
of the time, the growth rate (r of the standard growth
equations) is slightly negative for cave animals. This
was certainly the case for O. australis australis in Shelta
Cave and was likely true for many of the fish popula¬
tions studied by Poulson. Occasionally, growth rates
must be positive or these populations would become
extinct. The long life span and relatively high potential
investment in reproduction allows stygobionts and
troglobionts to take advantage of these opportunities
for growth.
See Also the Following Articles
Natural Selection
Adaptation to Low Food
Bibliography
Cooper, J. E. (1975). Ecological and behavioral studies in Shelta Cave,
Alabama, with emphasis on decapod crustaceans. Lexington, KY:
University of Kentucky.
Culver, D. C. (1982). Cave life. Evolution and ecology. Cambridge, MA:
Harvard University Press.
Culver, D. C., & Pipan, T. (2009). The biology of caves and other subter¬
ranean habitats. Oxford, U.K.: Oxford University Press.
Ginet, R., & Decou, V. (1977). Initiation a la biologie et a Ve'cologie sou-
terraines. Paris: J.-P. Delarge.
Mathieu, J., & Turquin, M. J. (1992). Biological processes at the popu¬
lation level, n. Aquatic populations: Niphargus (stygobiont
amphipod) case. In A. I. Camacho (Ed.), The natural history of bio¬
speleology (pp. 263—293). Madrid: Monografias, Museo Nacional
de Ciencias Naturales.
Niemiller, M. L., & Poulson, T. L. (2010). Subterranean fishes of
North America: Amblyopsidae. In E. Trajano, M. E. Bichuette &
B. G. Kapoor (Eds.), The biology of subterranean fishes
(pp. 169—281). Enfield, NH: Science Publishers.
Poulson, T. L. (1963). Cave adaptation in amblyopsid fishes.
American Midland Naturalist, 70(2), 257—290.
Trajano, E. (2001). Ecology of subterranean fishes: An overview.
Environmental Biology of Fishes, 62, 133—160.
Turquin, M. J., & Barthelemy, D. (1985). The dynamics of a popula¬
tion of the troglobitic amphipod Niphargus virei Chevreux.
Stygologia, 1, 109—117.
ENCYCLOPEDIA OF CAVES
MAMMOTH CAVE SYSTEM,
KENTUCKY
Roger W. Brucker
Cave Research Foundation
INTRODUCTION
Mammoth Cave System is located about 160 km
(100 miles) south of Louisville, Kentucky, and about
56 km (35 miles) northeast of Bowling Green, Kentucky.
Most of the cave lies within Mammoth Cave National
Park, a World Heritage Site and a part of the United
Nations program of International Biosphere Reserves. It
is the longest cave in the world by a factor of 3, with
about 631 km (392 miles) of surveyed passage (Fig. 1).
Between its historical discovery in the late 1700s and
1957, it was considered a single cave 59 km (37 miles)
long, although surrounded by several long caves.
RECENT AND ONGOING EXPLORATION
In 1957, the discovery of a connection between Floyd
Collins' Crystal Cave and Unknown Cave in nearby
Flint Ridge revealed the likelihood that additional dis¬
coveries of natural connections between these long
caves would increase the length of Mammoth Cave. In
1972, a team of explorers led by John Wilcox discovered
a natural passage connection between the integrated
Hint Ridge Cave System (connecting Colossal Cave,
Salts Cave, Unknown Cave, and Floyd Collins' Crystal
Cave) and Mammoth Cave, a system then totaling
232 km (144 miles) in length (Brucker and Watson,
1987). Between 1972 and 1983, connections were found
between Proctor Cave, Mammoth Cave, and Roppel
Cave (Bordon and Brucker, 2000), extending the under¬
ground labyrinth outside the boundaries of Mammoth
Cave National Park.
Exploration continues, and Mammoth Cave as of
2010 is separated from the 185 km (115 miles) Fisher
Ridge Cave System by only 90 m (300 feet). Explorers
familiar with the caves of the Central Kentucky Karst
region have predicted that the Mammoth Cave System
may eventually reach 1600 km (1000 miles) in length,
although it might require cave diving to realize this
potential. The park is visited by 750,000 annually.
About half of the park visitors go on guided cave trips
under the leadership of the National Park Service.
PREHISTORIC EXPLORERS AND MINERS
The first explorers of Mammoth Cave were indigenous
inhabitants of Eastern North America. Radiocarbon dating
of organic materials they discarded or lost underground
indicates an activity span of about 2400 years, starting in
2250 BC. These explorers entered the caves originally for
seasonal shelter and to explore the labyrinth of passages;
later they mined mirabilite, gypsum minerals including
selenite (euhedral and subhedral gypsum crystals), and
epsomite. These ancient cavers explored about 19 km of
passages using cane and dry weed-stock torches for light.
Extensive archaeological investigations have led to a
reconstruction of the eastern aboriginal diet, based primar¬
ily on plant materials recovered from human paleofeces
preserved in the cave (Watson, 1997). A few desiccated
aboriginal corpses (mummies) have been found, one of a
prehistoric miner crushed when a heavy rock shifted.
Since 1999 several large passages have been found
that were entered by the prehistoric Indians, but which
were not found by later explorers. These discoveries
include several trunk passages averaging 10 m wide
and which extend for over 250 m each. They contain a
rich assemblage of prehistoric soot markings, wall bat¬
tering, torch fragments, and other organic materials
with no evidence of historic or modern visitation.
While the earliest prehistoric explorations may have
been motivated by curiosity, it has been hypothesized
Encyclopedia of Caves.
469
© 2012 Elsevier Inc. All rights reserved.
470
M
FIGURE 1 The Mammoth Cave System as compiled [C] 2007. The map shows 587 km (365 miles) of passages mapped by the Cave
Research Foundation and Central Kentucky Karst Coalition. Map courtesy of the Cave Research Foundation. Used with permission.
that the later extensive mining of gypsum and other
minerals was for ceremonial uses.
HISTORY OF MAMMOTH CAVE
The oral history of Mammoth Cave begins with long
hunters, solitary woodsmen who roamed the forested
Midwest in search of game in the late 1700s. The cave
was mined for saltpeter, one ingredient of gunpowder,
by a succession of owners, culminating in a commercial
saltpeter mining operation with 70 slaves during the
War of 1812. Wagonloads of saltpeter were shipped to
Delaware and made into black powder by E. I. duPont.
With the spread of population westward into Kentucky
between 1795 and 1840, a road network and stagecoaches
brought tourists to visit Mammoth Cave. Mammoth
Cave was part of the "grand tour" by international visi¬
tors, along with Niagara Falls, NY (Olson, 2010). In 1838,
Franklin Gorin, an owner of the cave, brought his slave,
Stephen Bishop, to guide visitors through the cave. He
proved to be one of the most personable and celebrated
explorers and guides in a tradition of African-American
guides that extended unbroken until 1941. His cave map
shows about 40 km (25 miles) of the cave.
The building of the Louisville and Nashville Railroad
in 1858 swelled tourism and created opportunities
for entrepreneurs to exhibit nearby smaller caves.
Commercial rivalry intensified as the fiercely competi¬
tive "cave wars" characterized the struggle for tourist
dollars. The cave wars, fueled by the highly publicized
FIGURE 2 The Historic Entrance to Mammoth Cave. The Historic
Entrance to Mammoth Cave is one of 26 entrances to the cave sys¬
tem. Only three are natural: Historic Entrance, Salts Cave Entrance,
and Echo River Entrance (underwater.). Photo courtesy of the National
Park Service.
entrapment and death of cave explorer Floyd Collins in
Sand Cave in 1925 (Murray and Brucker, 1982), gradu¬
ally ebbed as marginal caves went bust and the move¬
ment to create Mammoth Cave National Park gained
momentum. The park was formally established in 1941
and dedicated in 1946. As of 2011, Mammoth Cave has
26 entrances (Fig. 2), a fact that enables explorers to
reach unexplored passages relatively quickly compared
with travel times in large one- or two-entrance caves.
ENCYCLOPEDIA OF CAVES
MAMMOTH CAVE SYSTEM, KENTUCKY
471
GEOGRAPHIC AND GEOLOGIC SETTING
The Mammoth Cave System is developed in
Mississippian limestone rocks on the west side of the
Cincinnati Arch, primarily the Girkin limestone, Ste.
Genevieve limestone, and St. Louis limestone. The
Mammoth Cave Plateau, or Chester Cuesta, is sepa¬
rated from the Pennyroyal Plateau by the Chester
Escarpment, which rises 60 m (200 feet) above the
Sinkhole Plain to the Mammoth Cave Plateau. The
Sinkhole Plain and Mammoth Cave Plateau contain
the Mammoth Cave System, which is drained by the
deeply incised and meandering Green River. The rocks
dip gently northwest (Palmer, 1981).
The Central Kentucky Karst is an area of about
390 km 2 (150 square miles), consisting of 30 or more
overlapping karst drainage basins. Sinkholes and sink¬
ing creeks funnel runoff and groundwater to caves.
Rolling farmland with hundreds of closed depression
sinkholes characterizes most of the area's topography.
The Mammoth Cave Plateau topography is developed
in a dendritic pattern of valleys and ridges oriented
generally perpendicular to Green River. The upland
represents a fluvial surface on the flat-lying Big Clifty
Sandstone, with remnants of overlying Upper
Mississippian bedrock on the highest ridges. The val¬
ley drainage breached the resistant sandstone and was
diverted into sinkholes, creating karst valleys below
the base of the sandstone (White et al., 1970).
The ages of Mammoth Cave passage levels and their
clastic fills generally correlate with the adjustment of
Green River elevations over a period of 3.5 million
years during the late Pliocene and Pleistocene (Granger
et al., 2001). Cosmogenic Al and Be in passage sedi¬
ments indicate at least four periods of downcutting and
at least three periods of aggradation, and the intervals
are in agreement with surface strath terrace level evi¬
dence. Such events as ice sheet advances and retreats,
drainage adjustments, and climate changes influenced
the paleo-Teays River, Ohio River, and Green River
water-table positions through time. Approximate ages
of fills located in passages on tourist routes are
Hippodrome 3.5 myr. Main Cave 2.35 myr, Violet City
2.27 myr. Forks of the Cave 1.78 myr, Cleaveland
Avenue. 1.34 myr, and Buchanans Way 0.8 myr. The
age of original passage development probably dates to
as much as 10 myr prior to the deposit of the fills.
Unlike rivers on the surface where drainage divides
can be delineated easily in the field, these karst basins
and valleys are dynamic. The subsurface consists of a
three-dimensional network of integrated conduits that
preserve the underground former flow paths or paleo-
drainage routes, cave passages that have migrated
downward through time. Basin low water boundaries
in the area have been determined by dye tracing, but
heavy rains overwhelm the constricted conduit system
and back up cave water stages to twice or more the
base level river stage. Under these conditions, water
may be pirated through earlier drains to adjacent
basins. The interconnected conduits not only create
spillover routes, but also serve as migration corridors
for cave-dwelling animals and leads for cave explorers
(White and White, 1989).
MAMMOTH CAVE PASSAGE PATTERNS
Passages in Mammoth Cave are generally arranged
in a tree-like dendritic or angulate pattern on several
levels. Truncations of passages and the development
of piracy conduits complicate interpretation of the
pattern. The farthest upstream ends are either narrow
canyon-like passages or vertical shafts (Fig. 3), cylin¬
drical voids that resemble the interior of grain storage
silos up to 12 m (40 feet) in diameter and 60 m (200
feet) deep (Brucker et al., 1972). Such shafts generally
have small drains relative to their volume. Vertical
shafts are often found at the edges of upland
FIGURE 3 Keller Well. A typical vertical shaft is cylindrical like
the inside of a silo. Shafts short-circuit groundwater from sinkholes
to small drains leading to the lowest levels. Many shafts contain
waterfalls during heavy rains and thin films of aggressive descend¬
ing water during drier times. A meter-long stick shows scale in
Keller Well. Photograph taken by the author.
ENCYCLOPEDIA OF CAVES
472
M
reentrants where water flows off the impermeable cap-
rock and sinks into the exposed limestone. Shaft catch¬
ment areas may be hundreds of square meters of
sandstone upland. High volumes of fast-moving water
descend the shaft walls to produce a rapid rate of
enlargement and material removal. Slower moving
films of water descend the walls continuously.
Vertically flowing groundwater in the shafts represents
a karst "chain saw" in comparison to the dissolving
capacity of slow-moving water flowing horizontally.
Slope retreat gradually isolates the larger shafts from
their catchment areas. Moving downstream, shaft
drains and sinkhole drains join to form downcutting
canyons that change to tubular passages with elliptical
cross-sections as they reach a temporary base level
(Fig. 4).
In their downstream extremities, the lowest active
level trunk passages sump at the Green River level. An
artificial pool of the Green River at 106 m (421 feet)
mean sea level is behind Dam No. 6 located at
Brownsville, KY. Built in 1893, the pool backflooded to
raise the normal pool stage in the cave about 1 m.
Cave divers have reported submerged elliptical cross-
section tubes from 2 to 10 m deep at the resurgences of
Turnhole Spring, Echo River Spring, and Pike Spring
on Green River. Additionally, water discharges
through root-like distributary systems below the pool
stage into the river.
Upper levels of the flat-lying cave were once contin¬
uous drains from the karst valleys and Sinkhole Plain
to the paleo-Green River. The highest elevation pas¬
sages are cut into segments by surface valleys that
break down passage ceilings and by vertical shafts that
short-circuit surface runoff and groundwater to the
lowest active levels of the cave.
FIGURE 4 Cleaveland Avenue. Cleaveland Avenue is an elliptical
tube in cross-section, suggesting shallow phreatic origin. It diverted
water from the higher original trunk passages. Abundant gypsum
crystals adorn walls and ceilings. Photo courtesy of the Notional Park
Service.
Some passages have been truncated by sinkholes or
valleys that formed after the active stage has passed.
Inside the cave, such segments are marked by terminal
breakdowns or fills. A cave map may show a continua¬
tion of the original passage, confirmed by elevation
data and stratigraphic correspondence. In the
Mammoth Cave System, several of these segmented
trunk passages can be traced for miles. One example of
a sequence of segments is Kentucky Avenues, extending
from the Frozen Niagara Entrance to a breakdown at
Grand Central Station, a missing segment, and then
continuing as Sandstone Avenue, another breakdown,
followed by a long segment starting at Violet City
Entrance and extending as Broadway to a distributary
net near the Historic Entrance. One of the distributary
passages is Houchins Narrows that leads to the Historic
Entrance. Yet another segment is Dixon Cave that termi¬
nates at a former big spring at the paleo-Green River.
One or more of the missing segments of this sequence
have been destroyed or may await discovery in the
future.
A primary reason why so many hundreds of miles
of passages are present in Mammoth Cave is that the
capping bed of clastic rock acts as a "roof" to the cave
system. The resistant Big Clifty Sandstone and a thin
bed of impermeable shale at its base constitute the cap-
rock that has protected much of the cave system from
erosion. A mapping program has revealed how pro¬
tected passage fragments were part of a set of evolving
drainage patterns that show how subterranean water¬
courses have developed through time. As can be seen
in the lowest water-filled passages as well as at upper
levels, drainage diversion or piracy is a ubiquitous
occurrence. Diversion routes often develop as trenches
cut in the bedrock floor of a tube or canyon. Their gra¬
dient deepens and they may depart from the original
passage to become tubes or canyons leading to lower
reaches of the system.
MAMMOTH CAVE MINERALS
Mammoth Cave is not noted for extensive displays
of speleothems, such as stalactites, stalagmites, col¬
umns, and flowstone. These depositional formations
are confined to parts of the cave system where the
clastic capping bedrock is thin or has been removed
and carbonate-saturated groundwater can descend.
The Frozen Niagara section shown to visitors contains
a significant display of these speleothems.
Various evaporative minerals can be seen in parts of
the cave where the clastic caprock above it remains
intact. These include gypsum crystals, needles, cotton,
flowers, massive crusts, and loose deposits resembling
drifted snow (Fig. 5). Less common are epsomite and
ENCYCLOPEDIA OF CAVES
MAMMOTH CAVE SYSTEM, KENTUCKY
473
FIGURE 5 Gypsum flowers are found where the sandstone and
shale caprock are intact over the cave passages. Photo courtesy of the
National Park Service.
mirabilite crystals. The origin of the sulfate may be pyrite
in the limestone or overlying beds, or dissolved gypsum
from overlying beds. Investigators believe that aborigines
gathered and used mirabilite as a laxative and perhaps
as food seasoning, gypsum as a paint ingredient, and sel¬
enite crystals as ceremonial objects.
MAMMOTH CAVE ORIGIN
As a significant and complex cave. Mammoth Cave's
origin in process terms has been studied since the
1920s. Early studies of the cave were mainly descriptive
and based on only about 10% of the cave system known
today. Inadequate cave maps also hampered early
investigations. A.C. Swinnerton observed that horizon¬
tal passages in Mammoth Cave were formed in the
"zone of discharge," a 30-m (100-foot) band located
between the highest stage of the base level Green River
and its lowest cave resurgence. At Mammoth Cave, this
"shallow phreatic zone" extends from 9 m (30 feet)
below pool stage to 21 m (70 feet) above pool stage. The
regional dip is toward the Green River, but some pas¬
sages veer across the strike. Hence, it appears that the
dip is not the only structural control (Palmer, 1981).
BIOLOGY OF MAMMOTH CAVE
The caves of the Mammoth Cave area have a very
diverse and well-studied biology. Currently, 41 species of
cave-adapted organisms are known from the Mammoth
Cave area. This number means that the park has one of
the highest subterranean diversities worldwide. Scientific
study of the cave biology of the Mammoth Cave System
began in the middle of the 1800s and actively continues
today. These studies have included taxonomic studies of
specific organisms, ecological studies of terrestrial and
aquatic systems, and evolutionary studies of the adapta¬
tion of cave animals. Poulson (1992) maintains that
Mammoth Cave is the best-studied and best-understood
cave ecosystem in the world. Notable species include
the endangered, endemic Kentucky Cave Shrimp
(Palaemonias ganteri Hay), two species of troglobitic fish,
and five co-occurring cave trechine beetle species. The
caves of the park are also home to two endangered spe¬
cies of bats. The caves of Mammoth Cave National Park
are the type locality of over 25 cave-adapted species.
Two aspects of this high species diversity are
explained here. First, the kinds of obligate and faculta¬
tive organisms are especially high for the aquatic envi¬
ronment; that is, we find fish, crayfish, shrimp, isopods,
amphipods, flatworms, copepods, worms, and snails.
Because of the gradients in the amount, kind, and qual¬
ity of organic input, and also water quantity, there are
replacement of species in several groups, for example,
four fish, two crayfish, two isopods, three amphipods,
many copepods, and two flatworms. Second, in geo¬
graphic scale terms. Mammoth Cave lies at the intersec¬
tion of several cave and karst underground and surface
species dispersal corridors along escarpments and sink¬
hole plains. Different species come from all directions.
Striking examples are the four fish and seven Carabid
beetles living in different parts of the cave.
Prominent among observable animals are cave
guests like bats and pack rats, cave lovers like sala¬
manders and cave crickets, and permanent cave dwell¬
ers like blindfish and beetles.
Since there is no light in the cave and hence no
growth of green plants, cave organisms must rely on
food brought in from above ground. Food chains in
aquatic habitats start with dead plant material washed
into the cave from cracks, sinkholes, and sinking
streams. Sinkhole streams and sinks above the oldest
vertical shafts, such as Mammoth Dome Sink and dry
ravines, take coarse particulate organic matter during
rare heavy rain events, but only fine particulate organic
matter and mostly dissolved organic matter enter via
the epikarst of the Sinkhole Plain and Mammoth Cave
Plateau karst valleys. Such material is serially reduced
in amount, size, and quality as it moves through the
epikarst. Decline in food quality also occurs through
deposition, leaching, and the use of decomposers from
active vertical shafts to master shaft drains to base level
rivers. This leads to a parallel decrease in density and
hence change in aquatic species, for example, two dif¬
ferent species of isopods, three of amphipods, two of
flatworms. Downstream of master shaft drains, one can
observe a change in species of cavefish, shrimp, and
snail as passages become wider and deeper with lower
velocity and more bottom silt.
ENCYCLOPEDIA OF CAVES
474
M
Decomposers such as bacteria and molds do the ini¬
tial breakdown of organic matters and animals eat the
decomposers. The most important food chains in terres¬
trial habitats start with the feces and eggs of cave crick¬
ets. Cave crickets roost inside cave entrances and at
night venture out to forage for bits of dead and decom¬
posing matter. One community of organisms depends
on the veneers of guano under the dense roosts of cave
crickets. Another community depends on the feces of
cave beetles that eat eggs that crickets lay in the sand
away from the cave entrance. Cave beetles search vigor¬
ously for cricket eggs using touch and smell to locate
the sand that crickets rake over their eggs.
From ceiling stains, it is evident that Mammoth Cave
used to contain millions of hibernating bats, but today
there are very few. Reasons may include the alteration
of microclimate due to entrance modification and visita¬
tion by saltpeter miners and hundreds of thousands of
modern tourists. Passages beneath the caprock are espe¬
cially dry, and this influences the biology, mineralogy,
and archaeology. Hydrated sulfate minerals, for exam¬
ple, epsomite, mirabilite, and gypsum, apparently
"suck" the moisture from living and dead organisms.
For the biologist these are the "Great Kentucky Desert,"
but for the archaeologist they are sites without decom¬
position and thus preserved the remnants of aboriginal
people who used the caves.
As with any aquatic life, the organisms of Mammoth
Cave are vulnerable to pollution. Several disastrous
petroleum and chemical spills have confirmed the sen¬
sitivity of species in large streams, especially long-lived
shrimp, crayfish, and fish at the end of food chains. An
emerging area for research is bioaccumulation and bio¬
magnification of toxins, such as mercury, that migrate
up the food chain, affect the reproduction of long-lived
animals, and are remobilized by floods. Another area of
concern is incompletely treated human waste. This
excess organic matter favors fast-reproducing species
whose numbers overwhelm the usual cave species and
so reduce species diversity. Awareness of pollution
threats is focusing keen attention on violations of the
Clean Air Act and Clean Water Act, and the effects of
development on the Tier III protected waters of
Mammoth Cave.
A MAGNET FOR INTERNATIONAL
CAVE RESEARCH
The modern era of cave investigation at Mammoth
Cave started in 1954 (Lawrence and Brucker, 2010) and
with the organization of the Cave Research Foundation
(CRF) in 1957. That group undertook a systematic
exploration and mapping program as the basis for a
scientific program in archaeology, biology, and
hydrogeology. More than a dozen books, dozens of the¬
ses, and hundreds of scientific papers and cave maps
have resulted from a cooperative partnership between
the National Park Service and CRF scientists.
University research and educational activities associ¬
ated with Mammoth Cave include three organizations
at Western Kentucky University, Bowling Green, KY:
the Center for Cave and Karst Studies, the Hoffman
Institute, and the Mammoth Cave International Center
for Science and Learning.
Bibliography
Borden, J. D., & Brucker, R. W. (2000). Beyond Mammoth Cave.
Carbondale, IL: Southern Illinois University Press.
Brucker, R. W., Hess, J. W., & White, W. B. (1972). Role of vertical
shafts in the movement of ground water in carbonate aquifers.
Ground Water, 10(6), 5—13.
Brucker, R. W., & Watson, R. A. (1987). The longest cave. Carbondale,
IL: Southern Illinois University Press.
Granger, D. E., Fabel, D., & Palmer, A. N. (2001). Pliocene—Pleistocene
incision of the Green River, Kentucky, determined from radioactive
decay of cosmogenic 26 A1 and 10 Be in Mammoth Cave sediments.
Geological Society of America Bulletin, 113, 825— 836.
Lawrence, J., Jr., & Brucker, R. (2010). The caves beyond: The story of the
Floyd Collins' Crystal Cave exploration (reprint of the 1955 edition).
Dayton, OH: Cave Books.
Murray, R. K., & Brucker, R. W. (1982). Trapped! The story of Floyd
Collins. Lexington, KY: University Press of Kentucky.
Olson, C. O. (2010). Mammoth Cave by lantern light: Visiting America's
most famous cave in the 1800s. Dayton, OH: Cave Books.
Palmer, A. N. (1981). A geological guide to Mammoth Cave National
Park. Teaneck, NJ: Zephyrus Press.
Poulson, T. L. (1992). The Mammoth Cave ecosystem. In A. Camacho
(Ed.), The natural history of biospeleology (pp. 569—611). Madrid:
University of Madrid.
Watson, P. J. (Ed.), (1997). Archeology of the Mammoth Cave area.
St. Louis, MO: Cave Books.
White, W. B., Watson, R. A., Pohl, E. R., & Brucker, R. W. (1970). The
Central Kentucky Karst. Geographical Review, 160(1), 88—115.
White, W. B., & White, E. L. (Eds.), (1989). Karst hydrology, concepts
from the Mammoth Cave area. New York: Van Nostrand Reinhold.
MAPPING SUBTERRANEAN
BIODIVERSITY
Mary C. Christman and M aja Zagmajster
Department of Statistics, University of Florida, ^Department
of Biology, University of Ljubljana
INTRODUCTION
The ability to describe and explain the spatial distribu¬
tion of subterranean fauna is becoming increasingly
more important as humans continue to expand their
reach and encroach on traditionally inaccessible areas.
Encyclopedia of Caves.
© 2012 Elsevier Inc. All rights reserved.
MAPPING SUBTERRANEAN BIODIVERSITY
475
Knowing where and under what conditions various
fauna are found is important from different ecological
and management perspectives. To that end, it is
important to understand the processes that have led to
and sustain such distributions, and in the light of
recent environmental changes, to help in understand¬
ing the impact of future environmental scenarios.
Additionally, it is important for developing appropri¬
ate conservation plans (Adams et al., 2000). Hence, an
increasingly urgent need is the ability to accurately
map the geographic distribution of various species and
ecosystems (and often as critical, the distribution of the
required habitat). Mapping includes not only the
graphical presentation of species occurrences and
derived parameters such as densities and species rich¬
ness over a region but, more importantly, it includes
the statistical spatial modeling that provides both the
map and the means of assessing its accuracy. Here we
present some statistical approaches that provide such
information which can be used in interpretation,
hypothesis testing, and decision making. Statistical
modeling provides the means for studying the spatial
process that generates the observed geographic distri¬
bution of species and hence, it is useful both for
research into the generating process as well as for
identifying areas in which high species richness might
be expected, although not yet reported.
DATA TYPES
To statistically model the spatial distribution of
species or species diversity, one needs to have geo-
referenced positions of the localities where the species
occur; that is, their exact position on the Earth's sur¬
face as determined with coordinates. Accuracy of coor¬
dinates can differ among datasets, referring to the
specific site (cave) at which species were observed or
to the centroid of a large subregion of the study area
in which the species is associated. Data that consist of
the locations of occurrence of species or individuals
are called a point pattern. Point patterns are usually
analyzed for inference concerning whether the loca¬
tions (points) tend to cluster in some meaningful
way. Mapping of these data usually is in the form of
locations of occurrence (presences) on a map.
More commonly, individual locations are not reported
(often for purposes of confidentiality), yet such informa¬
tion can be used for exploration of subterranean biodi¬
versity patterns. The data are spatially aggregated over
subregions within the study area and the resulting
counts or totals are reported at the level of the subre¬
gions. These areas can be artificially created, for example,
grid cells constructed by overlaying a grid on the study
region (Zagmajster et al., 2008a, 2010), or based on
geopolitical boundaries such as counties in the United
States (Christman and Culver, 2001a). These types of
data, referred to as lattice data (Cressie, 1993), are often
analyzed explicitly to determine if spatial autocorrelation
exists, that is, to study the form of the correlation among
neighboring areas. Hence, the level of spatial scale
should be considered as a characteristic of the dataset
used in analyses since it influences the methods used to
statistically model the spatial distribution of the species
or diversity measures.
The aforementioned presentations most often refer
to presence of the species, and such statistics as spe¬
cies richness derived from the data. Adding informa¬
tion on sampling effort increases the value of
biodiversity maps (Zagmajster et al., 2008a, 2010).
In this chapter we focus on the mapping of quantita¬
tive data, especially species richness, either point-
referenced or aggregated to subregions. Yet the
general ideas are the same, whether one is mapping
biodiversity or abundances of individual species or
species assemblages.
DECOMPOSITION OF THE
SPATIAL PROCESS
The process that generates the observed spatial
distribution of species richness can be decomposed
into several components (Cressie, 1993). First, there
is the large-scale trend that can be explained due to
environmental or geophysical features. These explana¬
tory variables might include such attributes as cave
length or depth, cave type, surface climate characteris¬
tics, altitude, latitude, distance from certain features
such as the maximal extent of an ancient ice sheet or
embayment, and others. For example, large-scale pat¬
terns of precipitation might help explain the spatial
arrangement of caves in karst areas and, hence, of cave
species.
Next is mesoscale or small-scale variation that is
not explainable by the large-scale trend. This is local
variability that is not typically captured over the geo¬
graphic scale of the large-scale trend. Such variation
might arise because of unknown factors that operate
at smaller scales, that is, elevation or slope aspect. It
might also be due to even finer spatial features such as
connectivity of caves. The variations at this scale are
often observed to have covariability, that is, adjacent
areas often have similar numbers of species, possibly
due to connectivity. As a result, this variation is often
modeled as having spatial autocorrelation among
nearby areas.
The final component is random error or the unex¬
plained variability that is unaccounted for after the
fitting of the other components. This variability could
ENCYCLOPEDIA OF CAVES
476
M
be due to sampling intensity or to factors not present
in the model, such as size of the cave entrance open¬
ing. It is usually assumed that these variations are
independent of one another, that is, the value at one
site neither influences nor is influenced by a value at a
nearby site.
Using this decomposition of the spatial variability,
we can write that Y(s), the number of species at site s,
is a linear combination of the components:
Y(s) = |i(s) + 6(s) + e(s)
where (i(s) is the large-scale trend at site s, 6(s) is the spa¬
tial correlation component describing the mesoscale vari¬
ability, and e(s) is the random noise or unexplained
variability. In order to do statistical analyses, it is assumed
that the three components are independent of each other.
When this assumption holds, the modeler can estimate
the precision of each estimated model component.
An example of the decomposition of actual data is
given in Figure 1. In this instance, Christman and
Culver (2001a,b) modeled the number of species
found in counties in the southeastern United States
as a function of the number of caves in that county
(large-scale trend), the autocorrelation between adjoin¬
ing counties (mesoscale variability), and random error
(unexplained variability). Counties without caves were
constrained by the model to have zero species; hence,
the flat regions on the graphs.
INVESTIGATING SPATIAL PATTERNS
When mapped, even visual inspection can offer
new insight into the patterns observed, but having geo-
referenced data allows us to explore for patterns that may
not be obvious in a map. Any statistical investigation
FIGURE 1 Perspective plot showing (A) the geographic distribution of the observed counts of stygobites (log-transformed) in counties in
the southeastern part of the United States; (B) the fitted regression estimate of counts for the large-scale component of the model in which log-
transformed number of stygobites is related to the number of caves (also log-transformed); (C) the spatial autocorrelation mesoscale compo¬
nent in which the log-transformed number of stygobites is predicted conditionally using the log-transformed counts of nearby surrounding
counties; and (D) the residual or unaccounted for variation in the data.
ENCYCLOPEDIA OF CAVES
MAPPING SUBTERRANEAN BIODIVERSITY
477
considers the spatial relationships of data, where observa¬
tions close in space are more similar than those farther
apart. Causes range from emigration and immigration,
common history or environmental features, or other
possible reasons. Such proximity-dependent similarity
in values of the variables of interest is called spatial auto¬
correlation. This autocorrelation is an important compo¬
nent of spatial processes and often drives the mesoscale
variation described in the previous section. It makes
spatial statistical analyses different from nonspatial
approaches. When autocorrelation exists, localities are not
considered independent and this must be taken into
account when testing hypotheses concerning the causes of
the observed distributions or whether changes have
occurred in the distribution between two time periods.
Exploratory analyses of the presence and strength of
spatial autocorrelation use autocorrelation indices such as
Moran's I (Moran, 1950), Geary's c (Geary, 1954), black-
white join count statistics (Cliff and Ord, 1981), and the
Mantel test (Mantel, 1967). Spatial autocorrelation can
exist only at certain distances or range of distances, and
not at others. These statistics can be applied at the differ¬
ent distances to identify the extent of spatial autocorrela¬
tion. Another approach that provides both a map and
measures of the spatial autocorrelation involve variogram
estimation and kriging.
INTERPOLATORS AND KRIGING
Besides point occurrence maps and/or presence
within bigger units, a better visual presentation that can
provide additional insight into the pattern observed is
based on interpolation techniques. Numerical interpola¬
tors (Fig. 2) are used to smooth data so that contour
plots or smooth three-dimensional perspective plots
of the variable of interest can be constructed. Several
kinds of interpolators are available, each providing a
slightly different method for smoothing data so that a
continuous surface is displayed. Techniques include
simple methods such as moving averages, locally
weighted smoothing, smoothing splines, and others.
FIGURE 2 Different graphical representations of the same data, number of obligate subterranean beetles (from families Cholevidae and
Carabidae) per 20 X 20-km quadrats in northwestern Balkans. (A) Filled contours based on inverse distance weighting; (B) filled contours
based on ordinary kriging; (C) prediction standard error map of ordinary kriging; (D) a three-dimensional presentation of the ordinary kriging
predictions. Scale refers to (A), (B), and (C). The first legend refers to (A), (B), and (D), and the second legend to (C). Mapping procedure as in
Zagmajster et al., 2008b.
ENCYCLOPEDIA OF CAVES
478
M
Moving-average smoothing involves estimating the
mean for a location using the set of observed points
within a given radius of the location to be predicted.
Locally weighted smoothing uses a weighted mean
instead of the arithmetic mean where the weights are
usually some function of the distances of the observed
points to the location to be predicted. A smoothing
spline replaces the weighted mean with the value pre¬
dicted by a local regression over the points within the
region.
One needs to be aware that the type of the surface
provided by the smoother depends on parameters
that are specified in the smoothing algorithm. For
example, at each point in the vertices of a grid over-
lying the study region, a moving average smoother
uses the average of the p% data points closest to the
point being estimated (p chosen before interpolating).
When p is small, only data located very close to the
prediction site are used; hence, the surface appears
spiky and very similar to the original data. If p is
large, then more of the surrounding data are used
and the result is a much smoother surface. The choice
of parameter values influences the final map display
and thus, like any graphic, can be manipulated to
emphasize or deemphasize features of these data.
These methods provide maps but are nonstatistical in
the sense that they are not models of the spatial pro¬
cess that generates these data and cannot provide
any means of assessing the bias, precision, or accu¬
racy of the map.
A technique that is statistically valid and uses the
spatial structure of the data is kriging (Cressie, 1993;
Haining, 1990) which provides measures of the error
of the map generated from a small set of observed
data. In kriging, the predicted value of Y(so), the num¬
ber of species at a prediction site, s 0 , is the sum of the
estimated large-scale trend plus the estimated autocor¬
relation component. In kriging with no large scale
trend component, the optimal (in the sense of smallest
variance) predictor for a new observation is a weighted
mean (linear combination) of the variable values (Y) at
the n sampled sites (s a , s 2 , ... , s„):
n
P(Y( so)) = ^ AiY(s,)
i=i
where the weights, X„ are functions of the auto¬
correlation between the observed value of Y at s, and
the value of Y at the new site. So- Kriging provides
a means of studying the fine-scale spatial structure
of subterranean biodiversity because autocorrelation
among sites can be ascribed to processes occurring at
the spatial level of individual sites. In addition, it can
be used to interpolate values at intervening locations
between observed sites and hence, to map biodiversity.
for example, over the whole study region. In fact, that
is its most common use, because the kriging procedure
will give a predicted value at an observed site equal to
the observed value at that site. In kriging, one can also
map the standard errors associated with the predicted
values in order to gain some sense of the precision of
the predictions.
Kriging requires some stringent assumptions be
met in order to be a valid method for modeling the
spatial distribution. These include stationarity and
isotropy. A spatial process is said to be second-order
stationary if the following is true: (1) the trend is
constant over the region, that is, the mean |i(s) = |i
is constant for every site s in the region; and (2) the
covariability between two sites may depend on
the distance and direction between them but not
on the location of the sites within the study region.
The second point indicates that the variance of the
Y variable is constant everywhere as well. A process
is said to be isotropic if the covariability between
two sites does not depend on the direction between
them.
When a large-scale trend does exist (i.e., |i(s) ^ p),
one approach is first to model the large-scale trend
and obtain estimates of the varying trend in the mean,
call this |l(s). Then the differences between the
observed values and the estimated large-scale trend,
Y(s) — |l(s) = e(s), which are called the residuals, are
modeled using kriging. The results of kriging the
residuals are estimates of the other two components of
Y, 6(s), and e(s). The estimates |i(s) and E(s) are then
added together to obtain estimates for the original
variable of interest. The standard error of the com¬
bined estimator can be approximated so that statistical
inferences can be explored.
A cautionary note concerning the use of kriging
for mapping subterranean biodiversity is required.
The geologic conditions necessary for caves and other
subterranean features needed by obligate subterra¬
nean species are not normally uniform over the study
region. Kriging assumes that the study region is a
convex polygon within which the sites at which data
are taken represent a sample from the continuous
spatial field and, hence, assumes that the conditions
are continuous over the entire study area. As a result,
it will predict nonzero estimates of species richness at
sites within the study area, even if those sites do not
have any appropriate habitat for cave fauna. In some
instances, this is easily resolved by masking, which
is the process of removing those locations from
possible interpolation or kriging. This is effective
only if the inappropriate area is small relative to
the study region. When habitable areas are patchier,
the correct modification to the modeling is not so
clear. Some work has been done using a method
ENCYCLOPEDIA OF CAVES
MAPPING SUBTERRANEAN BIODIVERSITY
479
called zero-inflated spatial models, but more research is
required to determine the best method of resolving
the spatial relationships under such situations. One
technique that has been used is to aggregate the point
data to another scale, for example, to sum up the
number of caves and total number of species found in
a given area.
MODELING ISSUES
Aggregation and Scale
One of the first issues that must be resolved when
performing spatial modeling is the scale at which the
analysis is to be done. One of the impediments can
be the positional accuracy of the localities; they can be
either poorly described in the literature (so determina¬
tion of exact position is difficult) or public access can
be restricted due to conservation. In that case, data
can still be used, but at a different scale. They can be
aggregated to the level of county, or into the units of
a grid that overlays the study area. Scaling has both a
positive and a negative impact on ecological studies.
First, it may be that different processes are acting at
different scales and the researcher should attempt to
partition the effects of the different scales. For exam¬
ple, similarities between adjacent caves could be due
to cave connectivity (a microscale process), altitude, or
aspect of the hillside on which the cave is located
(a mesoscale process), or the latitude at which the cave
is found (a macroscale process). On the other hand, if
data are aggregated in such a way that the aggregation
scale does not mimic the scale at which a process
is occurring, then it is not possible to study the
relationships at the appropriate scales. The explora¬
tory tests for spatial structure mentioned earlier
(Moran's I, Mantel test, etc.) are useful for identifying
possible spatial scales that are relevant to the problem
at hand.
Figure 3 is an example of an analysis done at
different spatial scales in order to identify the appro¬
priate level of aggregation of the point-level data.
The data on obligate subterranean beetles from the
Dinaric karst region were mapped at different spatial
scales of quadrats. Though generally the pattern of
highest species richness remained the same, there were
some differences. The quadrats of highest species rich¬
ness were not always overlapping among scales.
Additionally, the spatial autocorrelation among quad¬
rats with species numbers changed as the scale chang¬
ed. It increased up to the grid cell sizes of 20 X 20 km,
and began to decrease as the aggregation units got
larger. When the units are too big and cover too big
areas, the similarities among the quadrats start to
diminish. From this analysis it was suggested that the
suitable unit for data aggregation, based on the data
available, was at the scale of 20 X 20-km grid cells
(Zagmajster et ah, 2008a).
The size and shape of the aggregation units can
influence the analyses (Openshaw and Taylor, 1979).
For example, when square grid cells are used, the
neighboring quadrats are usually considered to be
those with a common boundary—this precludes the
grid cells on the diagonal. A possible solution to this
problem of different neighbors is the use of hexagons,
as developed by White, Kimerling, and Overton
(1992). Additionally, hexagons have the advantage of
retaining equal shape and equal area over the globe
(White et ah, 1992), which makes them especially
suitable for analyses when very large geographic areas
are considered. They have been widely used in map¬
ping biodiversity in the United States (Adams et ah,
2000 ).
Another problem of data aggregation is the choice of
the most suitable size of the unit for the problem being
studied. Relationships between variables can change as
the scale of data aggregation is changed, a phenomenon
called the ecological fallacy or the modifiable areal unit
problem (MAUP). A good review of the issues can be
found in Gotway and Young (2002).
When data are aggregated into blocks, the statistical
models used to analyze them and estimate the spatial
relationships differ as well. Now the model is based
on analyzing relationships among discrete units rather
than points. The decomposition remains the same but
the model is more flexible in that isotropy is no longer
required for analysis. In fact, it is common that the
spatial autocorrelation is different along different com¬
pass points in many biological applications. Models for
analyzing aggregated data are provided in Cressie
(1993) and Haining (1990), among others. In order to
make the model tractable, it is sometimes assumed
that the spatial correlation is independent of direction
between sites and location within the study region.
Other simplifying assumptions include constraining
the autocorrelation to exist only if the two sites are
within K miles (or kilometers) of each other or setting
the autocorrelation to be an inverse function of the dis¬
tance between two sites, for example, setting the auto¬
correlation to some function of d(s,t)~ p where d(s,t) is a
distance metric and p is the decay factor, indicating
how fast the autocorrelation drops off as the two sites
(s and t) become farther apart. Note that in this model,
one does not attempt to estimate species richness at
sites in between aggregated locations because, by defi¬
nition, there are no such sites between aggregated
blocks.
Figure 1 was generated using a lattice model with
the mean number of species in any county being a
ENCYCLOPEDIA OF CAVES
480
M
FIGURE 3 Number of obligate subterranean beetles (from families Cholevidae and Carabidae) in northwestern Balkans, mapped at
different scales (marked in lower left corner of maps). The classes are delimited relatively according to maximal number of species per unit
used (maxspp.). Only localities where positions determined with centroid of maximal 3-km radius were considered. See also Zagmajster et al.,
2008a.
ENCYCLOPEDIA OF CAVES
MAPPING SUBTERRANEAN BIODIVERSITY
481
function of the number of caves in that county and of
the number of species in surrounding counties whose
centroids were within 56 km of the county being esti¬
mated (Christman and Culver, 2001a). The spatial cor¬
relation parameterization used was
0, if ||s; — Sj\\ > 56 km
II s ) s y11 1
max{||s r — Stir 1 }’ if ll s ;-s/ll ^ 56km
r,t
where ||s, —s,|| is the Euclidean distance between the
two locations s, and s,- and max r f |s r — s t | is the great¬
est distance between any two locations in the dataset.
The ratio of the maximum distance to the local dis¬
tance constrains the parameter r| to fall within the
interval (—1, +1) and, hence, is interpretable as a corre¬
lation coefficient.
Excess Zeroes
Although the above models work well when the
assumptions are met and there are few zeroes, a com¬
mon problem with obligate subterranean data is the
excess number of zeroes. This occurs either because
there is no habitat available for cave species or because
there are no records of species within individual caves
or subregions. In the first case they can be regarded as
"true" zeroes, while in the second the situation is less
clear. It could be that the cave either has not been
explored yet or has been explored but no organisms
were found. In subterranean fauna, sampling can be
difficult so that species detectability and catchability
are often low and false negatives (species present, yet
not recorded) are common. One way of dealing with
the zeroes then would be to include information on
sampling effort, even if (or especially if) "unsuccess¬
ful." Unfortunately, this is rarely or not recorded in
the literature.
The problem of excess zeroes in mapping can be
partially diminished when data are aggregated into
bigger units, and not analyzed at the original scale
of individual caves. In the case of lack of habitat, the
lattice model can be adapted to require that any pre¬
diction be predicted on the existence of karst (or caves
or whatever other variable might be appropriate).
Some models have been used that allow for the fact
that a zero may be either due to lack of available habi¬
tat or simply no data have been observed ( cf. Ridout
et al., 2001). These models have been developed under
the assumption that there is no spatial autocorrelation
among sites and, hence, are applicable only if the
assumption holds (and therefore has limited usability
in studies of subterranean biodiversity). Some work
has been done to incorporate excess zeroes in a
kriging model (cf. Warren, 1997), but research is still
ongoing.
See Also the Following Article
Documentation and Databases
Bibliography
Adams, J. S., Stein, B. A., & Kutner, L. S. (2000). Biodiversity: Our
precious heritage. In B. A. Stein, L. S. Kutner & J. S. Adams
(Eds.), Precious heritage: The status of biodiversity in the United States
(pp. 3—18). New York: Oxford University Press.
Christman, M. C., & Culver, D. C. (2001a). The relationship between
cave biodiversity and available habitat. Journal of Biogeography,
28(3), 367-380.
Christman, M. C., & Culver, D. C. (2001b). Spatial models for
predicting cave biodiversity: An example from the southeastern
United States. In D. C. Culver, L. DeHarveng, J. Gibert & I. D.
Sasowsky (Eds.), Mapping subterranean biodiversity: Proceedings of
an international workshop held March 18 through 20, 2001, Laboratorie
Souterrain du CNRS, Moulis, Ariege, France (pp. 36—38). Charles
Town, WV: Karst Waters Institute.
Cliff, A. D., & Ord, J. K. (1981). Spatial processes: Models and applications.
London: Pion Limited.
Cressie, N. A. C. (1993). Statistics for spatial data (Rev. ed.).
New York: Wiley & Sons.
Geary, R. C. (1954). The contiguity ratio and statistical mapping. The
Incorporated Statistician, 5(3), 115—145.
Gotway, C., & Young, L. (2002). Combining incompatible spatial data.
Journal of the American Statistical Association, 97(458), 632—648.
Haining, R. (1990). Spatial data analysis in the social and environmental
sciences. New York: Wiley & Sons.
Mantel, N. (1967). The detection of disease clustering and a generalized
regression approach. Cancer Research, 27(1), 209—220.
Moran, P. A. P. (1950). Notes on continuous stochastic phenomena.
Biometrika, 37(1—2), 17—23.
Openshaw, S., & Taylor, P. J. (1979). A million correlation coeffi¬
cients: Three experiments on the modifiable areal unit problem.
In N. Wrigley (Ed.), Statistical methods in the spatial sciences
(pp. 127—144). London: Pion Limited.
Ridout, M., Hinde, J., & Demetrio, C. G. B. (2001). A score test
for testing a zero-inflated Poisson regression model against
zero-inflated negative binomial alternatives. Biometrics, 57(1),
219-223.
Warren, W. G. (1997). Changes in the within-survey spatio-temporal
structure of the northern cod ( Gadus morhua) population,
1985—1992. Canadian Journal of Fisheries and Aquatic Sciences,
54(Suppl.), 139-148.
Zagmajster, M., Culver, D. C., & Sket, B. (2008a). Species richness
patterns of obligate subterranean beetles (Insecta: Coleoptera)
in a global biodiversity hotspot—Effect of scale and sampling
intensity. Diversity and Distributions, 14, 95—105.
Zagmajster, M., Sket, B., & Culver, D. C. (2008b). The representation
of subterranean beetle biodiversity with the use of interpolation meth¬
ods. In D. Perko, M. Zorn, N. Razpotnik, M. Ceh, D. Hladnik,
M. Krevs, T. Podobnikar, B. Repe & R. Sumrada (Eds.), Geographical
information systems in Slovenia 2007—2008 (pp. 237—245). Ljubljana,
Slovenia: ZRC Publishing.
Zagmajster, M., Culver, D. C., Christman, M. C., & Sket, B. (2010).
Evaluating the sampling bias in pattern of subterranean species
richness: Combining approaches. Biodiversity and Conservation,
19(11), 3035-3048.
ENCYCLOPEDIA OF CAVES
482
M
MARINE REGRESSIONS
Claude Boutin and Nicole Coineau '
Universite Paid Sabatier, Toidouse, France,
^Observatoire Oceanologique de Banyuls,
Universite Paris 6, France
groups such as Decapoda, Isopoda, and Amphipoda.
In these groups the colonization of continental fresh-
waters by marine ancestral populations may have
occurred repeatedly at different periods, originating
groups of continental species that were different
from their distant marine ancestors.
BACKGROUND
The ocean milieu not only covers more than two-thirds
of the Earth's surface, but very likely has played a
major role in the appearance, the evolution, and diver¬
sification of the life on the whole planet, including the
different terrestrial and aquatic habitats existing on the
continents. Even caves and other subterranean aquatic
biotopes were often colonized by animal species
directly coming from marine habitats during the peri¬
ods of marine regressions. Because of this, the present
geographic distribution of some aquatic cave inhabi¬
tants is closely correlated with the past position of
shorelines before a recession of the sea. In some cases
cave or groundwater animal species are the sole evi¬
dence of past ocean presence, and could be used to
date the marine regression period.
FROM THE SEA TO FRESHWATER
AND LAND
Most biologists and paleontologists have long
agreed that life appeared first in marine environ¬
ments, and that all animal phyla (all the different
lineages of the animal kingdom) which later colo¬
nized continents first existed and evolved in the
sea. A good number of phyla are still exclusively
marine; some of them are well known, such as the
Echinodermata (sea urchins, starfishes) or the
Cephalocordata (the amphioxus), while other phyla
such as the Ctenophora or the Chaetognatha are
known only by zoologists. In contrast, a great num¬
ber of phyla have both marine branches and other
branches living in continental freshwaters or in ter¬
restrial habitats. This is the case of Arthropoda and
Vertebrata, which include many marine species and
many other species presently living in river and lake
freshwaters or in terrestrial habitats of continents.
While colonizing the freshwater habitats long ago,
many groups continued to live in the sea; they are
therefore represented both by marine and freshwater
species belonging to the same order, often to the
same family, and sometimes to the same genus. Such
situations may be observed in many crustacean
ADAPTABILITY TO DIVERSE HABITATS
Since the 1970s, biologists interested in the develop¬
ment of life on Earth and in the diversity of living
organisms have made significant advances in the field
of historical biogeography and evolutionary ecology.
Now we better understand why species occur where
they are presently distributed. These researchers have
also discovered the surprising ability of many species
to colonize a great variety of habitats, including the
most extreme and drastic environments. Such habitats
occur in polar regions covered by ice caps; in deep
parts of oceanic plains; around the hydrothermal
vents that form sea bottom oases; in the deep part of
continental caves; and in all subterranean waters pres¬
ent in continental caves, marine caves, and anchihaline
caves (a kind of coastal cave related to the sea but
filled with water also subjected to continental freshwa¬
ter influences), as well as in continental and fresh
groundwaters.
There is a general agreement for considering that
the ancestral populations of many troglobionts (species
living in caves and other terrestrial subterranean habi¬
tats) lived first in soils, humus, litter, or moss strata of
forest ecosystems. They lived mainly during the
Cenozoic and Pleistocene periods, in temperate
regions, and probably at any period from the Mesozoic
to the present in tropical regions. Thus it is shown that
terrestrial subterranean species or populations are
derived from surface ancestral species.
For the limnostygobionts, the subterranean fresh¬
water species presently living in caves or in conti¬
nental groundwaters, the question is less simple
because their origin is double. Some species clearly
belong to groups that presently live in surface fresh¬
waters and appear to be derived from surface fresh¬
water ancestral populations. But many other species
of limnostygobionts are representatives of marine
groups, totally absent in surface freshwaters, and
showing evident phylogenetic affinities with marine
species or groups belonging to the marine plank¬
tonic, benthic, or interstitial communities. With
reference to their origin, the first group of limnosty¬
gobionts was called limnicoid and the second group
tlwlassoid stygobionts.
Encyclopedia of Caves.
© 2012 Elsevier Inc. All rights reserved.
MARINE REGRESSIONS
483
MARINE REGRESSIONS
AND THE BIOGEOGRAPHY OF
THALASSOID STYGOBIONTS
The direct marine origin of the thalassoid limnostygo-
bionts has been postulated and acknowledged for a long
time by many stygobiologists who came first from France
and then from many different countries in the world.
Among the thalassoid limnostygobionts occurring
in continental subterranean habitats around the present
marine shores, it is possible to mention some Protista
such as the Foraminifera known from groundwaters of
Central Asia, the Sahara, and Morocco and a high num¬
ber of Metazoa (or Animalia phyla). Subterranean
Porifera and Cnidaria are thalassostygobionts still living
in littoral caves, but Platyhelminthes, Nematoda,
Annelida (Oligochaeta and Polychaeta), and Gastrotricha,
Mollusca, and Arthropoda (mainly representatives of
many Crustacea orders but also Flydracarina) often occur
within the freshwater thalassoid stygofauna.
The ecological and evolutionary processes that chan¬
ged the habitats of some ancestral marine populations
that originated the thalassoid limnostygobionts have
interested stygobiologists for the past 30 years (Coineau
and Boutin, 1992; Flolsinger, 2000). J. Stock (1980) was
the first to propose a regression model evolution which
provided an understanding of the occurrence and dis¬
tribution of stygobiontic amphipods living in ground-
waters of several Mediterranean countries and Atlantic
islands. The marine regression was considered the
event responsible for the origin of subterranean thalas¬
soid crustaceans. These species were simply viewed as
descendants of marine populations "stranded" by the
receding sea. Then several authors independently
described a two-step model of evolution and colonization
emphasizing the succession of two very different eco¬
logical changes. The first one, occurring in marine
coastal habitats, is called the "vertical transition." It is
the active colonization of subterranean milieus, either
in marine caves or more often in interstitial and crevicu-
lar environments, correlated with an adaptation to the
life in darkness and, in the latter habitats, in small size
interstices. The second step, sometimes called a horizon¬
tal transition, is an adaptation to life in brackish water
and progressively in freshwater. This may have
occurred passively, first in the intertidal zone, and then
continued on the spot in groundwater during and after
the sea recession (Boutin and Coineau, 1990; Flolsinger,
1994; and Fig. 1).
This way of viewing the origin of thalassoid limno¬
stygobionts occurring in caves or in groundwaters,
and the feeble ability for active dispersal observed in
interstitial and subterranean biota, allows us to under¬
stand the correlation that may be observed between
the extent of past marine gulfs and the present
distribution area of continental thalassoid stygobionts
(Fig. 2). Notenboom (1991) emphasized such correla¬
tions between the location of past marine gulfs and the
distribution of present stygobiotic amphipods from
Spain. Boutin (1994a) explained similar correlations
occurring with the metacrangonyctid amphipods from
Morocco. Many other distribution patterns of limnosty-
gobiontic peracarid crustaceans including a number of
thalassoid amphipods, isopods, and Thermosbaenacea
could probably be well understood in the same way
when considering the two-step model of evolution and
colonization.
WHEN THALASSOID
LIMNOSTYGOBIONTS ARE
THE SOLE EVIDENCE OF A MARINE
TRANSGRESSION-REGRESSION CYCLE
Usually subterranean biologists use the data pro¬
vided by the studies of regional historical geology and
paleogeography for the reconstruction of biogeo¬
graphic scenarios that explain the observed species
distribution pattern. However, an exactly opposite
approach, using biological and paleogeographical data,
is sometimes possible. Furthermore, it may happen
that some thalassoid limnostygobionts, belonging to a
well-known group of species, occur in a region charac¬
terized by an incompletely known paleogeographic
history. This can occur when all the Mesozoic and
Cenozoic rocks have been removed by recent erosion.
Such a situation occurs in southwestern Morocco
where two locations of the amphipod Longipodacrangonyx
are known in the groundwaters of the Tiznit region,
about 40 km south of the known southern limits of the
Eocene marine gulf of the Souss Valley. Farther south,
near the city of Guelmim, another amphipod belonging
to the most primitive lineage of the genus Metacrangonyx
occurs in many sites. Other amphipod species of
Longipodacrangonyx and species of the primitive lineage
of Metacrangonyx are well known in many other regions
of Morocco where they clearly exhibit a distribution
respectively related to the past marine embayments of
the Lutetian (Mid-Eocene) sea and with those of the
Cenomano-Turonian (Mid-Cretaceous). Therefore they
are considered the descendants of coastal ancestral popu¬
lations which colonized the continental sediments during
the regressions of the Late Lutetian (about 40 Ma BP) for
Longipodacrangonyx and during the Turonian (about
90 Ma BP) for one group of Metacrangonyx species. The
geological history of the regions of Tiznit and Guelmim
is far from well known because of a recent uplift of
the region resulting in the erosion of all marine sedi¬
ments of Mesozoic and Cenozoic age. There are no
data or marks of possible marine deposits in the
ENCYCLOPEDIA OF CAVES
484
M
FIGURE 1 Origin of two present thalassoid limnostygobiontic species S'l and S'2, in relation to two successive marine transgression-
regression cycles, from time T1 to the present, T3. The sea level first reached shoreline SLT1 at time T1 and then the sea receded. Later a new
transgression reached the shoreline SLT2 at time T2 and receded until the present shoreline SLT3. SI was a marine species that colonized the
coastal sediments at time T1 (vertical transition), S2 was a thalassostygobiont living during T2 period, and S3 is a present thalassostygobiontic
species. The present limnostygobionts, S'l and S'2, are derived from SI and S2 during the sinking of the water table up to the present water
table (PWT). WA, WB, and WC are three present wells. Two vicariant thalassoid stygobiontic species, S'l and S'2, occur in WB and WC since
their origin is related to the two different regressions that occurred after T1 and T2 (because of an impermeable black layer reducing the possi¬
ble genetic fluxes between the groundwater populations of WB and WC).
region that are more recent than 400 Ma BP (Choubert
and Faure-Muret, 1962; Andreu-Bousut, 1991). The
occurrence and location of Longipodacrangonyx is an
indication of an Eocene gulf more southerly extended
than supposed from observation of preserved sedi¬
ments, and the occurrence of Metacrangonyx east of
the present shoreline is evidence of the presence of
the Cretaceous sea during a period that ended with
the Turonian marine regression. The presence of this
ancient sea near Gelmim and Tiznit is not really sur¬
prising for geologists, since the Tarfaya Cretaceous
Basin is well known about 200 km in the south, as
well as the High Atlas Cretaceous Basin immediately
north of the Souss Valley, near Agadir, where marine
sediments are preserved. However, the continuity of
the Turonian Sea from the Tarfaya Basin to the High
Atlas was only a possibility before the discovery in
this region of a species of Metacrangonyx, which
clearly belongs to a group of species that settled in
continental groundwaters during the regression of the
Turonian shorelines.
Generally, after a marine regression, the presence
of the sea is usually marked on the continent by the
presence of sedimentary rocks containing classic fos¬
sils used for dating the transgression. But if a subse¬
quent erosion occurs, all sediments with their fossils
may be removed. In contrast, the thalassoid crusta¬
ceans—often called living fossils —settled during the
marine regression and later still dwelled in subterra¬
nean groundwaters. They may survive such erosive
destruction, as they are able to change habitat when
groundwaters, first embedded in superficial sedimen¬
tary rocks, shift into deeper and older rocks during a
regional uplift.
DATING MARINE REGRESSIONS
OR LAND EMERSIONS FROM THE
EXTANT STYGOFAUNA
It is noteworthy that the thalassoid limnostygobionts
may provide not only the evidence of an ancient sea in
the region where they occur, but that they can also, if
they belong to a well-known monophyletic group of
species, provide an indication about the date of the
marine regression that allowed the colonization of
ENCYCLOPEDIA OF CAVES
MARINE REGRESSIONS
485
FIGURE 2 Schematic map showing the geographic distribution of three species of thalassoid limnostygobionts Spl to Sp3, in relation to
three successive marine regressions. TR1SS to TR3SS are the seashore lines during the maximum of transgression at the periods Tl, T2, and
T3. Each transgression was followed by a marine regression as far as the present shore (PSH). PTSp is a present thalassostygobiontic species
still living in a coastal interstitial or cave environment. CLAD is a simplified cladogram showing the phylogenetic relationships within the
group of present stygobiontic species.
subterranean waters by their ancestral populations. The
above example in southwestern Morocco allows the
dating of two different marine regressions: one during
the Eocene, south of the present Souss Valley, and the
other during the Late Turonian in the Guelmim
Province.
The date of the most ancient exposed lands of Israel
and Palestine certainly is not Cenozoic, as generally
proposed. A species of Metacrangonyx belonging to a
lineage settled in continental subterranean waters dur¬
ing the marine regression of the Senonian (Late
Cretaceous) occurs near the Dead Sea, as well as in the
eastern Egyptian Sinai near the Eilat Gulf. Some geolo¬
gists considered the possible emergence of isolated
lands in this region during the Late Cretaceous from
indirect evidence, but these emerged lands or islands
are difficult to localize with precision, and in many
parts of Israel marine sediments indicate the presence
of the sea only from the beginning of the Cenozoic.
The characteristics of the two species of Metacrangonyx
known in the region indicate that these species are
part of a derived group of species that accomplished
the transition from a marine to a continental environ¬
ment during the Senonian. The suggested conclusion is
that the Late Cretaceous is the age of the most ancient
parts of the Levant (Boutin, 1997).
The emergence and the growing of an island, when
marine sediments are uplifted above the sea surface, is
an ecological event very similar to a marine regression
occurring on the margin of a continent. During the clo¬
sure of the eastern Mediterranean, all the lands of the
Levant (eastern Mediterranean countries) formed a
continuous emerged land. This was marked on its
margins by recent cycles of marine transgressions and
regressions. It appeared first as islands, and the great¬
est age of the land in these countries is that of the most
ancient emerged island.
Similarly, the age of the most ancient part of the
Canary Archipelago has been dated from the Senonian,
based on the study of the thalassoid amphipods occur¬
ring in wells of Fuerteventura Island (Boutin, 1994b). A
part of Fuerteventura is not volcanic, unlike the remain¬
der of the Archipelago; instead it is made of Cretaceous
sedimentary marine rocks forming a first emerged
island before the Cenozoic, while all studied volcanic
rocks are Cenozoic. This island emergence is quite simi¬
lar to a marine regression discovering lands above sea
level and at the same time developing available ground-
water habitats for marine crustaceans that are candi¬
dates for originating the thalassoid limnostygobiontic
fauna. It is noteworthy that all the wells housing the
Fuerteventura metacrangonyctid amphipod are dug in
ENCYCLOPEDIA OF CAVES
486
M
the metamorphic limestones of the "basal complex" and
never in the Cenozoic volcanic rocks. Metacrangonyx is
the sole evidence of the Senonian age of the emersion of
the most ancient part of the island of Fuerteventura.
EVALUATING THE RATE OF
EVOLUTIONARY CHANGES FROM THE
MARINE REGRESSION CALENDAR
Many observations made by paleontologists as well
as by neontologists suggest that the morphological
change in animal lineages is extremely slow in habitats
ecologically stable for millions of years, and rather
important and rapid in unstable habitats subjected to
repeated disturbances (Boutin and Coineau, 1991). As
for the molecular evolution, it is generally described
following the molecular clock model with the molecu¬
lar changes supposed to occur regularly and propor¬
tionally to the duration. But the average rate of
molecular changes fixed in generations forming a line¬
age is different according to the considered gene, or
molecule. Moreover, it is difficult to be sure that, dur¬
ing a long period, the fixation of occurring mutations
is always constant. This rate could be dependent on
repeated "bottleneck" occurring in animal populations
living in unstable habitats. Therefore it has been sug¬
gested (Boutin and Coineau, 2000) that it would be
especially interesting and informative to study, com¬
paratively, both at a molecular level and from a mor¬
phological point of view, some thalassoid
limnostygobionts belonging to different lineages of a
group that settled in continental groundwaters during
different marine regressions. As a matter of fact, after the
vertical transition the different lineages of interstitial lim¬
nostygobionts lived for a more or less long period in the
changing and unstable coastal biotopes before the hori¬
zontal transition to the more stable continental ground-
waters, which resulted from the repeated marine
regressions. Probably such comparisons will permit the
testing of molecular clocks and the respective value of
molecular and morphological characters for the recon¬
struction of phylogenies. These studies would be possi¬
ble when the dates of the successive marine regressions,
which originated the different lineages within a thalas¬
soid group of limnostygobionts, are well known
(Danielopol, 1980). The first studies with this aim have
just been performed and new results are expected by
stygobiologists.
Finally, the biogeography and the history of the tha¬
lassoid stygobiontic fauna are closely related to the
shifts of the shorelines, and the origin and some evolu¬
tionary aspects of these stygobionts result from marine
regressions.
See Also the Following Articles
Vicariance and Dispersalist Biogeogmphy
Invasion, Active versus Passive
Bibliography
Andreu-Boussut, B. (1991). Les ostracodes du Cretace-moyen
(Barremien a Turonien) le long d'une transversale Agadir-Nador
(Maroc). Strata, 24(Serie 2), 1—405 (in French).
Boutin, C. (1994a). Phylogeny and biogeography of metacrangonyc-
tid amphipods in North Africa. Hydrobiologia, 287, 49—64.
Boutin, C. (1994b). Stygobiology and historical geology: The age of
Fuerteventura (Canary Island), as inferred from its present stygo-
fauna. Bulletin de la Societe Ge'ologique de France, 165, 273—285.
Boutin, C. (1997). Stygobiologie et geologie historique: L'emersion
des terres de Mediterranee orientale datee a partir des
Amphipodes Metacrangonyctidae (Micro-Crustaces souterrains).
Geobios-Memoires Special, 21, 67— 74 (in French).
Boutin, C., & Coineau, N. (1990). Regression model, Modele biphase
devolution et origine des micro-organismes stygobies interstitiels
continentaux. Revue de Micropale'ontologie, 33, 303—322 (in French).
Boutin, C., & Coineau, N. (1991). Instability of environmental condi¬
tions and evolutionary rates. Example of thalassoid subterranean
microcrustaceans in Mediterranean countries. Bulletin de Vlnstitut
de Geologie du Bassin d'Aquitaine, 50, 63—69.
Boutin, C., & Coineau, N. (2000). Evolutionary rates and phyloge¬
netic age in some stygobiontic species. In H. Wilkens, D. C.
Culver & W. F. Humphreys (Eds.), Ecosystems of the world, 30:
Subterranean ecosystems (pp. 433—451). Amsterdam: Elsevier.
Choubert, G., & Faure-Muret, A. (1962). Evolution du domaine atlasi-
que marocain depuis les temps paleozoiques. In Livre a la memoire
du Professeur Paul Fallot (Tome 1) (pp. 447—527). Paris: Societe
geologique de France (in French).
Coineau, N., & Boutin, C. (1992). Biological processes in space
and time: Colonization, evolution and speciation in interstitial
stygobionts. In A. I. Camacho (Ed.), The natural history of biospe¬
leology (Monografias 7) (pp. 423—451). Madrid: C.S.I.C.
Danielopol, D. L. (1980). An essay to assess the age of the freshwater
interstitial ostracods of Europe. Bijdragen tot de Dierkunde, 50,
243-291.
Holsinger, J. R. (1994). Pattern and process in the biogeography of
subterranean amphipods. Hydrobiologia, 287, 131—145.
Holsinger, J. R. (2000). Ecological derivation, colonization, and speci¬
ation. In H. Wilkens, D. C. Culver & W. F. Humphreys (Eds.),
Ecosystems of the world, 30: Subterranean ecosystems (pp. 399—415).
Amsterdam: Elsevier.
Notenboom, J. (1991). Marine regressions and the evolution of groundwa¬
ter dwelling amphipods (Crustacea). Journal of Biogeography, 18,
437-454.
Stock, J. H. (1980). Regression model evolution as exemplified by the
genus Pseudoniphargus (Amphipoda). Bijdragen tot de Dierkunde,
50, 105-144.
MAYA CAVES
Andrea Stone and James E. Brady 1
University of Wisconsin-Milwaukee, ^California State
University, Los Angeles
Ancient Maya civilization extended across southern
Mexico, Guatemala, Belize, and western Honduras
Encyclopedia of Caves.
© 2012 Elsevier Inc. All rights reserved.
MAYA CAVES
487
FIGURE 1 Map of the Maya area with some important archaeo¬
logical cave sites.
(Fig. 1). Although this zone has a highland component,
it is the limestone-covered Maya Lowlands that saw
the rise of the great Classic cities, which are famous
for their hieroglyphic inscriptions, fine architecture,
and sculpture. The Lowland Maya offer an unusual
case study of one of the world's great civilizations
emerging in a tropical karst landscape. Moreover,
the prevalence of caves in the environment shaped
the development of the Classic Maya civilization.
Although caves served some practical ends, for
instance, sources of water in the Yucatan, they princi¬
pally provided religious sanctuaries and theaters of rit¬
ual activity. They also figured prominently in the
Maya's mental world of myth and symbol. In their
highly stratified society, the Maya's ritual use of caves
was an institution that cut across class boundaries.
This widespread cultural practice also had great antiq¬
uity, accompanying the rise of sedentary villages,
beginning in the Early Preclassic Period ( ca. 1200 b.c.).
Cave rituals became more elaborate during the Late
Classic fluorescence (600—900 a.d.) and continued in
northern Yucatan and the Maya Highlands through
the Postclassic Period (900—1550 a.d.). The fact that
cave ceremonialism continues today in the Maya area,
despite a legacy of colonial persecution, shows the cen¬
trality of caves in indigenous religion at the time of the
conquest.
GENERAL CHARACTERISTICS OF
ANCIENT MAYA CAVE UTILIZATION
The ancient Maya explored caves with the aid of
wooden torches. Bundles of split pine torches have
been recovered in cave entrances and burned remains
recovered in deeper passages. Charcoal torch strikes
on cave walls are also common. One of the few cave-
specific tools developed by the Maya is a ceramic tube
that served as a torch holder. Unlike Paleolithic cave
explorers who penetrated equal depths by firelight,
the ancient Maya never used oil lamps. Light appears
to have been a limiting factor in their speleological
exploits. Their record for penetration of a cave is 3 km,
set at Actun Chek, Belize. The longest caves in the
Maya area are the Chiquibul and Cave's Branch cave
systems of Belize, where individual caves reach 15 km
in length. Most caves explored by the ancient Maya
are more modest; however, small size did not preclude
their designation as sacred sites. Even the smallest
caves, cenotes, and rockshelters were ritually utilized.
Although the Maya generally explored the deepest
parts of caves, the heaviest concentration of artifacts is
usually found in the light to twilight zones near
entrances. Forays into the dark zone may have been
restricted to ritual specialists.
ANTIQUITY OF MAYA CAVE USE
Evidence for cave utilization in the Maya area
prior to the advent of sedentary villages, circa 1200 b.c.,
is extremely limited. The discovery at Loltun Cave,
Yucatan, of stone tools associated with extinct megafauna
offers some evidence for use during the Late Pleistocene.
Charcoal from a hearth in Carwash Cenote, Quintana
Roo, yielded a radiocarbon date of 8250 ± 80 years BP.
The picture is equally dim during the Archaic period but
begins to clarify during the Preclassic. At Copan,
Honduras, ceramic evidence from Gordon's Cave #3
indicates its use as an ossuary as early as 1000 b.c. Major
architectural modifications at the Cueva de las Pinturas,
Guatemala, have been dated to the end of the Middle
Preclassic (300 b.c.), and stone masonry at Naj Tunich,
Guatemala, dates to the Late Preclassic. An elaborate
figure carved at an entrance to Loltun Cave suggests that
the site was appropriated for elite use by the Late
Preclassic.
CAVE BURIAL
Human skeletal material in caves reflects both the
disposal of sacrificial victims and the burial of loved
ENCYCLOPEDIA OF CAVES
488
M
ones. Rockshelters were often used as cemeteries in the
southern Maya Lowlands from at least the Late
Preclassic (300 b.c. to 300 a.d.) to the end of the Classic
Period (900 a.d.). Important individuals were buried in
cave alcoves that were closed off with crude stone
walls. This practice appears to be confined to the
Preclassic except at Naj Tunich, which is unique in
having elaborate masonry tombs dating to the Classic
period.
A number of cave ossuaries have been reported,
generally in the Highlands, although the data are
sketchy. On the eastern edge of the Maya area, Gordon's
Cave #3 near Copan contains hundreds of cremated
remains from roughly 1000 b.c. that appear to have been
brought to the cave in cloth bags. On the western edge
of the Maya area in Chiapas, small, walled-up caves
yielded sealed ceramic vessels that contained cremated
human remains dating to the Postclassic (900—1550 a.d.).
Ethnohistorical sources from Chiapas mention the
worship of bundles in caves containing the remains of
ancestors, and ethnographic sources suggest that impor¬
tant males were interred in lineage caves up until the
close of the nineteenth century. If, as these sources sug¬
gest, the ossuaries are associated with ancestor worship,
it is interesting to note that many Maya still believe that
ancestors live within the Earth.
CAVE MODIFICATIONS
Almost all Maya caves show extensive breakage of
speleothems. Often broken formations have been
moved or taken from the cave, and speleothems are
frequently reported in cultural deposits at surface sites.
Modern indigenous terms for speleothems identify
them as congealed water or dripping water turned to
stone. Because water is so essential to agriculture, spe¬
leothems are directly related to fertility, and today
they are placed on household altars. Similar ideas
were probably held by the Classic Maya.
The Maya deployed various types of architectural
constructions to modify space within caves. Among the
most common are stone walls defining enclosures.
Walls with low doorways intentionally impede access
by forcing visitors to enter on hands and knees.
Retaining walls and dirt fill were used to create level
spaces. Plaster floors and stone pavements also demar¬
cated special areas. Altars, often nothing more than a
crude stack of stones, are frequently reported (Fig. 2).
Stairways leading down into caves appear most fre¬
quently in Yucatan and Quintana Roo. Dams have been
reported in several caves with seasonal flooding.
Finally, small architectural structures, probably reserved
for the most private rituals, were occasionally built
inside caves.
FIGURE 2 Altar-like construction from Actun Kabal, Belize. Photo
by Andrea Stone.
CAVE ART
With about 60 known decorated caves, the Maya
Lowlands have one of the most important cave art tra¬
ditions in the world. Pigment-based art consists of
paintings, drawings, and both positive and negative
hand- and footprints, often utilizing charcoal and clays
available in the cave. However, prepared mineral pig¬
ments, such as red (from hematite), blue, and yellow
occur in rare instances. The most typical carvings,
mainly found in Yucatan, Quintana Roo, and Belize,
are deeply engraved petroglyphs showing frontal
faces, meanders, and geometric elements. Another cat¬
egory of sculpted art is the modified speleothem, typi¬
cally shaft-shaped speleothems with crudely carved
faces (Fig. 3). These may represent spirit beings
embodied in stone. Artistic modification of spe¬
leothems was widely practiced among the Classic
Maya. Rare examples of painted petroglyphs and
sculptures modeled from clay are also known.
Although the sculpted art tends to be crude, some of
the painted art is refined and resembles elite art. This is
epitomized by the paintings at Naj Tunich, Guatemala,
dating to the seventh and eighth centuries. Both hiero¬
glyphic inscriptions and depictions of human figures,
which evince elite cave utilization, are present. The hier¬
oglyphs reveal that political relations among regional
sites were mediated by ritual cave use and that some
caves, such as Naj Tunich, held higher status than others
as regional pilgrimage destinations. The painted cave art
has been the subject of technical analysis including AMS
radiocarbon dating and multispectral imaging.
ETHNOHI STORY AND ETHNOGRAPHY
Colonial documents and modern ethnographic stud¬
ies of the Maya aid in the reconstruction of Classic
ENCYCLOPEDIA OF CAVES
MAYA CAVES
489
Maya cave use. An important colonial source is the
Popol Vuh, a K'iche' Maya text recounting the adven¬
tures of supernatural twins in an underworld realm
called Xibalba, "place of fright," a thinly veiled allu¬
sion to a cave. The Popol Vuh reveals thematic associa¬
tions of caves with danger, regeneration, and the
ballgame.
In indigenous cosmology the first appearance of
humans on the face of the Earth is from a cave, the
womb of the Earth. Thus, caves are the primary sym¬
bol of human and world creation. Not surprisingly,
caves play an important role in conceptions of sacred
geography, definitions of territorial boundaries, and
pilgrimage routes in today's Maya communities and
probably did so in the past.
As penetrations into the sacred Earth, caves are
the residences of ancestors and supernatural beings.
The most important indigenous deity in the Maya
Highlands today is known by a name that translates as
"Hill-Valley." He/she controls the fruits of the Earth
and is petitioned in caves. The cave's association with
water is equally important among contemporary
groups, many of whom believe that storms emanate
from caves. In dry areas, such as Yucatan, the presence
FIGURE 3 Modified speleothem from a cave in Guatemala.
Drawing by Anne Chojmcki.
of a cave or cenote influenced the location of settle¬
ments. Cave water was ritually pure and in Yucatan
was called suhuy ha or "virgin water." Archaeological
evidence of ritual activity near pools of water in caves
and the placement of ceramic vessels and stone
troughs under drips, often in remote areas, are testi¬
mony to the antiquity of this belief.
CAVES AND COMMUNITY
The identity of Maya communities with caves is so
close that the cave often gives its name to the commu¬
nity. In colonial rituals at the foundation of a new com¬
munity, the presence of a cave was considered
essential and lent authority to settlers' claims to
occupy and govern a territory. Because the cave sym¬
bolized political legitimacy in Mesoamerica, it is not
FIGURE 4 Artificial cave dug in volcanic ash under the central
plaza of the Pre-Columbian settlement of Utatlan, Guatemala. This
cave is an important religious site among contemporary highland
Maya. Photo by James Brady.
ENCYCLOPEDIA OF CAVES
490
M
surprising that the pre-Hispanic Maya elite incorpo¬
rated caves into their ritual activity and their elaborate
architecture. Caves have been found to underlie build¬
ing complexes, including pyramids, at such sites as
Chichen Itza, Mayapan, Tulum, Polol, and Dos Pilas.
Caves were so important that, in the nonkarstic area
of the Maya Highlands where caverns do not do occur
naturally, artificial caves were excavated, often beneath
pyramids or other architecture. Such tunnels have
been found at the sites of Utatlan, Mixco Viejo,
Zaculeu, Iximche, and La Lagunita in Guatemala
(Fig. 4). The fact that artificial caves have also been
reported at the Central Mexican sites of Teotihuacan,
Xochicalco, and Acatzingo Viejo suggests that caves
were a fundamental concern throughout Mesoamerica.
found in other parts of the biosphere (e.y., Whitman
et al., 1998). Reactive interfaces among air, water, and
rock offer suitable habitats for microbial activity,
including within cave and karst landscapes. As such,
cave and karst environments serve as reservoirs for
diverse microbes over extended periods of time. This
article describes the history of microbiological research
in karst, current methodological approaches being
used to understand microbial diversity and ecology, as
well as the roles of microbes in important geologic and
geochemical subsurface karst processes.
MICROBIAL DIVERSITY
AND METABOLISM
See Also the Following Articles
Ancient Cavers in Eastern North America
Cave Dwellers in the Middle East
Bibliography
Brady, J. E. (1997). Settlement configuration and cosmology: The role
of caves at Dos Pilas. American Anthropology, 99(3), 602—618.
Brady, J. E. (1999). Studies in Mesoamerican cave use. Sources for the
study of Mesoamerican ritual cave use (2nd ed. (Publication 1)). Los
Angeles: California State University.
Garcia-Zambrano, A. J. (1994). Early colonial evidence of Pre-
Columbian rituals of foundation. In M. G. Robertson & V. Field
(Eds.), Seventh Palenque Round Table, 1989 (pp. 217—227). San
Francisco, CA: Pre-Columbian Art Research Institute.
McNatt, L. (1996). Cave archaeology of Belize. Journal of Cave Karst
Studies, 58(2), 81-99.
Stone, A. J. (1995). Images from the underworld: Naj Tunich and the tradi¬
tion of Maya cave painting. Austin: University of Texas Press.
Stone, A. J. (1997). Pre-Columbian cave utilization in the Maya area. In
C. Bonsall & C. Tolan-Smith (Eds.), The human use of caves (BAR
International Series 667) (pp. 201—206). Oxford, England: Archaeopress.
Thompson, J. E. S. (1975). Introduction to the reprint edition. In H. C.
Mercer (Ed.), The hill-caves of Yucatan. Norman: University of
Oklahoma Press.
Vogt, E. Z. (1981). Some aspects of the sacred geography of Highland
Chiapas. In E. P. Benson (Ed.), Mesoamerican sites and world views
(pp. 119—142). Washington, DC: Dumbarton Oaks Research
Library and Collection.
MICROBES
Annette Summers Engel
University of Tennessee
INTRODUCTION
The Earth's continental subsurface contains abundant
and active microbial biomass that exceeds the biomass
Microorganisms in cave and karst settings have vary¬
ing metabolism. Microbial metabolism refers to how a
microbe obtains carbon, along with other nutrients, and
gains energy (Fig. 1). Carbon for cellular growth origi¬
nates by either converting inorganic carbon (CO 2 ,
HCO 3 ) to organic carbon as an autotroph, or assimilat¬
ing organic carbon initially produced by autotrophs.
Heterotrophs use already existing organic carbon for cel¬
lular energy and their carbon sources. The physiologi¬
cal mechanisms for capturing chemical energy are
diverse, and the distinction between a chemosynthetic
c
CO,
+ COo
j
1 ^
Aerobic pathway Anaerobic pathway
Oxic
T
Anoxic
FIGURE 1 Schematic representation of energy-yielding metabolic
reactions. Aerobic microorganisms, such as iron-oxidizing bacteria,
use dissolved oxygen (0 2 ) to generate energy. In contrast, anaerobes
use alternative terminal electron acceptors. For methanogenesis,
which occurs under the most reducing conditions, C0 2 is the elec¬
tron acceptor. C0 2 is required for all chemolithoautotrophic meta¬
bolic reactions, and H 2 is required by most anaerobes. Aerobic and
anaerobic metabolism can occur simultaneously, although there is
usually spatial and temporal separation of the reactions within a
microbial habitat. For instance, aerobic microorganisms colonize
uppermost portions at the air-water interface, while anaerobes
occupy inner regions of a mat or water-sediment interface.
Encyclopedia of Caves.
© 2012 Elsevier Inc. All rights reserved.
MICROBES
491
and a photosynthetic organism is based on whether the
initial source of energy is from inorganic chemicals
(litho ) or light (photo). Microbes that gain energy
through chemosynthesis and fix inorganic carbon are
chemolithoautotrophs (literally "self-feeding rock-eaters").
During chemosynthesis, microorganisms gain energy by
transferring electrons from one chemical (electron donor)
to another (electron acceptor) that can come from
groundwater or colonized rock and mineral surfaces.
Chemical electron donors can include, but are not lim¬
ited to, molecular hydrogen or reduced sulfur com¬
pounds; organic molecules ( organo ), such as acetate or
formate, can also be used. Chemolithoautotrophs can
also grow if organic carbon is present as mixotrophs, in
which both chemolithoautotrophy and heterotrophy are
expressed simultaneously. However, an organism is no
longer classified as a chemolithoautotroph if organic
compounds are used to gain energy or for a carbon
source. Organisms that gain cellular energy from chemi¬
cal transformations but use organic carbon compounds
for their carbon source are chemoorganotrophs.
Microorganisms are well suited to live in geo-
chemically or mineralogically extreme environments.
Extremophiles can be adapted to live in conditions that
are nutrient poor ( oligotrophic ), low pH ( acidophilic ), high
pH ( alkaliphilic ), high temperature ( thermophilic ) or low
temperature ( psychrophilic ), or sulfide-rich (sulfidophilic).
Classification of microbial metabolism is also based on
oxygen requirements and whether a microbe respires
aerobically, anaerobically, or ferments, all of which relates
to electron acceptor utilization (Fig. 1). Microorganisms
that require oxygen are aerobes, and oxygen serves as the
terminal electron acceptor for metabolic processes yield¬
ing energy, such as from the oxidation of reduced sulfur
compounds (e.g., hydrogen sulfide, thiosulfate), iron oxi¬
dation, or ammonia oxidation. In more reducing environ¬
ments, where organic carbon is rapidly consumed,
microbes that do not require oxygen (anaerobes) use alter¬
native electron acceptors for respiration in a sequence of
energetic, reduction reactions that occur along thermody¬
namic (and redox) gradients. After oxygen, electron
acceptors are reduced in the following order: NO 3 —>
Mn 4+ - Fe 3+ -> SO 2 - C0 2 .
Most anaerobes are obligate, but microaerophilic
organisms require low oxygen concentrations, while
facultative organisms grow in the presence or absence of
oxygen, using different, and the most energetic, elec¬
tron acceptors available. Some important anaerobic,
redox reactions include nitrate or dissimilatory nitrogen
reduction (N0 3 —► N 2 ), ferric iron reduction (Fe 3+ —>
Fe 2+ ), dissimilatory sulfate reduction (S0 4 —> H 2 S),
and methanogenesis (C0 2 —> CH 4 ). Microorganisms
involved in these processes can grow as chemolithoau¬
totrophs or chemoorganotrophs. There are several other
energetically favorable pathways that occur in the
absence of oxygen, including sulfide oxidation via
nitrate reduction and anaerobic ammonium oxidation
(anammox).
HISTORY OF MICROBIOLOGY STUDIES
The earliest microbiological studies from caves used
microscopy and culturing for specific metabolic
groups. Microbes were found to be prevalent in sedi¬
ments and water. By the early 1900s and through
the 1940s, many cave deposits, such as cave nitrates
(saltpeter or saltpetre) (e.g., Faust, 1949), carbonate
moonmilk, and other speleothems, were considered to
have a microbiological origin. Chemolithoautotrophic,
iron- and sulfur-oxidizing bacteria were identified
and sulfur-oxidizing bacteria were implicated in the
sulfuric acid-promoted dissolution of limestone (e.g.,
Principi, 1931). Incidentally, it has only been within
the past 5—10 years that modern methodology has
been able to prove that some of these deposits and spe-
leogenetic processes are the result of microbial activity.
By the mid-1960s, the ecological and geological roles
of microbes in caves started to come into focus (e.g.,
Caumartin, 1963; Symk and Drzal, 1964). However,
based on what was known of microbial diversity by
the 1960s through the 1980s, which was predominantly
from culture-based approaches, microbial communities
in the subsurface were considered to be a subset of
surface, such as soil, communities that were flushed
underground by meteoric drip waters, surface streams,
or air currents, or were carried by animals into
caves. All life in caves was thought to be dependent
on photosynthetically produced organic matter, and
therefore, life in caves was considered to be nutrient-
limited. Consequently, the primary ecological function
of microbes in caves was as heterotrophs, that
degraded organic matter and were food sources for
higher organisms. Because of oligotrophic conditions,
microbial biomass was considered to be low and insuf¬
ficient to impact most geological or geochemical pro¬
cesses, except in specialized habitats such as in sulfidic
aquifers or metalliferous sediments. There was some
recognition that laboratory enrichment and strain isola¬
tion by culturing was perhaps a problem because cul¬
turing selected for fast-growing rather than slow-
growing groups, and most cave habitats were found to
be oligotrophic.
Microbial primary production based on chemosyn¬
thesis, rather than photosynthesis, was introduced in
the late 1980s from the discoveries of the deep-sea
hydrothermal vents. This research coincided with the
development and application of molecular genetics
techniques in the 1980s and 1990s, and provided
timely release from the earlier, one-sided ecological
ENCYCLOPEDIA OF CAVES
492
M
views of microbes in ecosystems. For cave and karst
research, work in the early to mid-1990s from the
Movile Cave, Romania (Sarbu et ah, 1996), brought the
significance of microbes in chemosynthetically based
ecosystems into focus. Since then, it is clear that micro¬
organisms can have tremendous biomass and meta¬
bolic diversity in the subsurface, and that many of the
metabolic processes are not dependent on sunlight.
Cave and karst microbiological studies have kept pace
with technology, and are now based on applying geo¬
chemical models to genetic diversity data resulting from
automated DNA sequencing, high-throughput comput¬
ing analyses of environmental rRNA and functional gene
sequence phylogenies, or from oligonucleotide rRNA
probing, microautoradiography, and stable-isotope prob¬
ing. All of these methods evaluate biochemical diversity
and the evolution of genes, enzymes, and metabolites
(e.g., Hutchens et ah, 2004; Chen et ah, 2009; Dattagupta
et ah, 2009). Essentially, research efforts over the past
40 years have gone from answering the basic question,
"Who's there?," to a more difficult question to answer,
"What are they doing?"
METHODOLOGICAL APPROACHES
Techniques are rapidly changing and being improved.
Virtually any method has some type of error associated
with it, which should be considered when interpreting
results. The approaches described herein are a sampling
of possible strategies that could be used to understand
"Who's there?" and "What are they doing?" Without
question, a combination of different methods, while not
always feasible, still provides the best results.
Culturing and Metabolic Assays
Culturing involves growing microbes in the labora¬
tory by providing cells with all the essential nutrients
needed for growth, including electron donors and
acceptors. Even when the metabolic requirements are
known (or guessed), however, enrichment and pure
culture isolation are difficult. In fact, most microorgan¬
isms in nature have not been grown in culture, and it
has been estimated that less than 1 % of known
microbes are culturable with current techniques.
Standard culturing often introduces a selective bias
toward microorganisms that are able to grow quickly or
to utilize substrates provided in the medium more effi¬
ciently. Nevertheless, culturing, although highly selec¬
tive and difficult for unknown or poorly characterized
microorganisms, is still a reliable way to determine
metabolic capabilities of certain microbial groups. This
method is impractical for diverse communities, as it is
prohibitive to attempt to grow all possible microbial
metabolic groups that may be present in an ecosystem.
If pure or mixed cultures are obtained, then micro¬
organisms can be characterized or identified based on
a variety of growth-dependent assays. Biochemical
assays can also identify activities in natural samples
without cultures. Assays include, but are not limited
to, measuring the presence or absence of enzymes
involved in degradation of substrates in the growth
medium, production of gases (e.g., H 2/ C0 2 , H 2 S), or
the utilization of specific carbon substrates, and deter¬
mining the ability to grow under different physico¬
chemical conditions, such as growth over a range of
pH values or temperatures, or even under varying
oxygen concentrations. For many organisms, a few key
assays are all that is required for identification or char¬
acterization. Some more sophisticated assays include
measuring enzymatic activity or membrane lipids such
as phospholipid-derived fatty acids.
Molecular Methodologies
Molecular techniques that detect nucleic acids can cir¬
cumvent culture-based problems and provide a culture-
independent assessment of which microorganisms are
present in a community based on the evolutionary rela¬
tionships of gene sequences. To characterize microbial
taxonomy, most researchers have evaluated the 16S
rRNA gene sequence diversity using Sanger-sequencing
methods. This approach achieves varying levels of suc¬
cess depending on how many sequences are generated
from a sample and the community diversity; sample
coverage is generally high for low diversity materials,
but samples having high microbial diversity can have
very low coverage. Resulting gene sequences are evalu¬
ated in a phylogenetic framework to categorize groups.
An alternative metagenomics approach has been devel¬
oped recently that allows for evaluating exceptionally
high volumes of gene sequences by using next-genera¬
tion pyrosequencing chemistries, including an approach
referred to as 454 sequencing that involves high-through-
put sequencing platforms (e.g., Roche/454 Life Sciences)
(Marsh, 2007).
There are numerous well-known disadvantages to
these methods, including the original DNA extraction
methods that may bias microbial nucleic acid represen¬
tation in a sample. Other disadvantages are because
the methods rely on the polymerase chain reaction
(PCR) to amplify sequences from DNA, so PCR inhibi¬
tors within a sample, genomic complexity in a sample,
primer choice, chimera generation, and the interpreta¬
tion of multiple gene copies, can all affect the experi¬
mental results. Probably the most significant
disadvantage is that the phylogenetic affiliations that
ENCYCLOPEDIA OF CAVES
MICROBES
493
can be assigned are based on what is already known.
If a unique habitat or microbial community is exam¬
ined, then phylogenetic assignments of unknown
microbes may be weak or incorrect. Another signifi¬
cant disadvantage, especially in terms of understand¬
ing the metabolisms of a microbial population, is that
phylogeny rarely equals function. So, metabolic activ¬
ity must be inferred, if not tested separately, from
additional ecological, culturing, or molecular methods.
In recent years, and based on whole genome com¬
parative analyses from an ever-increasing number of
microbes, the characterization of functional genes is
becoming more important. If genes are sequenced
from DNA, then this allows for an evaluation of the
potential for a specific pathway to be done by certain
microbes. However, if genes are sequenced from
mRNA, then this allows for an evaluation of the tran-
scripted information, or more specifically, an evalua¬
tion of gene expression. This transcriptomics approach
can obtain key functional gene information for a com¬
munity without inferring metabolism from other
DNA-based methods.
Microarrays, which are microchips with attached
nucleotide probes, can also be used to study metabolic
activity and microbial diversity in complex systems
(Guschin et al., 1997). Probes can be designed for func¬
tional genes, as well as for species-, genera-, group-, or
domain-specific targets. The number of probes used in
an array is only limited by the size of the microchip.
Microarray methods circumvent PCR-based problems,
and can be done with DNA to examine gene presence or
with mRNA to examine transcripted activities. Although
the utility of microarrays is seemingly limitless, at pres¬
ent there are no cave or karst studies incorporating this
technique.
Stable Isotope Ratio Analyses and Labeled
Substrate Experiments
Microbial activity and ecosystem energetics in natu¬
ral or cultured samples can be evaluated from ratios of
stable isotopes in the sample compared to the surround¬
ings, or from using labeled substrates, either radioiso¬
topes or stable isotopes, such that both carbon fixation
and energy-conserving metabolism can be quantified.
These experiments can provide activity rate estimates
without knowing how many organisms are in the sam¬
ple, and allow a sensitive, quantitative assessment of
biogeochemical transformations in nature (Boschker
and Middelburg, 2002).
Stable isotope ratio analysis methods depend on the
creation and alteration of compounds by biotic and abi¬
otic processes. For example, the two carbon isotopes of
importance in studying carbon cycling are 12 C and 13 C.
Incorporation of carbon into living tissues invokes
significant kinetic isotope fractionation, such that bio¬
logical processes discriminate for the lighter isotope
( 12 C), leaving behind the heavier isotope ( 13 C).
Discrimination for 12 C is due to kinetic effects caused
by the irreversible enzymatic CC^-fixing reaction that
assimilates CO 2 into the carboxyl group of an organic
acid. Differences in the isotopic composition are
expressed in terms of the delta ©-notation of a ratio of
13 C/ 12 C in a sample relative to a standard, measured in
per mil (%»)• Biogenic carbon is isotopically lighter than
the inorganic reservoir, such as atmospheric CO 2 or dis¬
solved bicarbonate (HC0 3 ), and autotrophic fixation
pathways have some of the largest fractionation effects
during biological assimilation of carbon. The overall
isotopic fractionation is determined by subtracting the
carbon isotopic composition of an organism from the
isotopic composition of the carbon reservoir, with most
autotrophic fractionation values ranging between —20
and —40%o. Excretion, respiration, and heterotrophic
carbon cycling are (for the most part) considered negli¬
gible processes for carbon fractionation, and the isoto¬
pic composition of heterotrophic organic matter will be
the same or slightly higher than the source organic car¬
bon. Due to the diversity in carbon-fixation pathways,
as well as the level of recycling within a microbial detri-
tal loop, there can be a wide range in carbon isotope
values within an ecosystem.
The rate of certain metabolic processes can be deter¬
mined by measuring the uptake of one or more com¬
pounds over time and the subsequent buildup of other
compounds, such as metabolic by-products. Electron
acceptors are the most commonly radiolabeled sub¬
strates. For instance, sulfate reduction can be monitored
using radiolabeled 35 SC> 4 ~ compounds, whereby loss
of 35 SOf would indicate consumption, and a gain of
35 S 2_ as the by-product of sulfate reduction would pro¬
vide sufficient evidence that sulfate reduction is occur¬
ring. Carbon fixation rates can also be determined by
labeling carbon molecules, such as bicarbonate or ace¬
tate, with 14 C, and the evolved 14 C0 2 and 14 C-biomass
are measured. Even when multiple radiolabeled sub¬
strates are combined in an experiment, however, there
are several problems with carbon radioisotope studies,
with the most significant being how to decipher the
results. Carbon fixation or heterotrophic consumption of
organic matter can be overestimated or underestimated
due to sampling perturbation or poor understanding of
the community structure. Because natural samples can
be extremely sensitive to disturbance, the actual experi¬
ment may skew results and demonstrate that one meta¬
bolic process is dominant, when in fact, that process is
just less sensitive to disturbance than other processes.
Additionally, some microbes in a sample could be capa¬
ble of facultative or mixotrophic growth, acting as che-
molithoautotrophs in nature, but when provided with a
ENCYCLOPEDIA OF CAVES
494
M
rich organic carbon molecule, they will use it as che-
moorganotrophs or heterotrophs. Detailed culturing or
phylogenetic analyses of the microbial communities
should be coupled to radioisotope studies to aid inter¬
pretation. Recovering and sequencing the DNA or RNA
from the experimental biomass that incorporated the
labeled substrates can be done to investigate the active
microorganisms of interest as well.
New Uses for Microscopes
Cell morphologies and structural relationships
between cells and inorganic substrates can be visualized
directly by light microscopy or transmission or scanning
electron microscopy (TEM or SEM, respectively). But,
although microscopy can provide a wealth of information,
such as microbial mat structure and the nature of biomin¬
eralization, microscopy is not a reliable technique to iden¬
tify microbes, even with morphologically conspicuous
cells. General microscopy can quantify biomass, but does
not provide any information about metabolism, although
there are basic stains or dyes that can identify actively
metabolizing cells. The best use of microscopy is to com¬
bine it with geochemical, ecological, and microbiological
methodology, such as fluorescence in situ hybridization
(FISH). FISH probes are similar to those used in microar¬
rays, with a majority of them being based on 16S rRNA
targets. This method combines fluorescence microscopy
with oligonucleotide fluorescent probes. For low abun¬
dance or low rRNA content, catalyzed reporter deposition
(CARD) combined with FISH can enhance signal fluores¬
cence for group quantification. FISH combined with
microautoradiography (FISH-MAR) identifies not only
the microbe, but also what the microbe may be doing by
allowing for the incorporation of a labeled substrate
(Ouverney and Fuhrman, 1999). New methods have been
developed that utilize a high-resolution ion microprobe
using secondary ion mass spectroscopy (SIMS) technique
combined with HSH. At the single-cell level, this
approach simultaneously detects the probe hybridization
to inform of phylogeny, as well as the isotopic signal from
biomass incubated with labeled substrate (Li et ah, 2008).
MICROBIAL DIVERSITY OF CAVES
AND KARST
Current research from karst attempts to understand
the abundance, diversity, and activities of microbial
communities, their ecosystem function, the connection
between microbes and geological or geochemical pro¬
cesses. This section highlights some of the implications
of microbial diversity and distribution in karst sys¬
tems, based primarily on the retrieval of 16S rRNA
gene sequences retrieved from caves and karst. From
about 20 years of data (e.g., Engel and Northup, 2008;
Engel, 2010), roughly a third of the recognized bacte¬
rial phyla have been identified from caves, and slightly
less than half of the recognized archaeal phyla (Fig. 2).
Most of what is known of bacterial and archaeal diver¬
sity is from just a few systems worldwide, and from
focusing on specific types of caves or microbial groups
(Engel and Northup, 2008). The diversity of microbial
eukaryotes is limited to a few examples for each major
group (Fig. 2). We know much more about these
organisms from systematics rather than genetics,
despite the fact that some, such as fungi and algae, can
form structures seen by the naked eye (Fig. 3).
From these different molecular and culture-based
studies, the abundance and distribution of microbes
in caves and karst settings have to do with, at a
fundamental level, how a habitat originated and
the source(s) of formational fluids, including epigenic
(i.e., sourced from the surface) or hypogenic ( i.e.,
sourced from depth). In general, epigenic systems
have been studied more intensely, likely because these
are the most common cave type ( e.g ., Palmer, 2007).
Moreover, each major habitat zone (i.e., entrance, twi¬
light, and dark zone) has general physicochemical and
nutrient conditions that control the types of organisms
present. For instance, dry, air-filled passages can have
nearly constant temperatures and stable geochemical
conditions, but cave streams or underwater passages
are directly influenced by hydrology that can cause
seasonal or even episodic habitat variability. Subtle
physicochemical variability, speciation and availability
of redox-sensitive elements, and the types and loading
of organic carbon and other nutrients to a zone, further
influence the diversity and metabolism of (micro)
organisms that will colonize cave and karst habitats.
The juxtaposition of stable physicochemical conditions
with extreme geochemistry and physicochemical con¬
ditions has provided interesting microbiological res¬
earch avenues (Barton and Northup, 2007).
Probably the most recognizable macroscopic microbio¬
logical feature on moist cave walls in epigenic caves are
shiny gold, white, pink, or brown patches of large colo¬
nies of filamentous Actinobacteria, commonly and histor¬
ically referred to as actinomycetes (Fig. 4A,B,C). Water
condenses onto the cells, which are more hydrophobic
than the cave walls, thereby causing a sparkly appear¬
ance. Humid lava tubes also contain Actinobacteria, as
well as members of the phyla Chloroflexi, Verrucomicro-
bia, and Proteobacteria. The phylum Actinobacteria
includes the genera Streptomyces, Mycobacterium, and
Nocardia, and some species can be pathogenic. Some Acti¬
nobacteria produce secondary metabolites that are useful
antibiotic and antimicrobial agents. One secondary
metabolite of Streptomyces is geosmin, the substance that
causes the earthy smell common to many caves.
ENCYCLOPEDIA OF CAVES
MICROBES
495
FIGURE 2 Based on Engel (2010), and the represen¬
tation of microbial phyla, indicated by stars (★), from
cave and karst systems within the domains: (A) bacte¬
ria, (B) archaea, and (C) eukaryota. The bacterial and
archaeal topologies are based on trees created using the
NAST alignment of 16S rRNA gene sequences from
Greengenes (DeSantis et al., 2006). Bars indicate
changes per nucleotide. The schematic phylogeny for
eukaryotes is based on the topology from 18S rRNA
gene sequences, presented by Dawson and Pace (2002).
Another large group of microbes described from cave
and karst settings belongs to the phylum Proteobacteria,
which is subdivided into classes, including the Alplwproteo-
bacterin, Betaproteobacteria, Gammaproteobaderia, Deltaproteo-
bacteria, and Epsilonproteobaderia (Fig. 2). Iron bacteria from
the Betaproteobacteria, including the chemolithoautotrophic
Gallionella spp. and heterotrophic Leptothrix spp., are com¬
mon in the allochthonous sediments of noncarbonate rocks
in many epigenic caves; the sediments can contain higher
ferromanganese content compared to the carbonate host
rock. Iron- and manganese-oxidizing bacteria have also
been identified from corrosion residues on cave-wall sur¬
faces from Lechuguilla Cave and other caves in the Guada¬
lupe Mountains of West Texas and New Mexico. Among
the diverse bacteria, the ferromanganese deposits also con¬
sist of Crenarchaeota and Euryarchaeota, which have also
ENCYCLOPEDIA OF CAVES
496
M
FIGURE 3 (A) Entomopathogenic fungi on Triphosa dubitata;
moth ~2.5 cm long, Planinska Jama, Slovenia. (B) Fungi growing on
organic debris; acorn cap for scale, approximately 2 cm across, Sandy
Cave, Kentucky. (C) Algae, cyanobacteria, and other phototrophic
organisms growing in the entrance zone.
been retrieved from paleofill sediment deposits in Wind
Cave in South Dakota. Poorly crystalline manganese oxide
and hydroxide minerals in caves have been attributed to
microbial processes. Encrusted sheaths and stalks resem¬
bling modem iron-oxidizing bacterial cells have been seen
from inside stalactites, known as msticles, as well as within
lithified sediments, thick mineral cmsts, and boxwork
from numerous caves, including those in the Black Hills,
South Dakota.
Proteobacteria are also prevalent in active hypo-
genic caves and karst systems, especially sulfidic caves
(Engel, 2007). The classes most commonly identified in
sulfidic systems include the Epsilonproteobacteria, Gam-
maproteobacteria, and Betaproteobacteria, which consist of
many different types of sulfur-oxidizing bacteria that
are essential to the development of chemolithoautotro-
plrically based ecosystems. These bacteria form the
bulk of paper-thin, floating microbial mats and thick,
filamentous microbial mats in streams and pools in
active caves (Fig. 5A,B,C).
They may have played a role in the formation of the
biogenic speleothem pool fingers found from caves
such as Lechuguilla Cave and Carlsbad Caverns
among others. In active sulfidic caves, these microbes
contribute to sulfuric acid speleogenesis. Cave-wall
surfaces are another microbial habitat in active sulfidic
caves (Fig. 5D). Cave walls can be acidic, due in part
to gypsum mineralization and microbial activity, and
have condensation droplets of pH <2. Interestingly,
the phylogenetic diversity of cave-wall biofilms is sim¬
ilar from active caves around the world, consisting of
various members of the Proteobacteria, Actinobacteria,
and bacterial candidate divisions, especially TM6, as
well as filamentous fungi and protists.
Only recently has the microbial diversity of vermicu-
lations, also known as biovenniculations, been investi¬
gated (Fig. 4D). These features are broad areas of
discontinuous, geometric deposits of organic matter
intermixed with clay on cave-wall surfaces, and are
found in many different types of caves, including lava
tubes, as well as mines. Since the 1960s, they have been
considered (micro)biological. The dominant bacteria in
biovermiculations from the sulfidic Frasassi Caves, Italy,
include Betaproteobacteria, Gammaproteobacteria, as well as
members of the phyla Acidobacteria, Nitrospirae, and
Planctomycetes (Jones et al., 2008). Mites and nematodes
have also been described. The causes of the geometric
structure of biovermiculations are still mysterious, but
the patterning is being used by astrobiologists as a way
to identify life processes.
Knowing the phylogenetic composition of microbial
communities is useful to understand what microbes
may be doing in, and to, caves, including being able to
determine the health of a cave ecosystem from the
microbial community composition. But the problem
comes in interpreting the presence and absence of
microbial groups in communities. Does absence mean
that the microbes are not present in a system, or that
the sampling and analytical methods did not detect
them? Moreover, there is a big interpretation leap from
the information gained from DNA, and specifically 16S
ENCYCLOPEDIA OF CAVES
MICROBES
497
FIGURE 4 (A) Gold and white actinomycete colonies forming in the upper portion of a cave passage, approximately 1 m across, that is not
submerged by rising water, Planinska Jama, Slovenia. (B) Close-up of the colonies from (A); water droplets condense on the hydrophobic sur¬
faces of the colonies. (C) White actinomycete colonies with condensation. Cascade Cave, Kentucky. (D) Clay vermiculations on limestone cave
wall, being covered in gypsum, Frasassi Caves, Italy.
rRNA genes, to knowing the metabolic roles of most
microbes. Attempts to evaluate this dilemma have
yielded mixed results. Recent investigations of hydro-
logically connected caves in different karst basins
have indicated that microbial communities in cave
streams and the karst aquifers under low-flow condi¬
tions include members of the Deltaproteobaderia,
Acidobaderia, and Nitrospirae, but high-flow conditions
following storm events have more representation from
common soil microbial groups, including members
of the phyla Firmicutes, Bacteriodetes, and Chlorobi
(e.g., Pronk et al., 2009). However, genetically similar
16S rRNA gene sequences to these potential distur¬
bance groups, along with members of the Betaproteo-
bacteria, have been considered indigenous communities
in pristine and low-impact (from tourism) cave sedi¬
ments, pristine cave pool waters, and carbonate moon-
milk deposits. To complicate matters, the same
betaproteobacterial groups have been found in high
abundance from high-impact (from tourism) sediments
(Ikner et al., 2007). In short, these mixed results
highlight that it may be possible to determine the com¬
position of an endemic subsurface microbial commu¬
nity, as long as it is possible to distinguish which
microbes may indicate disturbance. This means that
broader spatial and temporal distribution studies need
to be done, including evaluation of overlying soil, drip
waters, air, and so on, as well as studies to understand
the metabolism of active microbial groups because the
presence of specific DNA sequences does not necessar¬
ily mean that the microbes are metabolically active and
affecting the system geologically, geochemically, or
ecologically.
THE FUTURE
There have been important advances in our under¬
standing of the diversity and roles of microbes in caves
and karst, but many challenges wait. It is important
that we determine not just the overall habitat biodiver¬
sity from many more caves and karst settings, but that
ENCYCLOPEDIA OF CAVES
498
M
FIGURE 5 (A) Floating microbial mats in Movile Cave, Romania; PVC grids are 10 cm X 10 cm. (B) White filamentous microbial mats in
the Frasassi Caves, Italy. (C) White mats in Lower Kane Cave, Wyoming. (D) Cave-wall biofilms on gypsum with condensation droplets with
pH <2, Lower Kane Cave, Wyoming. Scale bar is 5 cm.
future research addresses "What are they doing?" and
"Where do they come from?" More data are needed to
test distribution and dispersal hypotheses, and hypoth¬
eses related to how microbes may adapt (or have
adapted) to the physicochemical conditions of karst. Is
it possible that microbes are endemic to a cave or,
more generally, the subsurface? Once these research
directions take shape, and we know more about the
microbial communities at the base of karst ecosystems,
then it should be possible to better protect and con¬
serve the natural habitats through time.
See Also the Following Articles
Chemoautotrophy
Cave Ecosystems
Bibliography
Barton, H. A., & Northup, D. E. (2007). Geomicrobiology in cave
environments: Past, current and future perspectives. Journal of
Cave and Karst Studies, 69(1), 163—178.
Boschker, H. T. S., & Middelburg, J. J. (2002). Stable isotopes and biomar¬
kers in microbial ecology. FEMS Microbiologi/ Ecology, 1334,1—12.
Caumartin, V. (1963). Review of the microbiology of underground envir¬
onments. Bulletin of the National Speleological Society, 25(1), 1—14.
Chen, Y., Wu, L., Boden, R., Hillebrand, A., Kumaresan, D.,
Moussard, H., Baciu, M., Lu, Y., & Murrell, J. C. (2009). Life with¬
out light: Microbial diversity and evidence of sulfur- and
ammonium-based chemolithotrophy in Movile Cave. The ISME
Journal, 3(9), 1093-1104.
Dattagupta, S., Schaperdoth, I., Montanari, A., Mariani, S., Kita, N.,
Valley, J. W., & Macalady, J. L. (2009). A novel symbiosis between
chemoautotrophic bacteria and a freshwater cave amphipod. The
ISME Journal, 3(8), 935-943.
Dawson, S. C., & Pace, N. R. (2002). Novel kingdom-level eukaryotic
diversity in anoxic environments. Proceedings of the National
Academy of Sciences, 99(12), 8324—8329.
DeSantis, T. Z., Hugenholtz, P., Larsen, N., Rojas, M., Brodie, E. L.,
Keller, K., Huber, T., Dalevi, D., Hu, P., & Andersen, G. L. (2006).
Greengenes, a chimera-checked 16S rRNA gene database and
workbench compatible with ARB. Applied and Environmental
Microbiology, 72(7), 5069-5072.
Engel, A. S. (2007). Observations on the biodiversity of sulfidic karst
habitats. Journal of Cave and Karst Studies, 69(1), 187—206.
Engel, A. S. (2010). Microbial diversity of cave ecosystems. In L. Barton,
M. Mandl & A. Loy (Eds.), Geomicrobiology: Molecular & environmental
perspectives (pp. 219—238). Springer. doi:10.1007/978-90-481-9204-5_10.
ENCYCLOPEDIA OF CAVES
MINERALS
499
Engel, A. S., & Northup, D. E. (2008). Caves and karst as model sys¬
tems for advancing the microbial sciences. In J. Martin & W. B.
White (Eds.), Frontiers in karst research (pp. 37—48). Leesburg, VA:
Karst Waters Institute (Special Publication 13).
Faust, B. (1949). The formation of saltpeter in caves. Bulletin of the
National Speleological Society, 11, 17—23.
Guschin, D. Y., Mobarry, B. K., Proudnikov, D., Stahl, D. A.,
Rittmann, B. E., & Mirzabekov, A. D. (1997). Oligonucleotide
microchips as genosensors for determinative and environmental
studies in microbiology. Applied and Environmental Microbiology,
63(6), 2397-2402.
Hutchens, E., Radajewski, S., Dumont, M. G., McDonald, I. R., &
Murrell, J. C. (2004). Analysis of methanotrophic bacteria in
Movile Cave by stable isotope probing. Environmental
Microbiology, 6(2), 111—120.
Ikner, L. A., Toomey, R. S., Nolan, G., Neilson, J. W., Pryor, B. M., &
Maier, R. (2007). Culturable microbial diversity and the impact of
tourism in Kartchner Caverns, Arizona. Microbial Ecologi/, 53(1),
30-42.
Jones, D., Lyon, E., & Macalady, J. (2008). Geomicrobiology of biover-
miculations from the Frasassi Cave System, Italy. Journal of Cave
and Karst Studies, 70(2), 76—93.
Li, T., Wu, T. -D., Maze'as, L., Toffin, L., Guerquin-Kern, J. -L.,
Leblon, G., et al. (2008). Simultaneous analysis of microbial iden¬
tity and function using NanoSIMS. Environmental Microbiology, 10
(3), 580-588.
Marsh, S. (2007). Pyrosequencing applications. Methods in Molecular
Biology, 373, 15-24.
Ouverney, C. C., & Fuhrman, J. A. (1999). Combined microautoradio¬
graphy-16S rRNA probe technique for determination of radioiso¬
tope uptake by specific microbial cell types in situ. Applied and
Environmental Microbiology, 65(4), 1746—1752.
Palmer, A. N. (2007). Cave geology. Dayton, OH: Cave Books.
Principi, P. (1931). Fenomeni di idrologia sotterranea nei dintorni di
Triponzo (Umbria). Grotte d'ltalia, 5, 1—4 (in Italian).
Pronk, M., Goldscheider, N., & Zopfi, J. (2009). Microbial communi¬
ties in karst groundwater and their potential use for biomonitor¬
ing. Hydrogeology Journal, 17(1), 37—48.
Sarbu, S. M., Kane, T. C., & Kinkle, B. K. (1996). A chemoautotrophically-
based cave ecosystem. Science, 272(5270), 1953—1955.
Symk, B., & Drzal, M. (1964). Research on the influence of microor¬
ganisms on the development of karst phenomena. Geographia
Polonica, 2, 57—60.
Whitman, W. B., Coleman, D. C., & Wiebe, W. J. (1998). Prokaryotes:
The unseen majority. Proceedings of the National Academy of
Sciences, 95(12), 6578-6583.
MINERALS
Bogdan P. Onac
University of South Florida, Tampa, U.S.A.,
and “Emil Racovita” Institute of Speleology, Cluj, Romania
This article presents a short review of the physical,
chemical, and crystallographic properties of cave
minerals. In addition, their general modes of occur¬
rence and genesis under various cave settings are dis¬
cussed. To suit the needs of any particular reader,
nearly 30 of the most common cave mineral species
are described, providing the necessary backdrop for
everyone eager to know more about cave mineralogy.
INTRODUCTION
Caves are natural subterranean cavities, fissures,
and fragments of conduit systems that are accessible to
human exploration. The cave environment typically
maintains constant temperature, relative humidity, and
carbon dioxide partial pressure over long periods.
Solutions entering the caves, according to their primar¬
ily chemical composition or following the reaction
with different cave deposits (i.e., bedrock, cave sedi¬
ments, organic deposits) will precipitate a variety of
interesting, sometimes unique, cave minerals. Only
secondary mineral species (i.e., formed within the cave
from a physical process or a chemical process) are con¬
sidered true cave minerals.
Although several types of reactions may take place
in the cave environment (White, 1997), those that are
ultimately responsible for the deposition of common
cave minerals fall into one of the following categories:
(1) dissolution/precipitation (e.g., calcite, gypsum);
(2) hetero- or homogeneous acid/base reactions (the
Mg0-C0 2 -H 2 0 system, carbonic acid); (3) phase transi¬
tions (aragonite/calcite inversion); (4) hydration/
dehydration (e.g., mirabilite/thenardite or brushite/mon-
etite); (5) microbial processes (e.g., calcite precipitation);
and (6) redox reactions involving mainly sulfur and
manganese).
Moore (1970) published the first checklist of cave
minerals. He included 68 cave minerals formed
within and outside of the United States. Six years
later, another comprehensive compilation of nearly
80 minerals, primarily a review of U.S. cave mineralogy,
was undertaken (Hill, 1976). The first book to provide
worldwide coverage on this topic was the first edition
of Cave Minerals of the World (Hill and Forti, 1986).
This edition summed up 173 minerals, 86 of which
were either ore-related or being miscellaneous, un¬
common cave minerals. In the second edition of the
same book published in 1997 were reported 255 cave
minerals, 125 of which were precipitated under special
cave settings. Since then, 29 new cave minerals have
been added and the list is far from being completed
(Onac and Forti, 2011). Apart from these two books
specifically dedicated to cave minerals, more recent
reviews on this topic can be found in White (1988),
Moore and Sullivan (1997), Onac and Forti (2011),
and Ford and Williams (2007).
This exponential increase in the number of cave
minerals over the past 30 years is due to (1) advance¬
ments in analytical facilities (X-ray powder diffraction.
X-ray fluorescence spectrometry, scanning electron
microscopy, electron microprobe, and other tech¬
niques); (2) discovery of many new cavities carved by
sulfuric acid or hydrothermal ore-derived solutions.
Encyclopedia of Caves.
© 2012 Elsevier Inc. All rights reserved.
500
M
skarn-hosted caves, and so on (all displaying a diverse
and fascinating mineralogy); and (3) the increasing
interest for nondescript weathering crusts and various
earthy masses comprising the cave soils.
Minerals most likely to be encountered in a "nor¬
mal" cave environment ( i.e., limestone, gypsum, or salt
caves) belong to eight chemical classes. Our approach
will emphasize only the chief minerals of each class
that are represented in the cave environment following
Dana's New Mineralogy scheme (Gaines et al., 1997).
Less attention will be given to species (ore-related and
other miscellaneous minerals) that are rare, or that do
not ordinarily appear in caves except in a unique set of
conditions. Within each of Dana's classes, physical and
chemical properties of the more common cave miner¬
als are discussed, along with their general modes of
occurrence, genesis, and stability under various cave
settings. The crystallographic data for the common
cave mineral species are listed in Table 1. Organic
minerals are given a brief overview. The presentation
is arranged according to the abundance in the cave
environment of each chemical subclass.
CARBONATES
Approximately 29 carbonate species are reported
from caves worldwide. Of these, calcite and aragonite
are the two most common and abundant (Table 1).
Calcite is the thermodynamically stable form of cal¬
cium carbonate under temperature, pressure, and CO2
partial pressure found in caves. Crystal habits include
rhombohedrons, scalenohedrons (dogtooth), prismatic
(nailhead), or combinations of these (Fig. 1). Aragonite
is the metastable, orthorhombic polymorph of calcite
and occurs as short to long prismatic (along c), acicular
or tabular (chisel-shaped) crystals, often twinned. With
time (i.e., at geologic time scales) aragonite will inter¬
nally change its crystal structure to the stable one of
calcite; however, the external habit of aragonite
remains preserved. Both calcite and aragonite may
appear as monocrystals. More commonly these crystals
are associated in polycrystalline aggregates shaped
into a countless number of speleothem types (i.e., soda
straws, stalactites, and stalagmites). Mineralogical and
isotopic investigations in caves hosting perennial ice
deposits enabled the identification of cryogenic carbo¬
nates. These minerals form during freezing of carbon¬
ate-rich water, when expulsion of CO2 from solution
promotes subsequent precipitation of the rare carbon¬
ate minerals such as monohydrocalcite and ikaite (e.g.,
Scarisoara Ice Cave, Romania).
Although dolomite (a mineral containing magne¬
sium) is very common as a rock, it is rare as a second¬
ary cave mineral. Except for huntite, CaMg 3 (C03)4;
dolomite, CaMg(C0 3 )2; and hydromagnesite, the
Mg-rich cave minerals are quite uncommon. These
minerals, along with magnesite, MgC0 3 ; monohydrocal¬
cite, CaC0 3 ■ H 2 0; and nesquehonite, Mg(HC0 3 )(0H) •
2H 2 0 are the chief constituents of moonmilk deposits
appearing as microcrystalline aggregates of various
habits and shapes. They are also identified in crusts and
nodules. The Mg/Ca ratio of the solutions governs the
deposition of Ca-Mg carbonate species in caves. As the
ratio increases, a polymineral sequence precipitates along
the pathway best illustrated by coralloid multi-aggre¬
gates (Fig. 2).
Some secondary carbonates containing copper and/
or zinc, such as malachite; azurite, Cu 3 (C0 3 )2(0F1)2;
and rosasite, (CuZn) 2 (C0 3 )(0H) 2 , are extremely color¬
ful. The whole range of speleothems they form or stain
have great aesthetic value. These minerals, as well as
the relatively rare cerussite, PbC0 3 ; rhodochrosite,
MnC0 3 ; and smithsonite, ZnC0 3 , are the oxidation pro¬
ducts of sulfides (chalcopyrite, bornite, sphalerite, etc)
dispersed in the bedrock.
Carbonate mineral deposition in caves is dictated
by the availability of supersaturated percolation water
entering the cave galleries. Figure 3 summarizes the
chemical pathway of percolating waters passing
through the vertical sequence of air-soil-limestone-
cave. The meteoric water is already acidified by atmo¬
spheric carbon dioxide, CCUatm)- As these waters per¬
colate through decaying organic soil matter, dissolved
C0 2( bio) increases further. On entering the cave, degas¬
sing of CO2 occurs, causing the water to become super¬
saturated with respect to calcium, resulting in the
precipitation of the various speleothems of calcite
and/or other carbonate minerals. In addition, a grow¬
ing body of evidence points toward the role of bio¬
genic processes in precipitating calcium carbonates in
caves. Through their metabolic activities, a number of
bacteria favor CaC0 3 deposition by simply modifying
the supersaturation of the solution.
The simple linked series of reactions shown in
Figure 3 is far more complex in reality; the chemical
kinetics and various biogenic processes are discussed
in detail by White (1988), Ford and Williams (2007),
Palmer (2007), and Jones (2010).
SULFATES
Ca 2 , Mg 2 / and Na + ions, with lesser amounts of K,
dominate the circulating groundwater. As a result,
gypsum, epsomite, and mirabilite are the usual sulfate
minerals found in caves (Table 1). By far the most rep¬
resentative of this group and perhaps the second most
common cave mineral is gypsum. It forms a huge vari¬
ety of speleothem types ranging from the less
ENCYCLOPEDIA OF CAVES
MINERALS
501
TABLE 1 Crystallographic
Data for Common Cave Minerals
Mineral
Crystal system
Unit cell parameters
Observations
Chemical formula
flo (A)
ct(°)
b 0 (A)
P(°)
Co (A)
y(°)
CARBONATES
Calcite
Trigonal
4.99
Vigorous effervescence in cold dilute HC1, perfect
CaC0 3
-
rhombohedral cleavage; exhibits double refraction;
rhombohedral, prismatic (nailhead), scalenohedrons
17.061
(dogtooth) or twinned crystals.
Aragonite
Orthorhombic
4.959
Acicular; sinks in bromoform; columnar aragonite
CaC0 3
7.964
has cleavage parallel to elongation; twins are
common.
5.737
Hydromagnesite
Monoclinic
10.11
114.5
Common in moonmilk, nodules, earthy masses; feels
Mg 5 (C0 3 ) 4 (0H) 2 ■ 4H z O
8.94
like cream cheese when rubbed between the fingers.
8.38
Malachite
Monoclinic
9.48
Green color and streak; botryoidal habit; typically
Cu 2 C0 3 (0H) 2
12.03
98
banded.
3.21
SULFATES
Gypsum
CaS0 4 ■ 2H 2 0
Monoclinic
5.678
Low hardness; perfect cleavage after (010); swallow¬
tail twins; flexible; soluble in hot dilute HC1.
15.20
118.43
6.52
Epsomite
Orthorhombic
11.86
Acicular crystals elongated parallel to c, powdery
MgS0 4 ■ 7H 2 0
11.99
crusts; bitter taste; very soluble in water; vitreous,
fibrous, silky luster; frequently associated with
6.85
gypsum.
Mirabilite
Na 2 S0 4 ■ 10H 2 O
Monoclinic
11.51
Bitter and salty taste, very soluble.
10.37
107.80
12.85
Celestine
Orthorhombic
8.37
Tabular and fibrous crystals; characteristic pale blue
SrS0 4
5.35
color; brilliant red flame coloration.
6.87
Barite
Orthorhombic
8.88
Tabular and prismatic crystals; high specific gravity;
BaS0 4
5.45
green flame coloration.
7.15
Alunite
Trigonal
6.981
Crystals typically pseudocubic or tabular; fibrous to
KA1 3 (S0 4 ) 2 (0H) 6
—
columnar, porcelaneous, commonly granular.
17.331
PHOSPHATES
Hydroxylapatite
Hexagonal
9.41
Compact to coarse granular in crusts; various shades
Ca 5 (P0 4 ) 3 (0H)
_
of brown and red; indistinct cleavage.
6.87
(Continued)
ENCYCLOPEDIA OF CAVES
502
M
TABLE 1 ( Continued )
Mineral
Crystal system
Unit cell parameters
Observations
Brushite
Monoclinic
6.36
Colorless to ivory yellow; needle-like or prismatic
CaHP0 4 ■ 2H 2 0
15.19
118.45
crystals; earthy or powdery.
5.18
Ardealite
Monoclinic
5.72
Fine-grained powdery masses; white to pale yellow;
Ca 2 (HP0 3 0H) (SO 4 ) ■ 4H 2 0
30.95
117.29
Associated with gypsum and brushite; described
from Cioclovina Cave, Romania.
6.26
Taranakite
Trigonal
8.71
Nodules, flour-like masses; white, gray, yellowish;
K 3 Al5(P0 3 0H) 6 (P04)2- 18H 2 0
-
acidic environment indicator; first reported as
minervite.
95.1
Variscite
Orthorhombic
9.822
Earthy masses, nodules, and crusts; grayish-white.
AIPO 4 ■ 2H 2 0
9.630
pale-yellow to brown.
8.561
OXIDES & HYDROXIDES
Goethite
Orthorhombic
4.62
Blackish brown to yellow or reddish; typically
FeO(OH)
9.95
massive, botryoidal, with internal radiating fibrous
or concentric structure; pseudomorphs after pyrite.
3.01
Ice
Hexagonal
4.52
Melts at CPC; form sublimation crystals.
h 2 o
_
7.36
Birnessite
Monoclinic
5.174
Red-brown streak; soluble in concentrate HC1;
(Na,Ca,K ) 0 6 (Mn 4 + ,Mn 3 + ) 2 0 4 ■ 1.5H 2 0
2.85
103.18
commonly earthy and admixed with other Fe
and Mn oxides and hydrates.
7.33
SILICATES
Opal
Si0 2 ■ nH 2 0
Amorphous
Botryoidal crusts, banded; conchoidal fracture; brittle.
Quartz
Trigonal
4.91
Prismatic crystals terminated by hexagonal
Si0 2
-
dipyramid or rhombohedra faces; practically any
color; high hardness; soluble only in HF; no cleavage.
5.40
Allophane
Amorphous
Earthy masses, moonmilk-like deposits; white.
Al203(Si02)l.3— 2 . 0 '2.5— 3 H 2 O
pale-cream, blue, yellow.
HALIDES
Halite, NaCl
Cubic
5.64
Salty taste, perfect cubic cleavage, low hardness.
Fluorite, CaF 2
Cubic
5.46
Isometric; fluorescent in UV, no effervescence in HC1.
NITRATES
Niter
Orthorhombic
5.41
Cooling taste, soluble in water, violet flame.
kno 3
9.16
6.43
ENCYCLOPEDIA OF CAVES
MINERALS
503
FIGURE 1 Common crystal habits of calcite: (A) rhombohedron,
(B) scalenohedron (dogtooth), (C) prismatic (nailhead). Adapted from
Onac, 2000.
Mg/Ca increases -►
calcite - Mg-caicite-aragonite - huntite - hydromagnesite
FIGURE 2 Sequence of polymineral precipitation in corraloid
multi-aggregates. Adapted from Maltsev, unpublished manuscript.
spectacular white, soft moonmilk deposits composed of
microcrystals, to curving clusters, to prismatic crystals
up to 12 m long (Cave of the Giant Crystals, Naica,
Mexico), and to enormous and spectacular stalactites
(the so-called chandeliers) exposed in Lechuguilla (New
Mexico) and Cupp-Coutunn (Turkmenistan) caves.
Two other common cave sulfates that usually occur
as colorless to white fibrous efflorescences or crusts on
cave walls and sediments are mirabilite and epsomite.
- H* + HCO- | ■
2 HCO- -CaCO, + H,0 + Cok
FIGURE 3 Reactions in the air-soil-rock-cave system and carbo¬
nates deposition. Modified from Onac, 2000.
Both minerals are extremely sensitive to variations in
relative humidity and temperature. As the relative
humidity drops below ~70% mirabilite will dehydrate
to thenardite (Na 2 S 04 ) whereas epsomite transforms into
hexahydrite (MgS 04 ■ 6 H 2 O) and kieserite (MgS 04 ■ H 2 O)
respectively. Samples of mirabilite and epsomite removed
from the cave environment will decompose within
minutes into a white milky powder. Cave passages with
low relative humidity, high temperatures (above 20° C),
or that are well ventilated appear to be ideal locations for
the deposition of bassanite ( 2 CaSC >4 ■ H 2 0) and anhydrite
(CaS0 4 ).
Barite and celestine (Table 1), are relatively insoluble
in comparison to other sulfates. Only under special con¬
ditions does direct precipitation of these minerals in
caves from meteoric waters take place. Both sulfates
commonly occur as millimeter- to centimeter-size
crystals, crusts, and speleothems in caves invaded by
thermal waters or hydrothermal solutions. Famous
occurrences of celestine are known from Crystal Cave
(Ohio), Lechuguilla Cave (New Mexico), Cango Cave
(South Africa), and Valea Rea Cave (Romania). Barite is
known from caves in England, Italy, Hungary, France,
Russia, and the United States.
Although less common compared to the above cave
sulfates, alunite (Table 1 ) is a key mineral used to trace
the origin of caves and their ages. It is a by-product
material resulting from the reaction between cave clays
and sulfuric acid in the process of cave formation.
Because alunite can be dated using radiometric meth¬
ods ( i.e ., K-Ar or Ar-Ar), it can provide age estimates
for sulfuric acid caves (Polyak et al., 1998).
The other 60 sulfates described from caves (e.g., alu-
nogen, halotrichite, jarosite, chalcantite, melanterite,
etc.) are known only from warmer caves in which Al,
Fe, and Cu were supplied by ore bodies or hydrother¬
mal solutions.
Four mechanisms are ultimately responsible for the
deposition of sulfates in caves. Most commonly cited is
simple precipitation by evaporation and sulfuric acid
reaction with the carbonate bedrock or cave sediments.
With respect to the origin of sulfates in caves, when
these minerals do not occur in the overlying limestone,
the following sources have been proposed: oxidation
ENCYCLOPEDIA OF CAVES
504
M
of sulfides ( e.g., pyrite or marcasite), presence of bat
guano, or post-volcanic activities ( i.e ., fumaroles).
PHOSPHATES
Phosphates are the second largest group of cave
minerals after sulfates. The (PO 4) 3 radical combines
with some 30 elements to form over 300 phosphate
minerals, out of which 57 have been found in different
cave settings. Hydroxylapatite, brushite, ardealite, tara-
nakite, and variscite are the most abundant (Table 1);
all the other phosphates are rare. They occur whenever
a cave contains fresh or fossil bat guano or significant
accumulations of bone deposits. Concentrations of bat
guano are found primarily in caves situated in low-
latitude humid areas. The most diverse phosphate
assemblages are to be found there. Many of the rare
phosphate minerals were identified in caves from
South Africa, Australia, Namibia, and Romania. Unlike
other cave minerals, the phosphates do not form
spectacular speleothems, but occur as crusts, nodules,
lenses, and earthy or powdery masses.
Depending on whether the percolating water passing
through guano reacts with carbonate rocks or clay
minerals, Ca-rich or Mg- and Al-rich phosphates will be
deposited. Less commonly, the reaction of guano with
ore-derived metals or spontaneous combustion of guano
can produce some rare phosphate minerals (Fig. 4).
Many cave phosphate minerals include the ammo¬
nium ion (e.g., biphosphammite, H 2 (NH 4 )P 04 ; stru¬
vite, (NH 4 )Mg(P0 4 ) ■ 6H 2 0; taranakite, etc.), which
is derived from decomposition of bat urea. After
leaching of alkali ions, the more stable minerals (e.g.,
hydroxylapatite or brushite) persist. Members of the
apatite group, Ca 5 (P0 4 )3(0H,F,Cl), are common cave
minerals. Fluorine, chlorine, and the hydroxyl ion
can mutually replace each other to form almost
pure end members: fluorapatite, chlorapatite, and
hydroxylapatite. Of these, the last one is the most
thermodynamically stable phosphate mineral under
ordinary cave conditions, occurring on almost all
coatings and crusts. Brushite is stable under acidic
(pFI <6) and damp conditions, occurring as nodular
masses or prismatic crystals. It loses water readily,
converting to monetite (CaHP0 4 ). Brushite is iso-
structural with gypsum, so it is not surprising that
ardealite should exist, particularly in view of their
frequent association. A decrease of organic content
and an increase of the Ca/P ratio characterize the
mineral sequence brushite-whitlockite-hydroxylapatite.
Like brushite, taranakite is stable in acidic environ¬
ments and precipitates near contacts between guano
and clays, under poorly drained conditions. It usually
occurs as white-yellowish soft nodules, veins, or
FIGURE 4 Development of phosphate minerals in an idealized and simplified cave environment.
ENCYCLOPEDIA OF CAVES
MINERALS
505
flour-like masses. Variscite, another fairly common
phosphate, occurs under acidic conditions as veins
and nodules in, or at the contact with, highly phos-
phatized Al-rich cave sediments. Other products
with the interaction of bat guano with various rocks
and loose sediments in the cave environment are
illustrated in Figure 4.
OXIDE^AND HYDROXIDES
Among the most common oxide and hydroxide
minerals frequently precipitated in caves are those
containing iron (e.g., goethite, hematite, Fe 2 0 3 ; lepido-
crocite, Fe 3+ O(OFQ) and manganese (e.g., birnessite,
pyrolusite, Mn0 2 ; and romanechite, (Ba,H 2 0) 2 (Mn 4+ ,
Mn ’ ) 5 0| 0 ); Table 1). Of these, goethite occurs in caves
worldwide, forming a variety of speleothems colored
reddish-brown. Dark-brown to black fine-grained mix¬
tures (chiefly birnessite, romanechite, and todorokite)
have been reported covering cave walls, stream clasts,
and other speleothems in caves such as Vantului
(Romania), Jewel and Jasper (U.S.A.), and Zbrasov
(Czech Republic). The term wad is frequently used as a
field term for composite mixtures consisting mainly of
manganese oxides, in much the same sense as limonite is
used for hydrous iron oxides. All minerals mentioned
above are formed under oxidizing conditions.
Pyrite (FeS 2 ) and other sulfides are the primary
source for iron oxide and hydroxide minerals in caves.
Manganese may be released from impure limestone or
from within the soil zone where it accumulates from
decaying of plants. Some bacterial species are known
to catalyze the oxidation and precipitation of Fe and
Mn in the cave environment. Furthermore, pH and/or
redox conditions within the aqueous solution control
this biologically mediated process.
The crystalline forms of solid water (ice) also belong
to this group of cave minerals. Ephemeral ice spe¬
leothems form in the entrance passages of caves dur¬
ing cold seasons, whereas perennial ice deposits
accumulate when particular airflow conditions main¬
tain low temperatures even during the warm season.
Freezing of dripping or flowing water creates similar
speleothems to those of calcite. Spectacular hexagonal
platy ice crystals and hoarfrost may form by sublima¬
tion on the cave walls and ceiling from moist, warm
air entering the cave.
SILICATES
The silicates comprise about a third of all mineral
species. Although many of these minerals are uncom¬
mon in caves, others make up some particular cave
deposits (e.g., opal, quartz, and allophane) (Table 1).
The solidified colloidal silica, opal, is the most com¬
mon mineral among the cave silicates. It occurs mostly
in lava tubes and cavities within volcanic or metamor-
phic rock, but may also be found in limestone caves.
Opal typically forms botryoidal crusts, coralloids, and
can be intermixed with calcite in different speleothems.
In time it can gradually transform to chalcedony or
even quartz. The most abundant opal speleothems
are to be found in lava tubes (e.g., Algar do Carvao,
Azores). The presence of quartz in caves, usually as
euhedral crystals lining the walls, indicates at least one
hydrothermal episode in the cave's history. Spectacular
quartz crystals are displayed in Wind and Jewel caves
(South Dakota), Chiricahua Crystal Cave (Arizona), and
Kristallkluft Cave (Swiss Alps). Fluid inclusion studies
on quartz allow for estimating the temperature of the
hydrothermal solutions from which the crystals were
precipitated.
The third most frequent silicate mineral in caves is
allophane, an amorphous hydrous aluminum silicate.
It has been reported mostly as white-pale cream to
yellow-orange moonmilk-like deposits, but also in
the form of different speleothems. Among the best
displays of allophane are those from Mbobo Mkulu
(Transvaal, South Africa) and Iza (Romania) caves,
where it forms spectacular blue flowstones. Most allo¬
phane occurrences are related to the action of acidic
waters on cave sediments.
Most clay minerals (hydrated silicates) in caves are
thought to be derived from the bedrock or were trans¬
ported in along with other clastic sediments by the
cave streams or percolating waters. Others, however,
are definitely of authigenic origin. Among these are
dickite, halloysite, montmorillonite, and sepiolite.
In addition, a long list of other cave silicates can be
compiled (e.g., benitoite, clinochlore, epidote, fraipon-
tite, ilvaite, etc.); however, their secondary origin is
questionable. All these minerals were deposited either
as crystals, granular aggregates, or earthy masses
mostly under particular settings (i.e., hydrothermal
and ore-related environs, lava tubes, etc.).
HALIDES
The list of halides, compounds in which a halogen
element is the sole or principal anion, comprises some
100 species. Only 11 of these minerals were docu¬
mented from different cave environments. The more
common members are simple in composition and fall
into either the anhydrous (halite; sylvite, KC1; fluorite,
etc.) or hydrous (carnallite, KMgCl 3 ■ 6H 2 0; atacamite,
Cu 2 C1(OH) 3 , etc.) halide group. They form mostly in
caves located within extensive salt beds (Mount
ENCYCLOPEDIA OF CAVES
506
M
Sedom, Israel; Meledic Plateau, Romania) or in lava
tubes from Iceland, Italy, the United States, and
Australia.
Halite is by far the most common water-soluble
halide mineral (Table 1), displaying an overwhelming
variety of speleothem types, especially when precipi¬
tated in salt caves. Typically, halite may appear as
massive, coarsely granular or compact aggregates, or
as cubic crystals, often with cavernous and stepped
faces (hopper crystals). It is colorless, white, gray,
yellowish (colored by discrete impurities), blue, or
purplish (color due to lattice imperfections).
Fluorite is another secondary halide mineral form¬
ing various-sized euhedral cubic crystals (sometimes
twinned) and coarse-to-fine granular aggregates of
blue, purple, green, or violet color (Table 1). Fluorite
has been described from a number of ore-related
cavities such as Cupp-Coutunn (Turkmenistan), Treak
Cliffs (U.K.), Kootnay (British Columbia, Canada),
Glori and Blanchard (New Mexico, U.S.A.). In all these
locations fluorite was interpreted as being precipitated
from hydrothermal solutions circulating along fissures
or channels. Rare occurrences are known for fluorite
precipitated from low-temperature percolating waters
(caves in Arizona and Wyoming) (Hill and Forti, 1997).
Except for the halides occurring in caves within or
near ore bodies (fluorite and bromargyrite) or those
precipitated from fumarole vapors (e.g., galeite and
kainite), all are rather soluble salts that are leached by
percolating waters from the overburden rocks and
redeposited in caves during evaporation of the solvent.
Therefore, these minerals form mainly in caves in dry
climate areas that experience little precipitation.
NITRATES
Minerals of this group are structurally similar to the
carbonates with plane triangular (NO3) - groups in
place of (C0 3 ) 2- ■ Since N is more electronegative than
C, the anionic group, (NO3) - , is less stable, restricting
the occurrence of nitrate cave minerals much more
than the carbonates. Nitrates are highly soluble, form¬
ing only in warm, dry (low humidity), and well-
ventilated caves. They commonly appear along with
gypsum, epsomite, halite, and other salts as efflores¬
cence, tiny crystals disseminated within cave soil and
dry guano deposits.
There are 11 nitrate minerals found in caves, with 5
of these being known only from particular cave
environs (sveite, mbobomkulite, hydrombobomkulite,
nickelalumite, and gwihabaite). Although niter; nitra-
tine, NaN0 3 ; nitrocalcite, Ca(N0 3 ) 2 ■ 4H 2 0; and nitro-
magnesite, Mg(NC> 3) 2 • 6H 2 0, were documented in
several caves worldwide, none of these can be consid¬
ered common for the cave environment.
Most of the nitrate minerals are concentrated in cave
soils after being carried inside by water percolating
through rich organic soils. Alternatively, bat and rat
guano or basic volcanic rocks can also supply nitrates in
caves. Ultimately, microorganisms (particularly bacteria)
play major roles in the deposition of nitrate minerals in
caves. Their precipitation is affected by environmental fac¬
tors that influence microbial activity, such as temperature,
moisture, and organic matter availability.
NATIVE ELEMENTS
Although several chemical elements occur as minerals
in the Earth's crust, so far only sulfur has been found in
a number of caves from the United States, Italy, Mexico,
Iceland, Romania, and Russia. Sulfur occurs as canary-
yellow powder, granular to massive aggregates, or it can
take a variety of speleothem types (crusts, stalactites,
euhedral crystals, cave rafts, etc.). Most commonly, native
sulfur occurs in sulfuric acid caves, in which its origin
was related to either hydrogen sulfide oxidation (caves
from the Guadalupe Mountains, United States; Cueva de
Villa Luz, Mexico; Sant Cesarea and Grotta dello Zolfo,
Italy; Diana, Romania) or reduction of sulfate ion in the
presence of sulfur bacteria (Frasassi, Grotta di Cala
Fetente, Italy) (Hill and Forti, 1997).
ARSENATES AND VANADATES
The arsenates and vanadates are oxysalts character¬
ized structurally by independent (AsCh) 3- or (VO 4) 3
anionic groups linked through intermediate cations.
The majority are mostly low-temperature, hydrous
minerals. Only eight arsenate and six vanadate minerals
were identified in caves known to have sulfuric acid or
hydrothermal genesis. Typically, they form at the
expense of primary arsenic or vanadium-rich minerals
by weathering and oxidation processes; some may have
been precipitated in reducing hydrogen sulfide-rich
environments. None of these minerals is common for
normal cave settings; however, when they occur as
small crystals or scaly crusts, they are easily noticed
due to their bright-yellow or green color. Common
members of each of these two subclasses include: coni-
chalcite and talmessite (arsenate group) and tyuyamu-
nite, carnotite, and calciovolborthite (vanadate group).
Most of the above minerals were identified in caves
from Tyuya-Muyun region (Kyrgyzstan); Valea Rea
Cave (Romania); Sonora Caverns (Texas); Lechuguilla,
Carlsbad, Spider, and other caves in the Guadalupe
Mountains (New Mexico).
ENCYCLOPEDIA OF CAVES
MINERALS
507
SULFIDES
This group of cave minerals comprises compounds
in which the large atoms S, As, and Bi are combined to
one or more of the metals (Fe, Zn, Pb, Cu, Hg, and Sb).
Thirteen minerals ( e.g., pyrite, marcasite, cinnabar,
metacinnabar, stibnite, etc.) were described from
skarn-hosted and hydrothermal-related caves. Most of
the later ones are associated with Mississippi Valley-
type lead-zinc ore deposits (Hill and Forti, 1997; Ford
and Williams, 2007). The usual occurrences of sulfides
consist of crusts, inclusions, druses, or crystals. With
few exceptions, all sulfides were precipitated in preex¬
isting metasomatic- or meteoric-dissolved karst chan¬
nels from low-temperature hydrothermal solutions.
ORGANIC MINERALS
The presence of large bat or bird colonies in some
caves from arid and semi-arid regions of Western
Australia, Namibia, Israel, and the United States is ulti¬
mately responsible for the deposition of nine highly solu¬
ble organic minerals. These include oxalates, mellates, and
purines, which normally form pale-yellow, red, or light
brown crusts and efflorescences. Almost all organic miner¬
als were derived from the reaction of urine or animal
excreta with carbonate bedrock, clays, or cave detritus.
WHY STUDY CAVE MINERALS?
Cave minerals can yield a wealth of information.
Foremost, they simply retained their scientific value as
mineral bodies occurring within the cave. Much infor¬
mation, however, is gained by studying speleothems,
which are made up of one or more cave minerals.
Examinination of such information reveals that cave
minerals are valuable data repositories, which allows
for the investigation of dozens of interesting problems
that cover a broad range of topics; some of these are
summarized below:
1. Caves are natural underground laboratories where
crystal growth processes can be observed and
monitored.
2 . Studies of crystallography and mineralogy of cave
minerals provide invaluable insight into the
chemical and physical (sometimes biological too)
conditions existing within given cave environments
at various times during mineral formation.
3. Some caves provide a unique set of conditions that
allow the deposition of a suite of exotic minerals;
many of these are restricted to only such particular
cave settings and are never found in the outside
world.
4. The mineral assemblage present in a given cave or
caves within a region can sometimes be used as
diagnostic criterion of the cave speleogenetic
pathway.
5. Calcite speleothems in caves are well suited for
uranium-series dating based on which an absolute
chronology can be derived. Furthermore, carbon
and oxygen isotopic variations in carbonate
speleothems combined with changes in calcite
crystal fabric and luminescence of growth lamina
are potentially powerful tracers of changes in
Quaternary climate and possibly vegetation.
Aragonite and calcite speleothems in littoral caves
can resolve sea level changes and tectonic histories.
6 . Oxygen isotope analyses (8 ls O) and fluid inclusions
of hydrothermal-related cave minerals (e.g., quartz,
barite, calcite, malachite, etc.) provide solution
temperature at the time of deposition.
7. The ability to reliably differentiate hypogene from
supergene cave sulfates (e.g., gypsum, barite,
celestine, etc.) by studying their isotopic composition
(8 34 S) is critical to understanding sulfur source,
hydrological, and speleogenetic processes.
8 . Some cave minerals (e.g., calcite, alunite, etc.) play
an important role in reconstructing landscape
evolution (e.g., valley incision rate, mountain uplift)
or determining the age of caves.
9. Supporting evidence shows some mineral assemblages
were precipitated in caves with the aid of various
microorganisms. Deciphering the geomicrobiological
processes involved on their deposition may help our
understanding of the caves as life environments
on Earth or beyond.
Apart from the above issues, the cave minerals can
provide an accessible window into geochemistry, pale¬
ontology, archaeology, soil science, ore deposits, and
other Earth science disciplines. It is expected that their
use will become even more widespread, stimulating
further interest and excellence in future multidisciplin¬
ary research carried into these fascinating subterranean
worlds.
See Also the Following Article
Speleothems: General Overview
Bibliography
Ford, D. C., & Williams, P. W. (2007). Karst hydrogeology and geomor¬
phology (2nd ed.). Chichester, England: John Wiley & Sons.
Gaines, R. V., Skinner, C., Foord, E. E., Mason, B., & Rosenzweig, A.
(1997). Dana's new mineralogy. New York: Wiley & Sons.
ENCYCLOPEDIA OF CAVES
508
M
Hill, C. A. (1976). Cave minerals. Huntsville, AL: National Speleological
Society.
Hill, C. A., & Forti, P. (1986). Cave minerals of the world. Huntsville,
AL: National Speleological Society.
Hill, C. A., & Forti, P. (1997). Cave minerals of the world (2nd ed.).
Huntsville, AL: National Speleological Society.
Jones, B. (2010). Microbes in caves. In H. M. Pedley & M. Rogerson
(Eds.), Tufas and speleothems. Unravelling the microbial and physical con¬
trols (336, pp. 7—30). Geological Society of London Special Publication.
Moore, G. W. (1970). Checklist of cave minerals. National Speleological
Society News, 28(1), 9—10.
Moore, G. W., & Sullivan, N. (1997). Speleology. Caves and cave envi¬
ronment (3rd ed.). St. Louis, MO: Cave Books.
Onac, B. P. (2000). Geology of karst terrains. Bucuresti: Didactica si
Pedagogica.
Onac, B. P. , & Forti, P. (2011). State of the art and challenges in cave
minerals studies. Studia UBB Geologia, 56(1), 33—42.
Palmer, A. N. (2007). Cave geology. Dayton, OH: Cave Books.
Polyak, V. ]., McIntosh, W. C., Giiven, N., & Provencio, P. (1998).
Age and origin of Carlsbad Cavern and related caves from
40 Ar/ 39 Ar of alunite. Science, 279(5358), 1919-1922.
White, W. B. (1988). Geomorphology and hydrology of karst terrains.
New York: Oxford University Press.
White, W. B. (1997). Thermodynamic equilibrium, kinetics, activation
barriers, and reaction mechanisms for chemical reactions in karst
terrains. Environmental Geology, 30(1—2), 46—58.
MODELING OF KARST AQUIFERS
Qeorg Kaufmann, Douchko Romanov, and
Wolfgang Dreybrodt'
Free University of Berlin, Germany, ^University of Bremen,
Germany
INTRODUCTION
Groundwater flow through a karst aquifer is prone to
contamination because of the very nature of the karst-
ified host rock: Fissures and bedding partings in the
rock are enlarged by chemical dissolution over time
and provide preferential flow paths, through which
water is transferred rapidly and almost unfiltered
from input points such as sinks and dolines to output
points such as large karst springs. The enlarged frac¬
tures and bedding partings are responsible for a very
heterogeneous distribution of permeability within the
karst aquifer. Enlarged passages can be very conduc¬
tive (1 —10ms _1 ) but have low storage capacity. The
surrounding rock is orders of magnitude less conduc¬
tive (10 -8 ms -1 ), but can provide significant storage.
This large-scale heterogeneity in conductivity makes
it difficult to assess the karst aquifer properties from
field studies such as borehole pumping, packer, and
slug tests. Monitoring spring discharge, on the other
hand, provides only an integral picture of the karst
aquifer. A different approach to understanding a
karst aquifer and its spatial and temporal evolution
are numerical models. This field has evolved dramati¬
cally over the past decades, and is described in this
article.
CHEMISTRY
In karst landscapes, soluble rocks are dissolved by
water enriched with carbon dioxide. In the case of gyp¬
sum, the dissolution is a physical dissociation process,
in which only water and gypsum are involved. In the
case of limestone or dolomite, the water enriched
with carbon dioxide, which originates from the atmo¬
sphere and possibly the organic soil cover, reacts to
form carbonic acid, which then dissociates and chemi¬
cally dissolves the rock. On the surface, the bedrock is
removed under conditions open to the atmosphere,
and the dissolved material is then transported away in
the aqueous solution. This process, termed karst denu¬
dation, lowers the surface with time. Within the karst
aquifer, bedrock is removed from the walls of narrow
fractures and bedding partings, enlarging the voids in
which the water circulates. Depending on the situation
within the aquifer, dissolution proceeds either under
open-system conditions (soil, epikarst, vadose zone) or
under closed-system conditions with solution com¬
pletely filling the voids (phreatic zone). The enlarge¬
ment of fractures and bedding partings in the aquifer
enhances flow by a positive feedback loop, as the
enlarged voids provide a much larger conductivity
than the initial bedrock.
Calcium Equilibrium
The maximum amount of dissolved limestone or
gypsum, described by the calcium concentration c
[molm -3 ], is controlled by the calcium equilibrium con¬
centration, c, tj [molm 3 ]. When the solution reaches this
concentration, it is saturated with respect to calcite or
gypsum and dissolution stops. For gypsum, the calcium
equilibrium concentration is almost independent of the
water temperature and amounts to c ai = 15.4 molm -3 .
For limestone in its pure form, calcite, the calcium
equilibrium concentration can be derived as an analy¬
tical expression (e.g., Dreybrodt, 1988); and it depends
strongly on carbon-dioxide pressure and also on water
temperature.
As seen in Figure 1, a solution undersaturated with
calcium is located in the undersaturated field to the
left of the equilibrium lines (white dot on the left
axis). Depending on the boundary conditions, disso¬
lution proceeds either along the horizontal line, until
the equilibrium is reached (upper white dot on the
equilibrium curve), thus keeping the carbon-dioxide
Encyclopedia of Caves.
© 2012 Elsevier Inc. All rights reserved.
MODELING OF KARST AQUIFERS
509
FIGURE 1 Calcium equilibrium as a function of carbon-dioxide
pressure. The equilibrium line is shown for two temperatures,
T = 10°C and T = 20°C. Also shown are the different pathways for
solution toward equilibrium (white dots). In the open system, disso¬
lution proceeds along the horizontal line, while in the closed system
dissolution proceeds along the sloping line. The effect of mixing two
saturated solutions equally (black dots) results in a newly undersatu¬
rated solution (gray dot).
pressure in the solution constant (open system), or if the
closed-system condition holds, the carbon-dioxide pres¬
sure decreases, and the calcium equilibrium line is
reached at a lower value (lower white dot on the equi¬
librium curve). The dissolution of calcite is also tempera¬
ture-dependent, and a decrease in temperature results in
an increase in dissolved calcium. Thus, a cooling of
water through daily variations in surface temperature or
cooling of rising thermal water can dissolve additional
calcite. While the former process is limited to the first
few meters in juvenile karst aquifers, thermal corrosion
can be locally important in regions of hydrothermal
water supply. Here, large temperature gradients are
responsible for a significant amount of additional calcite
dissolved. A temperature drop of T = 10°C will result in
around 0.5 mol m” 3 additional calcium dissolved.
The mixing of two solutions saturated with respect to
calcite, which have different carbon-dioxide concentra¬
tions (black dots), results in a solution which, again, can
dissolve calcite (gray dot). This alternative form of cor¬
rosion, termed mixing corrosion (Bogli, 1980), is a result
of the nonlinear relation between the calcium equilib¬
rium concentration and the carbon-dioxide pressure. As
a consequence, water with different chemical properties
flowing in the karst aquifer through the network of
small fissures and cracks, all hydrologically connected,
will mix, and results in additional corrosive power. As
water with different partial carbon-dioxide pressure is
the normal case within a drainage area due to differ¬
ences in surficial origin of the solution, in surface cover
and soil on a scale of meters as well as daily variations
of organic activity, mixing corrosion can be a significant
FIGURE 2 Low- and high-order flux rates as a function of cal¬
cium concentration. Flux rates are given for different temperatures T
and carbon-dioxide pressures P, and a film thickness of d = 0.001 cm.
process on a local scale, but the additional amount of
dissolved calcium rarely exceeds 0.2 mol m ’.
Calcium Flux Rate
The flux rate F (mol/m -2 s -1 ) describes the removal of
material from the surface per unit area and per time. It is
controlled by several potentially rate-limiting processes
on the surface: (1) the surface reaction of water with the
bedrock; (2) mass transport of the dissolved species by
diffusion; and (3) conversion of carbon dioxide into car¬
bonic acid, in the case of limestone. Flux rates have been
measured experimentally for limestone and calcite and
for gypsum, and have been predicted numerically (see
references in Dreybrodt et al., 2005; Kaufmann and
Dreybrodt, 2007). For both gypsum and limestone, they
can be approximated by a nonlinear rate law. In the case
of limestone, for large undersaturation (c < 0.3c e J, an ini¬
tial fast linear kinetics is present, with very high removal
capability. For more saturated solution ( 0.3c eij < c <
0.9c eq ), a slower linear kinetics takes over, which removes
material about one order of magnitude slower than
the initial rate. For concentrations close to the equilibrium
(c > 0.9c e q), the high-order kinetics is established, and the
rate becomes nonlinear and drops fast. The reason is the
accumulation of impurities on the reactive surface, which
significantly inhibits the dissolution process.
For the flux rate of limestone, several key aspects can
be seen (Fig. 2): Firstly, flux rates depend strongly on the
carbon-dioxide pressure in the solution. The larger the car¬
bon-dioxide pressure the more calcite can be dissolved.
Secondly, flux rates decrease with temperature, with
warmer solutions being able to dissolve less calcite.
Thirdly, calcium flux rates decrease rapidly with increas¬
ing concentration for strong undersaturation (c < 0.3c eq ),
from then on the decrease is slower, but still linear, for
ENCYCLOPEDIA OF CAVES
510
M
calcium concentrations below 90% of the equilibrium con¬
centration. For c >0.9c ecj , dissolution rates drop sharply
and the higher-order power law controls the flux rates.
This nonlinear effect is important for the evolution of karst
aquifers during their early phases, as the reduced flux
rates close to equilibrium increase the distance undersatu¬
rated solution can penetrate into the aquifer significantly.
Gypsum shows linear kinetics up to 0.94c,, £/ and then
becomes nonlinear.
THE SINGLE FRACTURE
Fractures and bedding partings in the karst aquifer
are enlarged by chemical dissolution, following the
chemical reactions described above (Fig. 3). The fracture
aperture d (m) is a function in time of the initial aperture
d{ (m) and the calcium flux rate F in a given fracture. A
first generation of models introduced conceptual ideas
and started developing karst evolution models as a one¬
dimensional single fracture (see references in Dreybrodt
and Gabrovsek, 2003; Dreybrodt et al., 2005).
An example of the evolution of a circular fracture in
limestone is shown in Figure 4. The fracture has a
length of 1 km and the initial fracture aperture is
di = 0.5 mm; the head difference between the left and
the right side is 10 m. The upper panel shows the evo¬
lution of the aperture with time. The undersaturated
solution ( Ci n = 0 mol m -3 ) enters the fracture from the
left side. As dissolution proceeds in the early phase,
only the entrance section is enlarged, whereas the
downstream section grows only slowly (red lines).
This is a result of the slow high-order kinetics, as the
calcium concentration quickly reaches levels close to
saturation farther downstream, which is shown in the
middle panel. Thus, efficient dissolution only occurs
within the entrance section. When enlargement pro¬
ceeds, more highly undersaturated solution is carried
farther downstream, and the fracture finally grows at a
relatively constant pace (blue lines).
The importance of high-order kinetics close to the satu¬
ration of the solution, c ecj , becomes obvious when we dis¬
cuss the calcium-concentration along the fracture. At the
onset of evolution, linear kinetics in the entrance section
results in solution close to saturation within the first few
centimeters. As a consequence, modeling only low-order
linear kinetics would result in enlargement limited to a
few centimeters to meters from the entrance, depending
on fracture geometry and film thickness. By taking high-
order kinetics into account, when the solution is close to
saturation, the calcium flux rate is reduced by several
orders of magnitude, and slightly undersaturated solution
can be carried much father downstream. Dissolution is
occurring up to several kilometers within a fracture. This
mechanism is largely responsible for enlarging fractures
FIGURE 3 Sketch of a single fracture along the junction of a fis¬
sure and a bedding plane.
in limestone and gypsum during the early phase of karsti-
fication. The widening of the fracture at its exit increases
with increasing aperture width. This causes a positive
feedback loop. The amount of water flow through the
fracture is shown in the lower panel. Firstly, the flow rate
increases slowly with time, but due to the increasing wid¬
ening of the fracture aperture width at the exit, the feed¬
back becomes so strong that suddenly a dramatic increase
in the flow rate is observed. This event is called break¬
through. After breakthrough the amount of water, which is
limited, determines the further evolution of the fracture.
MODELING THE EVOLUTION
OF KARST AQUIFERS
Based on the single fracture, more sophisticated
two-dimensional karst aquifer network models were
assembled (see references in Dreybrodt et al., 2005),
which were used to describe the development of karst
aquifers both in horizontal and vertical scenarios, the
latter with the inclusion of a varying water table.
These two-dimensional models have then been used
to describe regional case examples, for example, the
gypsum maze caves from the Western Ukraine (Rehrl
et al., 2008), and the bathyphreatic loop caves in the
Jurassic limestone of the Swabian Alb in southwestern
Germany (Kaufmann and Romanov, 2008). Another
focus has been the safety and integrity of water reser¬
voirs. Flere, karst modeling is used to assess the tem¬
poral evolution of leakage and the development of
subsurface voids and sinkholes in the vicinity of large
hydraulic structures (e.g., Romanov et al., 2003).
The most recent development is the extension of karst
aquifer evolution models into the third dimension
(Annable, 2003; Kaufmann, 2009; Kaufmann et al., 2010).
With these models, it is possible to simulate a karst
ENCYCLOPEDIA OF CAVES
MODELING OF KARST AQUIFERS
511
FIGURE 4 Evolution of a single fracture. Flow is
driven from left to right. (A) Fracture aperture evo¬
lution; the black line indicates the initial fracture
width, red lines are enlarged fracture widths before,
blue lines after low-order kinetics is established.
Lines are plotted at time steps of 2000 years. (B)
Calcium concentration along fracture before (red)
and after (blue) breakthrough. (C) Flow rate at the
exit of the fracture for laminar (red) flow only and
for both laminar and turbulent flow (blue).
aquifer in all dimensions, including a time-dependent
water table.
As an example, we demonstrate a state-of-the-art
karst evolution model for limestone in three dimensions,
which depicts an idealized karst aquifer with horizontal
dimensions of 742.5-m length and 375-m width, and a
vertical extent of 150 m. The aquifer is discretized to
provide a fracture spacing of 7.5 m. The model consists
of 107,100 nodes, 99,000 matrix elements, and 287,831
fracture elements. The fine fractures have initial aperture
values following a log-normal distribution with mean
d am , = 0.15 mm and standard deviation of d rT = 0.20 mm.
The rock matrix has conductivities of K m = 10 10 m s 1 .
How in the karst aquifer is driven by a fixed-head
boundary condition, with heads fixed on the left and the
right sides. The head difference is 100 m. The water infil¬
trating the karst aquifer is aggressive, c,„ = 0 mol m a .
As climatic conditions we have chosen a temperature of
T = 10°C and a carbon-dioxide pressure of P = 0.05 atm,
resulting in a calcium equilibrium concentration of
c ecj = 2.12 mol nU 3 .
In Figure 5, the evolution shortly before and
some time after breakthrough is shown. Initially,
several competing sets of fractures begin evolving
from the left side. These evolving fractures push
the hydraulic heads toward the right side, as seen in
the figures before breakthrough. The right part of the
modeling domain is still a bottleneck for flow. One of
the enlarged fracture systems will be the first to
reach the right side, and the model experiences
breakthrough after around 900 years, which can
be seen by the pressure lines bouncing back. The
enlarged flow path connecting the high-pressure side
with the low-pressure side is sinusoidal, because
fractures, which initially have a larger aperture, are
less frequent. After breakthrough, a zone of enlarged
fractures evolves around the continuous flow path,
and other branches experience secondary break¬
through events, when they either connect to the exist¬
ing flow path or manage to continue growing to the
low-pressure side, depending on the more favorable
flow path.
ENCYCLOPEDIA OF CAVES
512
M
heads(m)
160 ► 180 200 220 240
^-' T . . . .
250
0.001
fractures (m)
0.01
FIGURE 5 Evolution of simple model with sta¬
tistical fracture distribution. Fixed head boundary
conditions are shown as red (high pressure)
and blue (low pressure) blocks. Heads are con¬
toured every 10 m (color code as left legend).
Enlarged fractures are shown as tubes, with the
color indicating the aperture (color code as
right legend). (A) Situation before breakthrough;
(B) situation after breakthrough.
See Also the Following Article
Hydrogeology of Karst Aquifers
Bibliography
Annable, W. K. (2003). Numerical analysis of conduit evolution in karstic
aquifers (Ph.D. dissertation). Ontario, Canada: University of Waterloo.
Bogli, A. (1980). Karst hydrology and physical speleology. Berlin:
Springer-Verlag.
Dreybrodt, W. (1988). Processes in karst systems. Berlin & New York:
Springer.
Dreybrodt, W., & Gabrovsek, F. (2003). Basic processes and mechan¬
isms governing the evolution of karst. Speleogenesis and Evolution
of Karst Aquifers, 1(1), 1—26; www.speleogenesis.info.
Dreybrodt, W., Gabrovsek, F., & Romanov, D. (2005). Processes of spe¬
leogenesis: A modeling approach. Postojna, Slovenia: Karst Research
Institute, ZRC Publishing, ZASU.
Kaufmann, G. (2009). Modelling karst geomorphology on different
time scales. Geomorphology, 106(1—2), 62—77.
Kaufmann, G., & Dreybrodt, W. (2007). Calcite dissolution kinetics in
the system CaCC> 3 —H 2 0—CO 2 at high undersaturation. Geochimica
et Cosmochimica Acta, 71(6), 1398—1410.
Kaufmann, G., & Romanov, D. (2008). Cave development in the
Swabian Alb, South Germany: A numerical perspective. Journal of
Hydrology, 349(3-4), 302-317.
Kaufmann, G., Romanov, D., & Hiller, T. (2010). Modeling three-
dimensional karst aquifer evolution using different matrix-flow
contributions. Journal of Hydrology, 388(3—4), 241—250.
Rehrl, C., Birk, S., & Klimchouk, A. B. (2008). Conduit evolution in
deep-seated settings: Conceptual and numerical models based on
field observations. Water Resources Research, 44. doi:10.1029/
2008WR006905.
Romanov, D., Gabrovsek, F., & Dreybrodt, W. (2003). Dam sites in
soluble rocks: A model of increasing leakage by dissolutional
widening of fractures beneath a dam. Engineering Geology, 70
(1-2), 17-35.
MOLLUSKS
David C. Culver
American University
INTRODUCTION
Clams, snails, and their relatives comprise the phylum
Molluska. They occur in subterranean habitats and are
sometimes common in aquatic habitats, including
Encyclopedia of Caves.
© 2012 Elsevier Inc. All rights reserved.
MOLLUSKS
caves and springs, and more occasionally in terrestrial
habitats. Nearly all obligate cave-dwelling mollusks
are gastropods. All gastropods have a muscular foot,
visceral mass, and distinct head region. A fleshy man¬
tle covers the viscera and secretes a calcium carbonate
shell. Bivalves have hinged calcium carbonate shells.
They are represented by a tiny handful of species in
caves, and there is but one undoubted cave-limited
clam, Congeria kusceri. Morphological characteristics of
obligate subterranean species include a thin, often
translucent shell that is usually white, with a depig-
mented body, and depigmented and reduced eyes.
The aquatic species of gastropods are usually very
tiny. This is true not only for species in small cavity
habitats such as the underflow of streams, but also for
large cavity habitats such as caves. Features associated
with the miniaturization of cave snails include com¬
plex coiling of the intestine, loss or reduction of gills,
simplification of gonadal morphology, and loss of
sperm sacs. Although these characteristics represent an
interesting case of convergent evolution, they also pose
a difficult challenge for any attempt to reconstruct the
phylogenetic history of cave snails. Some terrestrial
cave species can be quite large. The largest reported
terrestrial snail is that of Paraegopis oberwimmeri, a
TABLE 1 Subterranean Families and Genera
513
pulmonate snail whose shell reaches a diameter of
nearly 3 cm.
Almost nothing is known about the ecology of cave
mollusks. Because their shells remain long after the
death, many collections of cave mollusks consist of dead
shells only. This has resulted in a great deal of biogeo¬
graphic information but a lack of ecological information.
Most gastropods in surface waters are either detritus
feeders or algal feeders. Presumably, aquatic subterra¬
nean gastropods are detritus feeders. Most bivalves are
deposit or filter feeders, and the few cave bivalves are
likely also deposit or filter feeders. An amphibious habitat
seems especially common among cave snails. Nowhere is
this more apparent than in the genus Zospeum found in
Dinaric caves. Most species appear to be primarily terres¬
trial but Z. exiguum, which is common in Krizna Jama in
Slovenia, appears to be primarily aquatic.
TAXONOMIC PATTERNS OF
CAVE MOLLUSKS
Nearly all subterranean mollusks are gastropods.
The only undoubted obligate cave bivalve is the
Higher Classification
Families
Genera
Bivalvia: Lamellibranches
Dreissensiidae
Congeria
Sphaeridae
Pisidium
Gastropoda: Prosobranchia: Archaeogastropoda
Hydrocenidae
Georissa
Gastropoda: Prosobranchia: Mesogastropoda
Hydrobiidae
Over 50 genera
Pomatiopsidae
Akiyoshia, Moria, Saganoa
Assimindeidae
Cavernacmella
Cyclophoridae
Pholeoteras, Opisthostoma
Gastropoda: Pulmonata: Basommatophora
Acroloxidae
Acroloxus
Ellobiidae
Zospeum, Carychium
Gastropoda: Pulmonata: Stylommatophora
Pupullidae
Argna, Speleodiscus, Klemmia, Virpazria
Orculidae
Speleodentorcula
Enidae
Speleoconcha
Clausiliidae
Sciocochlea
Ferussaciidae
Ceciloides, Cryptareca
Subulinidae
Open
Helicodiscidae
Helicodiscus
Zonitidae
15 genera
Trigonochlamydidae
T roglolestes
Polygyridae
Mesodon, Polygyra
From Bernasconi and Riedel (1994).
ENCYCLOPEDIA OF CAVES
514 m
"living fossil" Congeria kusceri from the Dinaric karst. family Unionidae are extremely diverse in the southeast-
Minute clams in the genus Pisidium may also be cave- ern United States, which is also an area of extensive
limited species. In this genus it is very difficult to distin- cave development. Yet no clams are found in these
guish between stygobionts and surface-dwelling species caves.
that occasionally enter caves. Four species of Pisidium Nearly all aquatic subterranean gastropods are pros¬
have only been found in caves: three from caves in the obranchs. The major morphological difference between
Caucasus Mountains in Abkhazia and one from Turkey prosobranchs and pulmonates is that prosobranchs
(Bole and Velkovrh, 1986). The rarity of bivalves in possess a gill and a horny or calcareous operculum,
caves may be the result of the greatly reduced motility Pulmonates use a modified portion of the mantle cav-
of bivalves and their general inability to withstand con- ity as a lung and lack an operculum. Acroloxus from
ditions of reduced oxygen. It seems unlikely that the the Dinaric Mountains in Bosnia and Herzegovina,
absence of bivalves in caves is due to the absence of Croatia, and Slovenia and Hydraphrea from New
potential surface ancestors. For example, clams in the Zealand are pulmonates. Conversely, all troglobionts
TABLE 2 Obligate Cave-Dwelling Snails from Caves of the United States
Order Family
Species
State
Aquatic Mesogastropoda Hydrobiidae
Amnicola cola
AR
Amnicola stygia
MO
Antrobia culveri
MO
Antrorbis breweri
AL
Antroselates spiralis
KY, IN
Balconorbis uvaldensis
TX
Dasyscias franzi
FL
Fontigens antroecetes
IL, MO
Fontigens proserpina
MO
Fontigens tartarea
WV
Fontigens tnrritella
WV
Holsingeria unthankensis
VA
Phreatoceras conica
TX
Phreatoceras taylori
TX
Phreatodrobia coronae
TX
Phreatodrobia micra
TX
Phreatodrobia nugax
TX
Phreatodrobia plana
TX
Phreatodrobia punctata
TX
Phreatodrobia torunda
TX
Stygopyrus bartonensis
TX
Physidae
Physa spelunca
WY
Terrestrial Stylommatophora Carychiidae
Carychium stygium
KY, TN
Endodontidae
Helicodiscus barri
AL, GA, TN
Zonitidae
Glyphyalinia pecki
AL
Glyphyalinia specus
AL, GA, KY, TN, WV
Pristiloma cavato
CA
State distributions are shown in the final column.
From Hershler and Holsinger (1990).
ENCYCLOPEDIA OF CAVES
MOLLUSKS
515
are pulmonates, except for the prosobranch Pholeoteras
enthrix. By far the most common aquatic subterranean
species are in the family Hydrobiidae, comprising 97%
of all species. There are over 350 described species of
aquatic obligate subterranean mollusks. The number
of terrestrial cave-limited species is much less,
probably less than 50. A summary of molluskan troglo-
bionts and stygobionts is given in Table 1. What is
noteworthy is the extraordinarily high diversity of
hydrobiid gastropods at both the generic and species
levels.
BIOGEOGRAPHIC PATTERNS OF
CAVE MOLLUSKS
In the United States, there are 23 described species
of aquatic obligate cave snails and 5 species of terres¬
trial obligate cave snails (Table 2). Within the United
States, a hotspot of biodiversity is the Edwards
Aquifer in Texas. This deep karst aquifer, which is
intersected by numerous wells and caves, includes
two endemic genera, Phreatodrobia and Stygopyrus. Of
the 23 described aquatic cave snails, 11 are from
the Edwards Aquifer (Hershler and Holsinger, 1990). The
reasons for this high diversity are unclear, especially
which ecological conditions allow the coexistence of
this many snail species. It is known that unlike other
U.S. cave areas (but similar to some regions of north¬
ern Mexico), snails in the Edwards Aquifer arose not
only from old freshwater groups as did other U.S.
cave snails, but they also arose from marine ancestors
of groups during the late Cretaceous ocean embay-
ments. These species appear to be stranded relicts
that resulted from the regression of the Cretaceous
seas. The Edwards Aquifer is also a hotspot of subter¬
ranean biodiversity for other groups, especially
amphipods.
On a worldwide basis, two areas are likely to be hot¬
spots of subterranean snail biodiversity. One is the
Dinaric Mountains, ranging along the Adriatic Coast from
extreme northeast Italy to Montenegro. In the Slovenian
part of the Dinaric Mountains, for example, there are 37
aquatic obligate cave snails, 1 aquatic cave clam, and 11
species of terrestrial obligate cave snails. The entire
Dinaric Mountain region has several times that many spe¬
cies. Surface freshwater in this region is also extraordi¬
narily rich in mollusk diversity, especially among aquatic
species. A possible explanation for the richness of surface
species is the long freshwater It istory combined with the
extremely fragmented nature of surface waters. This frag¬
mentation occurs because of the high frequency of surface
streams sinking into subterranean watercourses, resulting
in increased chances for speciation. In addition, diversity
of mollusks may be higher in regions of carbonate rocks.
FIGURE 1 The hydrobiid snail Marstoniopsis croatica. Photo by
Boris Sket.
because this greatly increases the availability of calcium
carbonate for the construction of shells. The high diversity
of subsurface species in the region may be explained by
the same factors, as well as the high density of caves and
the development of extensive subterranean streams and
rivers. A sampling of the diversity of Dinaric cave snails
is shown in Figures 1 and 2.
The second hotspot of subterranean snail diversity is
East Asia, although this region has been studied less
than the Dinaric Mountains. In East Asia it is the terres¬
trial cave snail fauna that is especially rich in species.
The surface terrestrial fauna in karst areas is especially
rich as well. A single karst hill in East Asia may con¬
tain between 60 and 100 species. A single cave on an
island in Halong Bay in Vietnam yielded 17 terrestrial
cave snails, 15 of which belong to a single family
(Vermeulen and Whitten, 1999).
COMMENTS ON SELECTED SPECIES
The hydrobiid cave snail Fontigens tartarea occurs in
a dozen or so caves in West Virginia. As far as we
know, its habits are typical, at least for the genus. The
maximum linear extent of its range is 200 km, and
there is considerable variation among populations,
enough so that the northern species were at one time
considered to be a separate species. Variation includes
ENCYCLOPEDIA OF CAVES
FIGURE 2 A sampling of stygobiotic mollusks from caves in the Dinaric Mountains. (A) Anagastina hadouphylax, 4 mm; (B) Bagalivia kar-
amani, 1.7 mm; (C) Belgrandiella kusceri, 2.4 mm; (D) Cilgia dalmatica, 1.8 mm; (E) Dahnatella sketi, 2.2 mm; (F) Erytliropomatiana erythropoma-
tia, 1.5 mm; (G) Hcidziella ephippiostoma, 1.6 mm; (H) Hauffenia tellinii, 1.8 mm; (I) lstriana mirnae, 2 mm; (J) Lanzaia vjetrenicae, 1.8 mm;
(K) Marstoniopsis croatica, 3 mm; (L) Neolioratia subpiscinalis, 2.4 mm; (M) Paladilhiopsis robicicinn, 4 mm; (N) Plagigeyeria mostarensis, 3.3 mm;
(O) Zoospeum exiguum, 1.6 mm; (P) Arcoloxus tetensi, 4.5 mm. Adapted from J. Bole and F. Velkorvh (1986).
ENCYCLOPEDIA OF CAVES
MORPHOLOGICAL ADAPTATIONS
517
both size and shape of the shell, and indeed this may
be a complex of species. Shell length ranges from 1 to
2.3 mm, so even large F. tartarea are tiny. In place in a
cave stream, they look like large sand grains. It is very
sporadic in its occurrence, occupying less than 10% of
the cave streams within its range. Within cave streams
it is also very sporadic in occurrence; it is very com¬
mon in a few spots and absent in most of the stream.
Some of this patchiness may be the result of inade¬
quate collecting, but Fontigens has a reputation among
cave biologists as being patchily distributed in the
extreme. A possible explanation for this patchiness is
that they may be associated with patchily distributed
biofilms. Concentrations of F. tartarea are often on ver¬
tical faces of rocks in streams, yet another puzzle in its
distribution.
The most unusual and unique cave mollusk is cer¬
tainly the Dinaric cave clam Congeria kusceri (Morton
et al., 1990). It is related to the well-known zebra mol¬
lusk. The genus Congeria was common in the countries
of the western Balkans, Hungary, and Romania in the
late Miocene, and its near demise occurred when the
ancestral Mediterranean Sea dried up (the Messinian
salinity crisis). C. kusceri escaped extinction by coloniz¬
ing the subterranean waters that exited into the
Mediterranean Sea. Living specimens are only known
from Herzegovina and Croatia (shells are known from
Slovenia). It attaches to cave walls typically in terminal
lakes deep underground. Shell length reaches over
12 mm, and the mollusk often occurs in clusters. Shells
are often covered by precipitated calcium carbonate and
the tubes of the unique cave polychaete Marifugia cavati-
ca. It appears to be able to emerge from the water, prob¬
ably kept moist by water dripping from above. This
seasonal drying is the probable cause of the growth
rings observed on the shells of C. kusceri. Based on this
inference, C. kusceri lives to an age of 25 years, com¬
pared to 1 or 2 years for surface-dwelling species in the
same family. It also shows other demographic adapta¬
tions to cave life, possibly including internal fertilization
and brooding of eggs.
Bibliography
Bernasconi, R., & Riedel, A. (1994). Mollusca. In C. Juberthie &
V. Decu (Eds.), Encyclopedia biospeologica I (pp. 54—61).
Moulis, France: Societe de Biospeologie.
Bole, J., & Velkovrh, F. (1986). Mollusca from continental subterra¬
nean aquatic habitats. In L. Botosaneanu (Ed.), Stygofauna mundi
(pp. 177—206). Leiden, The Netherlands: E. J. Brill.
Hershler, R., & Holsinger, J. R. (1990). Zoogeography of North
American hydrobiid snails. Stygologia, 5, 5—16.
Morton, B., Velkovrh, F., & Sket, B. (1998). Biology and anatomy of
the "living fossil" Congeria kusceri (Bivalvia: Dreissenidae) from
subterranean rivers and caves in the Dinaric karst of former
Yugoslavia. Journal of Zoology, London, 245, 147—174.
Vermeulen, J., & Whitten, T. (1999). Biodiversity and cultural property
in the management of limestone resources. Lessons from East Asia.
Washington, D.C.: The World Bank.
MORPHOLOGICAL ADAPTATIONS
Kenneth Christiansen
Grinnell College
INTRODUCTION
Morphological adaptation is any evolutionary modifi¬
cation of the morphology of lineages of organisms
associated, in this case, with their existence in caves.
Stewart Peck (1998) usefully divided these changes
into two groups, regressive and progressive. Regressive
adaptations involve reduction or loss of systems
that occur in surface-dwelling organisms. Progressive
adaptations involve enlargement, modification, or devel¬
opment of systems not seen in most surface-dwelling
organisms.
TROGLOMORPHY
The term troglomorphy (Christiansen, 1962) designates
both regressive and progressive evolutionary features
associated with cave life. Though the term was origi¬
nally used only for morphological features, subsequent
study has shown that it applies equally well to behav¬
ioral and physiological features. In aquatic forms, where
underground waters are frequently not associated with
caves, the equivalent term s tygomorph is sometimes
used. These two terms serve to identify cave-adapted
organisms without the unprovable, and sometimes erro¬
neous, designation of troglobite. Troglomorphic features
are essentially the same as those discussed under the
rubric Le Facies Cavemicole (Vandel, 1964). Though char¬
acteristic of cave animals, they are by no means limited
to cave animals. Many of the regressive features associ¬
ated with the absence of light are seen in other environ¬
ments where light is absent or greatly reduced, such as
soil and the abyssal benthos. Some environments, such
as soil and microcaverns, serve as recruiting grounds
for cave organisms. Because of the lack of light, organ¬
isms in these environments rely on nonvisual methods
of food location that, combined with the common habit
of feeding on dead vegetable matter or fungi, preadapt
them to survival in caves.
Not all cave organisms display characteristics of
troglomorphism. Organisms with troglomorphic char¬
acteristics have been called cave-dependent, and those
organisms without troglomorphic features are called
Encyclopedia of Caves.
© 2012 Elsevier Inc. All rights reserved.
518
M
cave-independent. The existence of these two types of
characteristics facilitates the study of evolution within
cave organisms. Cave-independent features enable the
determination of the lineages to which different spe¬
cies belong, and the cave-dependent features, which
always display polarity, enable the determination of
their relative evolutionary position within a lineage.
EVOLUTION OF TROGLOMORPHY
Where troglomorphy does occur, it involves parallel
or convergent evolution of different lineages. This con¬
vergence can be seen clearly in the Collembolan genera
Sinella and Pseudosinella of the family Entomobryidae,
where similar foot modifications occur in many
European, North American, and Japanese lineages
(Fig. 1). Convergence can also be seen in the head flat¬
tening in different fish and amphibian families (Fig. 2)
and in physogastry and pseudophysogastry, which
occur in beetles (Fig. 3) and at least one Hemipteran.
FIGURE 1 Parallel evolution in caves of the foot complex of cave
Collembola. (A) Pseudosinella in the United States; (B), (C), (D) differ¬
ent lineages of Pseudosinella in Europe; (E), (F) different lineages of
Sinella in the United States; (G) Sinella in Japan. Least troglomorphic
species above and most troglomorphic species below.
Troglomorphic changes vary in their generality.
Many troglomorphic changes are limited to one
order or even to one or a few families, including
the foot modifications seen in Collembola (Fig. 1)
and in Hawaiian cave planthoppers. Another taxo-
nomically limited troglomorphy is the develop¬
ment of a ridge on the ventral surface of the
thorax in the tribe Leptodirini of the beetle family
Leiodidae. The clear paedomorphosis seen in the
Urodele amphibia and the pseudophysogastry of
beetles of the subfamilies Catopocerinae and Cho-
levinae of the family Leiodidae, and the subfamily
Trechinae of the family Carabidae (Fig. 3) are other
examples.
Somewhat more widespread is scale loss in some
species of four families of teleost fish. The reduction of
the swim bladder is more widespread, reduced in
some members of three families and totally lost in two
families.
Some troglomorphic features are very widespread.
Table 1 summarizes some of the major morphological
(F)
FIGURE 2 Extreme convergent troglomorphic head flattening in
cave vertebrates. (A), (B) lateral and dorsal view of the head of
Ophisternon inferniale (Synbranchidae); (C) Synocyclocheilus hyalinus
(Cyprinidae); (D) Spleoplatyrhinus poulsoni (Amblyopsidae); (E) Lucifuga
subterraneus (Bythitidae); (F) Typhlomolge rathbuni (Plethodontidae).
Adapted from Christiansen, 2004 ; and Weber, 2000.
ENCYCLOPEDIA OF CAVES
MORPHOLOGICAL ADAPTATIONS
519
FIGURE 3 Convergent pseudophysogastry in different lineages of cave beetles: (A) Trechinae, Aphaenops pluto (after Vandel, 1964);
(B) Cholevinae, Glacicavicola bathysciodes (courtesy Stewart Peck); (C) Bathysciinae, Antroherpon scutulatum (courtesy Pier Mauro Giachino); and
(D) Bathsciinae, Leptodirus hohemvarti. After Vandel, 1964.
troglomorphic features. These features include head
flattening in vertebrates (Fig. 2); thinning of the cuticle
in terrestrial arthropods; and the elongation of appen¬
dages and body form seen in some fish, amphibia, and
arthropods. Size increase is a troglomorphic character¬
istic in several orders of small arthropods, as is the
reduction or loss of wings in most troglobitic pterygote
insects. Some troglomorphic characteristics appear to
be almost taxonomically unlimited, occurring in virtu¬
ally all species that show significant troglomorphy.
This category includes pigment loss and eye reduction,
which are the most common regressive adaptations.
An increase in putative nonvisual sensory systems is
also virtually universal.
REGRESSIVE EVOLUTION
Much of the controversy surrounding the study of
morphological change in caves has been about regressive
modifications. Only a few results of evolution have cre¬
ated a wealth of theories and hypotheses comparable to
those devoted to explaining the effects of regressive evo¬
lution. The exponents of these theories often use cave
organisms as primary model systems to support their
arguments. At least 14 hypotheses have been advanced
to explain the mechanisms of regressive changes. Most
of these hypotheses are discussed in more detail in Barr
(1968; pp. 69—80). Ten such hypotheses that are of histor¬
ical or current interest will be dealt with in this article.
ENCYCLOPEDIA OF CAVES
520
M
TABLE 1 Common Troglomorphic Characteristics
Specialization of sensory organs (e.g., touch, chemoreception)
Elongation of appendages
Pseudophysogastry
Reduction of eyes, pigment, wings
Compressed or depressed body form (Hexapoda)
Increased egg volume
Increased size (Collembola, Arachnida)
Unguis elongation (Collembola)
Foot modification (Collembola, planthoppers)
Scale reduction or loss (teleost fishes)
Loss of pigment cells and deposits
Cuticle thinning
Elongate body form (teleost fishes, Arachnida)
Depressed, shovel-like heads (teleost fishes, salamanders)
Reduction or loss of swim bladder (teleost fishes)
Decreasing hand femur length, crop-empty live
weight ratio (crickets)
Five of these hypotheses are primarily of historical
interest. Three of them rely on theories that directly
oppose the now well-established core tenets of the Neo-
Darwinian thesis. These theories are (1) Lamarckism
and Neo-Lamarckism, which rely on the inheritance
of acquired characteristics; (2) orthogenesis; and (3)
Vandel's organicism. Organicism states that lineages go
through three stages: birth, specialization, and senes¬
cence. Under this model, degenerative changes are
characteristic of lineage senescence. Thus, Vandel
hypothesizes that all cave species are in the senescent
evolutionary stage. They show regressive evolution, not
because of environment or selection, but because of
their age status in the lineage. No recent evidence has
been presented to support Vandel's theory.
Two other hypotheses of historical interest are
Lankester's escape hypothesis and Ludwig's trap
hypothesis. The escape hypothesis states that weak-
eyed organisms falling into caves were unable to
escape. The trap hypothesis holds that weakly pig¬
mented surface forms would be sensitive to light and
thus would concentrate in caves. These two hypothe¬
ses were never widely supported; instead they were
dealt a deathblow by the work of Fong and Culver
(1985) which showed large-eyed individuals of a
highly variable amphipod population were actually
more rather than less photophobic than small-eyed
individuals. None of these five hypotheses is exten¬
sively studied today, but a few are being pursued.
Two explanations still being considered have
received no recent support. The first of these is mate¬
rial compensation. This hypothesis comes in many
forms but all of them presuppose that, in the food-
poor cave environment, selection will favor the most
economical use of the energy furnished by food and
thus will favor reduction of useless structures or pro¬
cesses. This hypothesis has run into difficulties. Recent
discoveries show that regressive evolution can occur in
cave environments where there is an abundant supply
of food. These environments include some tropical
caves and caves where chemoautotrophic bacteria fur¬
nish ample food supply. Even in caves with limited
food, the evidence to support a theory of material com¬
pensation is weak. As Poulson (1963) noted in his
detailed study of this problem, the calculations do not
clearly support or disprove the idea of selective reduc¬
tion of "useless" traits as having a positive effect on
energy economy. In the cavefish Astyanax the reduc¬
tion of the eyes does not appear to result in energy
savings as measured by egg yolk consumption in
embryos. Moreover, the caves where blind fish are
found are reported to be food-rich. More data on the
genetics of cave organisms are needed to settle this
debate.
Another explanation is Fleuts' (1951) negative
allometry hypothesis, in which he posits that organs
appearing early in development and growing rapidly
at first, such as eyes, are more likely to be negatively
affected by lowering growth rates. This thesis is sup¬
ported by some data but is generally more descriptive
than explanatory.
Most recent studies have involved either the
increase in the number of neutral genes having a dis¬
ruptive and/or reductive effect on functionless organs,
associative selection with adaptive structures, or some
combination of the two. Poulson (1986) and others
have strongly supported the disruptive effect of the
accumulation of neutral mutations as the most plausi¬
ble explanation for regressive evolution. The fact that,
in the course of increasing troglomorphy, progressive
features tend to show less variability than regressive
ones would tend to support the accumulation of genes
in regressive structures. Sket (1985) challenges the gen¬
erality of this increase in variability. Culver and
Wilkens (2000), in their recent work on the topic, sug¬
gested that the accumulation of mutated genes and
associative selection reinforce each other in regressive
evolution. Flypotheses of this sort afford the best avail¬
able explanation for regressive evolution.
There are still problems. First, it is questionable
whether any single hypothesis can explain all of the
aspects of regressive evolution. Moreover, a number of
ideas exist concerning regressive evolution that has not
been tested in cave animals. One such idea is Regal's
ENCYCLOPEDIA OF CAVES
MORPHOLOGICAL ADAPTATIONS
521
(1977) noise suppression hypothesis. This hypothesis
posits, on a basis of cybernetic theory, that natural
selection is constantly acting to squelch nonessential
messages within developing organisms. This action
reduces metabolic noise and hence improves evolution¬
ary fitness. In addition, a few earlier hypotheses, such
as orthogenesis, have recently been revisited and
revised. Several students of evolution suggest that this
concept is misunderstood and actually denotes that
there is a limited array of variations, which can occur
and be functional in any group of organisms. This
results in the fact that very similar characteristics occur
under different ecological conditions and independently
of selection. Regressive evolution remains a fruitful
field for future research. The study of progressive modi¬
fications in cave life has produced less controversy.
PROGRESSIVE TROGLOMORPHY
AND ADAPTATION
Though some progressive evolutionary modifica¬
tions may be nonadaptive, they most putatively
increase their owner's chances of survival or competi¬
tiveness in caves. There have been a number of
works dealing with such adaptations in specific groups
(Christiansen, 1965; Poulson, 1963; Hobbs, 2000;
Niemiller and Poulson, 2010). The general topic of cave
adaptation has also received a number of reviews
(Vandel, 1964; Barr, 1968; Culver, 1982; Culver et ah,
1995; and Culver and Pipan, 2009). External morpholog¬
ical adaptation has been studied far more than internal.
Regressive troglomorphy is commonly found in
troglophilic forms, but progressive troglomorphy is
usually limited to troglobitic or nearly troglobitic forms
(those very rarely found outside caves). Progressive tro¬
glomorphy is, however, not universal among such cave
organisms. In order for it to occur, three factors have to
be present: (1) selection pressure for the development
of a particular characteristic; (2) the genetic and physio¬
logical or behavioral ability of the organism to respond
to this pressure; and (3) sufficient time evolving in
caves to develop the adaptations. Many cave organisms
lack one or more of these factors. Indeed, the majority
of troglophiles and some troglobites show little or no
progressive troglomorphy, probably because one or
more of the factors necessary to morphological adapta¬
tion is absent. Thus, there are many species of fish and
crayfish found in caves that show no troglomorphism.
Many extremely edaphic Collembola, such as the mem¬
bers of the family Onychiuridae, already unpigmented
and eyeless in all surface habitats, rarely show further
troglomorphy, although they have many troglobitic
species. In many other groups, troglomorphy is ques¬
tionable or inconsistent (Culver, 1982). Troglomorphy is
also usually absent in some cave environments, such as
localized guano piles or large masses of organic debris
that are extremely energy-rich. Peck (1973) pointed out
that troglomorphy should not be normal in these cases,
because strong troglomorphy only occurs when organ¬
isms are exploiting large-volume spaces such as cave
wall or floor surfaces or large bodies of water. The
cases where troglomorphy occurs in food-rich regions
(some tropical caves and chemoautotrophic caves) most
commonly involve only the regressive features of eye
and pigment reduction. This fact supports the hypothe¬
sis that the mechanisms of regression differ from those
of progressive adaptation. It is interesting that Boutin
and Coineau (2000) have found something similar to
troglomorphy in phreatobites, which live in narrow
spaces, where small size and elongate body form are
developed convergently.
EFFECTS OF ENVIRONMENTAL
FACTORS ON ADAPTATION
There is a strong correlation between the environ¬
mental factors found in caves and the development of
progressive troglomorphy. While earlier works sug¬
gested non Neo-Darwinian explanations for adaptation
in cave animals (Vandel, 1964), almost all recent work
has been done with the clear assumption of the applica¬
bility of the Darwinian core tenets. A possible recurrence
of a view similar to that held by Vandel can be seen in
the emphasis on phenotypic plasticity espoused by
Romero (2009); however, the fact that similar positive
troglomorphic features evolve in many different cave
regions and lineages implies an adaptive control over
their genetic development.
The nature of this adaptive value is sometimes
completely unknown, as with the mesosternal carina
development in beetles and the flattening of heads in
fish and salamanders. In some cases, various untested
hypotheses for these adaptive values have been
advanced. For example, the large air space under the
elytra in pseudophysogastry has been theorized to
serve a respiratory function by increasing the air expo¬
sure of the thin abdominal membrane and by serving
as a bubble lung during flooding. It has also been
explained as a flotation device. In other cases the adap¬
tive value appears to be self-evident, as with the
increase in motion sensory organs in fish and elonga¬
tion or hypertrophy of tactile organs generally. In some
cases these apparently adaptive functions are associated
with specific habits and/or habitats. Thus, the
extremely delicate or gracile body form and elongate
antennae of the cave crayfish (Fig. 4) Troglocambarus
maclanei is associated with its normal habitat on walls
or ceilings of flooded cave passages.
ENCYCLOPEDIA OF CAVES
522
M
FIGURE 4 Extreme troglomorphy in the crayfish Troglocambarus
maclanei. Adapted from Hobbs, 2000.
There is general agreement that cave habitats are
very specialized and, in some respects, very demand¬
ing. The normal scarcity and sporadic availability of
food, the absence of light, and the difficulty of long-
range dispersal pose problems. On the other hand, the
relatively high and constant humidity, absence of
wind currents, and limited predation and parasitism
represent considerable advantages. All of these lead to
selection pressures, which do not occur in surface
habitats, and release from some selection pressures,
which do occur there. These changes are consistent
among caves all over the world.
EVIDENCE OF TROGLOMORPHIC
ADAPTATION
Many recent studies have furnished direct evidence
to support the adaptive value of troglomorphic changes
(Christiansen, 1965; Culver et al., 1995; and Poulson,
1963). These works indicate that those troglomorphic
FIGURE 5 The species of Amblyopsidae showing increasing
neuromast number: (A) Chologaster comutus; (B) Forbesichthys agassizi;
(C) Typhlichthys subterraneus; (D) Amblyopsis spelaea; (E) Amblyopsis
rosae ; and (F) Speoplatythinus poulsoni. Adapted from Weber, 2000.
features that have been studied are subject to selection
and are, in fact, adaptive.
In spite of all this, few studies have produced data
to support the precise adaptive functions of specific
troglomorphic developments. One of the best-studied
examples of the adaptive function of troglomorphic
change is that of the cavefish of the family Amblyopsi¬
dae. Here the great increase in neuromast motion sen¬
sory organs (Fig. 5) is associated with an increased
ability to locate prey and with greater success in cap¬
turing it. This adaptation, combined with a variety of
behavioral, physiological, and developmental changes,
increase the chances of survival in cave waters. A sec¬
ond example occurs in the beetles of the subfamily
Trechinae. Here the cave species that are crevice fee¬
ders tend to be small and flattened, whereas the
ENCYCLOPEDIA OF CAVES
MORPHOLOGICAL ADAPTATIONS
523
cursorial feeders tend to be larger pseudophysogastric
types (Fig. 3). Still another example is in the family
Entomobryidae of the Collembola, where laboratory
studies have shown that the foot adaptations seen in
increasing troglomorphy are associated with behav¬
ioral changes that first allow for better movement over
wet stone surface, then make for increased efficiency
in moving over wet clay, finally resulting in an ability
to walk on water surfaces (Fig. 6). In the course of this
evolution, each change in behavior (the need to walk
on wet stone, wet clay, or water) precedes a change in
the direction of the development of the foot structure.
Thus every troglomorphic characteristic in these cases
is associated with the ability to deal with the particular
environmental conditions found in caves.
The function of troglomorphic modification is also
shown by analogous modifications in other habitats.
For example, the extremely troglomorphic form of
Collembola is most closely approached in two very dif¬
ferent surface habitats. The foot structure is closest to
that seen in aquatic Collembola, but the body shape
and size are most similar to those seen in forms (largely
tropical) that live in trees above the litter or soil. The
FIGURE 6 Morphological and behavioral changes seen in the
process of adaptation of the foot complex of cave Collembola in the
genus Pseudosinella: (A) typical weak penetration of clay seen in non-
troglomorphic species; (B) intermediate troglomorphy showing
increased penetration of clay; (C) extreme troglomorphy, capable of
walking on water surfaces; (D), (E) nontroglomorphic position on
smooth, wet, hard surfaces; (F) sideways position in forms having
intermediate troglomorphy; and (G) typical foot position on hard,
wet surfaces in highly troglomorphic species.
foot structure is clearly related to the widespread occur¬
rence of wet clay in caves, which makes movement dif¬
ficult, and to the rimstone or other water pools, which
act as traps for many Collembola. The body shape and
size changes are probably associated with a method of
predator escape. The freedom of movement furnished
by open cave surfaces is more similar to conditions
found in the arboreal habitat than to the normal surface
habitat in soil or leaf litter where cave forms are
recruited. Another example of analogous modification
is the hypertrophy of organs for nonvisual location of
food in both deep-sea and cave-dwelling fish. Wing
reduction or loss, which is characteristic of cave insects,
is also seen in environments such as soil and high
alpine areas where wings are nonadaptive. The reduc¬
tion of eyes and pigment in cave forms occurs also in
dark edaphic and microcavernicole habitats. The loss of
the epicuticular wax layers seen in caves is also seen in
edaphic habitats where humidity is consistently high,
as it is in habitable caves.
OCCURRENCE OF MORPHOLOGICAL
TROGLOMORPHY
One of the mysteries of the evolution of cave organ¬
isms is the fact that only certain groups of any major
taxon evolve morphologically within caves. In some
cases the opportunity to evolve is simply lacking due
to dietary restrictions or lack of entrance to caves. Yet
there are many instances where neither restriction
applies. In Collembola, most genera do not reproduce
in cave depths, but many occasionally do in entryways
where large amounts of organic debris have accu¬
mulated. Members of the families Isotomidae and
Sminthuridae are often found in deeper reaches of
caves where piles of organic debris have accumulated,
yet no genus of the Isotomidae and only one genus of
Sminthuridae ( Arrhopalites ) have evolved a lineage of
cave species. Similarly, the order Psocoptera, with
more than 3800 species, has only 8 cavernicole species,
none of which is clearly troglomorphic. In beetles,
the highly variable families Scarabidae, Ptiliidae, and
Tenebrionidae have no troglomorphic species; there is
no a priori reason to expect this circumstance. In the
widely varied family Carabidae, only 4 or 5 of the 40
tribes have developed troglomorphic forms.
The majority of the troglomorphic species of the
beetle family Carabidae belong to a single tribe, the
Trechini. Similarly, in the 17 tribes of the beetle family
Leiodidae, only 3 tribes harbor troglobitic species
and most of these belong to the tribe Leptodirini
(Bathyscini). In freshwater teleost fish, about half
the families found in caves show no species with tro¬
glomorphic features or any evidence of derivative
ENCYCLOPEDIA OF CAVES
524
M
forms with them. Similarly, in the crustacean order
Amphipoda, the large Holarctic family Gammaridae,
with over 50 genera and over 300 species, has only
14 genera and 33 species that are troglobitic or stygobi-
tic. In contrast, the family Crangonyctidae, with only
6 extant genera and about 145 species, has 5 troglo-
morphic genera and 116 such species. Even more strik¬
ing is the Palearctic family Niphargidae with 8 genera
and 207 species, all troglobitic or stygiobitic.
One possible explanation of these cases of failure in
the evolutionary exploitation of caves is the absence of
necessary exaptations or preadaptations. It has been
generally agreed that preadaptation or exaptation is
required for successful passage from trogloxene to
troglophile. This idea is largely based on inference; lit¬
tle strong evidence to support it has been produced. It
is true that many terrestrial troglobite lineages are
recruited from habitats such as soil and microcaverni-
cole cavities where suggested preadaptations or exap¬
tations are developed. For example, soil organisms
have to deal with survival in the absence of light and,
and for most groups, poor food sources. Both of these
conditions could serve as preadaptations for cave life.
In addition, soil habitats usually have high humidity,
as do most biologically active cave habitats.
Four facts argue against this idea as a major explana¬
tion. First, each of these environments is significantly
different from that of caves. For example, food supply
in soil is generally widely and continuously available
but also very dispersed. In caves, the supply is usually
concentrated and locally available but sporadic or peri¬
odic. Second, movement in soil is severely limited,
where it is virtually unlimited for small organisms in
caves. Third, there are many cases where the most prim¬
itive troglobite members of a cave lineage do not show
morphological evidence of these preadaptations or exap¬
tations. Fourth and finally, even if these exaptations or
preadaptations are adequate to explain the successful
invasion of caves, there are many cases of groups with
similar potential preadaptations that are not successful.
Examples exist in the Carabid beetles and in the
Collembola. In the Carabidae, members of the tribe
Anillini are eyeless and depigmented. Although most
are edaphic, only one species in Alabama appears to be
a true troglobite. This contrasts with the tribe Trechini,
which has hundreds (if not thousands) of troglobitic
species. In the Collembola, the family Entomobryidae
has hundreds of troglobites, some showing the highest
degree of troglomorphy. In contrast, the Isotomidae,
with hundreds of depigmented, reduced-eye, or eyeless
edaphobites, has only two troglobitic species, neither of
which is strongly troglomorphic.
A possible explanation for these two cases is that
physiological and/or behavioral preadaptations rather
than morphological changes are crucial for survival in
caves. The best evidence to support this explanation
comes from the fish families with a large percentage of
stygiobiont forms. In the family Amblyopsidae, where
5 of the 6 species are found in caves, the single non¬
cave species lives in stable habitats, shows low activity
level, is nocturnal, and has the ability to feed and ori¬
ent itself in the dark. All these features could be adap¬
tive in a cave environment. The surface-living forms of
the family Synbranchidae include about 17 species,
4 of which are found in underground waters and show
similar probable preadaptations. They are mostly light¬
avoiding species feeding at night and able to move
through small spaces. Another piece of evidence to
support the idea of nonmorphological preadaptation is
the widespread existence of opportunistic troglophile
species that lack morphological troglomorphism and
are able to survive in caves. These species thrive where
they do not face competition from troglomorphic spe¬
cies and where food resources are abundant.
More definitive evidence concerning the role of
preadaptation or exaptation in the development of
troglomorphy could be obtained by comparing groups
successful in exploiting caves with those that are not
successful. For instance, compare the behavior and
physiology of surface species of the highly troglo¬
morphic fish family Amblyopsidae and Synbranchidae
with those of the family Centrarchidae that has 18 spe¬
cies, 11 of which are found in caves, none showing any
troglomorphy. In the family Cichlidae, with over 900
species with only 2 reported in caves, there are no
troglomorphic species. For a terrestrial example, com¬
pare the following: the collembolan genus Folsomia,
with 190 primarily soil species, many with reduced
eyes and pigment, has no troglomorphic cave species,
and the genus Pseudosinella contains 280 species, the
majority of which are troglomorphic troglobites.
Where troglomorphy does occur, it allows for a sepa¬
rate analysis of those features clearly affected by the
cave environment (cave-dependent) and those that are
unaffected by it (cave-independent). The cave-depen¬
dent features show a great deal of convergence and par¬
allelism, giving clear polarity for phylogenetic analysis.
This permits a measure, or least an indication, of the
degree of cave adaptation of different groups of organ¬
isms. An unfortunate corollary is that the large amount
of resultant homoplasy makes cladistic analyses of cave
animals difficult. Nevertheless, the existence of these
two types of features makes caves an excellent natural
laboratory for studying evolutionary processes.
The superficial underground environment (MSS) is a
specialized microcavernicole habitat that is frequently
in direct contact with caves. Here we find many, but
by no means all, of the fauna of caves. In a sense, this
can be looked on as an extension of the cave habitat
but with an absence of bodies of water or large open
ENCYCLOPEDIA OF CAVES
MORPHOLOGICAL ADAPTATIONS
525
spaces. Because this environment is not well studied, it
furnishes a good field for future research.
EXAMPLES OF DETAILED ANALYSIS OF
TROGLOMORPHIC EVOLUTION
Wherever progressive troglomorphy occurs, it is
clear that understanding the factors that produce it
requires detailed study of the organisms involved. In
the next section, two such studies will be considered
and what they tell us about the development of these
adaptations will be discussed.
The first study deals with interspecific troglomorphic
changes among the members of the North American tel-
eost fish family Amblyopsidae. This small family has six
known species (Fig. 5), all limited to the southeast quad¬
rant of the United States: one surface species ( Chologaster
cornuta), one troglophile ( Forbesichthys agassizi), and four
troglobite species ( Typhlichthys subterraneus, Amblyopsis
spelaea, Amblyopsis rosae, and Speoplatyrhinus poulsoni).
The single surface species is not found near any region
occupied by the troglophile or troglobite species of the
family, but it is found in slow streams and swamps on
the coastal plain from southern Virginia to southern
Georgia. The other species have either overlapping or
nearly adjunct ranges (Fig. 7). Although many have
studied these fish, Thomas Poulson did the definitive
work (1963). There is a clear gradient of increasing trog¬
lomorphy as we move from C. cornuta to S. poulsoni. The
troglomorphic changes involve eye, pigment, and optic
lobe reduction; increase in neuromast number, size, and
exposure; hypertrophy of tactile, olfactory, and equilib¬
rium receptors and their associated brain centers; and
enlargement and flattening of the head. The single
genetic study of the members of this family points to a
multiple origin for different populations of the least
troglomorphic of the troglobite species, T. subterraneus.
This conclusion is reinforced by the wide and disjunc¬
tive distribution of this species. Thus, the evidence is
strong that troglomorphic features seen in T. subterraneus
were developed by parallel evolutionary changes.
Whether the increasingly troglomorphic features of the
remaining troglobite species were developed from some
Typhlichthys- like ancestor or from now extinct troglophilic
forms remains in doubt. The existence of dines
in A. spelaea from south to north with the most troglo¬
morphic condition occurring in the southern populations
would seem to support the former hypothesis. In addi¬
tion, the extreme troglomorphy of S. poulsoni argues for
ancestry from some already troglomorphic form. In any
case, the widely disjunctive distributions of the highly tro¬
glomorphic A. rosae and A. spelaea strongly support the
parallel evolution of their troglomorphic features.
FIGURE 7 Distribution by county of Amblyopsidae in the eastern United States. Only the swampfish Chologaster cornuta is found outside
the Ozarks and Interior Low Plateaus. From Niemiller and Poidson, 2010.
ENCYCLOPEDIA OF CAVES
526
M
FIGURE 8 Members of the Pseudosinella hirsuta lineage: (A) P. dubia; (B) P. hirsuta ; and (C) P. christianseni.
A second notable feature is the lack of congruent
development of troglomorphic features. The overall
gradient for troglomorphy is clear but only one feature,
olfactory rosette size, follows this gradient strictly.
Eye size shows a sharp reduction from C. cornuta to
S. poulsoni, but the eyes of T. subterraneus are smaller
than those of the generally more troglomorphic
A. spelaea. Similarly, although there is a striking
increase in neuromast number from C. cornuta to
S. poulsoni, the number in Amblyopsis is smaller than
that of Typhlichthys. A third notable feature is the
increasingly limited distribution with increasing troglo¬
morphy in the troglobitic species (Fig. 7). Typhlicthys
subterraneus, the least troglomorphic of these, is scat¬
tered over five states. In contrast, S. poulsoni, the most
troglomorphic, is found in only one cave system.
The second study involves the troglobitic collembolan
Pseudosinella hirsuta, studied in great detail by
Christiansen and Culver (1968). This study involved the
geographic variation of 25 populations of P. hirsuta
using 12 cave-dependent and 3 cave-independent mor¬
phological features as well as one behavioral feature.
This species belongs to a lineage of three species (Fig. 8)
including P. dubia and P. christianseni. P. dubia shows lit¬
tle progressive troglomorphy and has a widely disjunc¬
tive distribution of three caves in Washington County,
Arkansas, and one cave in Dent County, Missouri.
Pseudosinella hirsuta is common and widespread in caves
of central Kentucky and Tennessee. It has some popula¬
tions in caves of northeast Alabama and two in adjacent
northwest Georgia. In addition, there are populations in
eastern Tennessee and the westernmost part of Virginia
(Fig. 9). Assuming that the troglophile ancestor of P. hir¬
suta resembled P. dubia, P. hirsuta already shows signifi¬
cant increase in troglomorphy. In some other features,
such as the foot structure, P. hirsuta shows a wide range,
varying from one similar to P. dubia to one similar to P.
christianseni. The most troglomorphic of the species, P.
christianseni, occurs in small, scattered locations in the
eastern part of the range of P. hirsuta.
All cave-dependent features of each cave population
show similar states of advanced or primitive cave
adaptation, with the exception of the extremely
Dripping
Region
Eastern Highland Rim v::
Central Basin
Cum berla nd Pla l ca u
Valley & Ridge
Centers of Novelty & Pathways
probable
possible o»itt>
Possible Mixing M
Caveless Areas □
FIGURE 9 Centers of novelty and pathways of evolutionary
distribution of Pseudosinella hirsuta.
phenotypically plastic mesothorax. On this basis, each
population can be placed in one of several sequences
of advanced to primitive conditions. Each sequence is
largely or entirely limited to a single geological region
with a single evolutionarily advanced region and one
to five primitive regions, generally marginal (Fig. 9).
The populations' cave-independent features do not fol¬
low the same pattern, but do give some indication of
decreasing variability from advanced to primitive
populations. Each sequence of caves contains most or
all of the variations seen within the species.
The main conclusions from this study regarding
troglomorphism are as follows.
1. The characteristics of Pseudosinella hirsuta have evolved
independently in at least four different places.
2. Pseudosinella christianseni represents a continuation
of the evolutionary trends seen in P. hirsuta. The
two species have never been found in the same
ENCYCLOPEDIA OF CAVES
MORPHOLOGICAL ADAPTATIONS
527
cave, but where they exist in caves near each other,
the populations of P. hirsuta are usually among the
most highly troglomorphic seen in the species.
3. Dispersal of P. hirsuta occurs primarily via underground
routes. There is no indication of morphological
discontinuity between populations in a given cave
sequence.
4. There is morphological evidence for genetic
discontinuity between different cave sequences.
5. Most importantly, it appears that the extremely
troglomorphic P. christianseni evolved at four
different times from P. hirsuta -like ancestors.
6. The ranges of the troglobitic species tend to reduce
with increased troglomorphy.
7. The putative ancestral species is no longer found
near the ranges of the more troglomorphic species.
These patterns are all seen in many other groups of
cave organisms but seldom with the detail and clarity
seen in these two examples.
CONCLUSIONS
1. Morphological cave adaptation or troglomorphy is
seen in many, but not all, troglobitic animals.
2. The development of troglomorphy requires a number
of conditions and is still not clearly understood.
3. Troglomorphy can be either regressive, involving
the reduction or loss of features, or progressive,
involving alteration or expansion of preexisting
features or the development of new features.
4. Progressive and regressive troglomorphy probably
involve different combinations of mechanisms.
5. All troglomorphic development involves extensive
parallelism and convergent evolution.
6. The occurrence of these troglomorphic
(cave-dependent) features alongside
nontroglomorphic (cave-independent) features,
which show no parallelism or convergence, makes
caves an ideal environment in which to study
evolution.
7. These trends are associated with the unusual, and
often demanding, environmental conditions of
caves, which are very different from those in
surface habitats but similar in the great majority of
caves throughout the world.
8. The common result of these environmental
conditions is that the initial development of
troglobitic forms results in equivalent levels of
troglomorphy evolving in many different places.
9. The higher the level of troglomorphy within
troglobitic groups, the more restricted the
distribution of the species and the more limited
their geographic range.
10. Forms showing low levels of troglomorphy are
often troglophilic, whereas those showing high
levels are almost always troglobitic.
11. Although it appears likely that development of
troglomorphy is facilitated by preadaptations or
exaptations for cave life, the reasons why only a
few groups of organisms develop this remain
unclear.
See Also the Following Articles
Adaptation to Lozv Food
Astyanax mexicanus: A Model Organism for Evolution
and Adaptation
Gammarus minus: A Model System for the Study of
Adaptation to the Cave Environment
Natural Selection
Bibliography
Barr, T. C. (1968). Cave ecology and the evolution of troglobites.
Evolutionary Biology, 2, 35—102.
Boutin, C., & Coineau, N. (2000). Evolutionary rates and phyloge¬
netic age in some stygobiontic species. In H. Wilkens, D. C.
Culver & W. F. Humphreys (Eds.), Subterranean ecosystems
(pp. 433—451). Amsterdam: Elsevier.
Christiansen, K. A. (1962). Proposition pour la classification des ani-
maux cavernicoles. Spelunca, 2, 76—78 (in French).
Christiansen, K. A. (1965). Behavior and form in the evolution of
cave Collembola. Evolution, 19, 529—537.
Christiansen, K. A. (2004). Adaptation: Morphological (external).
In J. Gunn (Ed.), Encyclopedia of cave and karst science (pp. 7—9).
London: Fitzroy Dearborn.
Christiansen, K. A., & Culver, D. C. (1968). Geographical variation
and evolution in Pseudosinella. Evolution, 2, 237—255.
Culver, D. C. (1982). Cave life. Cambridge, MA: Harvard University
Press.
Culver, D. C., Kane, T. C., & Fong, D. W. (1995). Adaptation and natu¬
ral selection in caves. Cambridge, MA: Harvard University Press.
Culver, D. C., & Pipan, T. (2009). The biology of caves and other subter¬
ranean habitats. Oxford: Oxford University Press.
Culver, D. C., & Wilkens, H. (2000). Critical review of relevant theo¬
ries of the evolution subterranean animals. In H. Wilkens, D. C.
Culver & W. F. Humphreys (Eds.), Subterranean ecosystems
(pp. 389—407). Amsterdam: Elsevier.
Fong, D. W., & Culver, D. C. (1985). A reconsideration of Ludwig's
differential migration theory of regression evolution. Bulletin of
the National Speleological Society, 47, 123—127.
Heuts, M. J. (1951). Ecology, variation, and adaptation of the blind
African cave fish Caecobarbus geertsi. Annales de Societe Royal
Zoologie Belgique, 82, 155—230.
Hobbs, H. H. (2000). Subterranean organisms—Crustacea.
In H. Wilkins, D. C. Culver & W. F. Humphreys (Eds.),
Subterranean ecosystems (pp. 95—108). Amsterdam: Elsevier.
Niemiller, M. L., & Poulson, T. L. (2010). Subterranean fishes of
North America: Amblyopsidae. In E. Trajano, M. E. Bichuette &
B. G. Kapoor (Eds.), Biology of subterranean fishes (pp. 169—280).
Enfield, NH: Science Publishers.
Peck, S. B. (1973). A systematic revision and the evolu¬
tionary biology of the Ptomaphagus ( Adelops ) beetles of North
America (Coleoptera, Leiodidae, Catopinae), with emphasis on
ENCYCLOPEDIA OF CAVES
528
M
cave-inhabiting species. Bulletin of the Museum of Comparative
Zoology, 145(2), 29-162.
Peck, S. B. (1998). Cladistic biogeography of cavernicolous
Ptomaphagus beetles (Leiodidae, Cholevinae: Ptomaphagini) in the
United States. In: Proceedings of the 20th International Congress of
Entomology (pp. 235—260). Torino, Italy: Museo Regionale di
Scienze Naturali.
Poulson, T. L. (1963). Cave adaptation in amblyopsid fishes.
American Midland Naturalist, 70(2), 257—290.
Poulson, T. L. (1986). Evolutionary reduction by neutral mutations:
Plausibility arguments and data from amblyopsid fishes and liny-
phiid spiders. Bulletin of the National Speleological Society, 47,
109-117.
Regal, P. J. (1977). Evolutionary loss of useless features: Is it molecu¬
lar noise suppression? American Naturalist, 111, 123—133.
Romero, A. (2009). Cave biology. Life in darkness. Cambridge, U.K.:
Cambridge University Press.
Sket, B. (1985). Why all cave animals do not look alike: A discussion
of the adaptive value of reduction processes. Bulletin of the
National Speleological Society, 47(2), 78—85.
Vandel, A. (1964). Biospeleologie—La Biologie des Animaux Cavernicoles.
Paris: Gauthier Villars (in French).
Weber, A. (2000). Subterranean organisms—Fish and amphibia.
In H. Wilkens, D. C. Culver & W. F. Humphreys (Eds.),
Subterranean ecosystems (pp. 109—113). Amsterdam: Elsevier.
MULTILEVEL CAVES AND
LANDSCAPE EVOLUTION
Darlene M. Anthony
Roane State Community College, Harriman, TN
MULTILEVEL CAVES
Cave explorers were among the first to note that
some caves were found at different levels within a
hillside or valley wall. The opportunity to enter a
cave via one entrance, journey through extensive hori¬
zontal passageways connected to each other by nar¬
row canyons, and exit via a second, lower entrance
was considered by many to be the ultimate trip in
sport caving. The notoriety of multilevel (or tiered)
caves increased with the overall depth of the entire
system, or with the number of levels or entrances that
could be traversed during a single trip underground.
When multilevel cave passages were surveyed and
drafted by cartographers, the map often resembled a
jumble of ribbon in plan view. However, when
viewed from the side, or in vertical profile, the pas¬
sages fell neatly into narrow zones of horizontal
development connected in a few key places by verti¬
cal passages. This particular pattern of cave develop¬
ment is found in many of the world's karst
landscapes, where scientists are interpreting the role
of multilevel caves in landscape evolution.
CAVES AND THE WATER TABLE
Why do some caves develop levels, whereas others
do not? What characteristics do multilevel caves share?
How many years does it take to form a level as large
as a subway tunnel? A basic understanding of the rela¬
tionship between horizontal conduits and the regional
hydrologic system is vital to answering these ques¬
tions. In a karst terrain, caves typically form where
slightly acidic water finds flow paths (such as joints,
bedding planes, or faults) in carbonate rocks, directing
surface water downward to the regional water table.
Caves often provide the fastest (and sometimes short¬
est) path for surface water to reach the regional base
level, which is represented by major regional rivers.
Because caves are formed by flowing water, cave
development is strongly controlled by base-level eleva¬
tion. In the vadose zone (above the base level), under¬
ground streams carve narrow canyons that lead
downward in the fastest manner possible until the
phreatic zone (at or below the base level) is reached.
At the phreatic zone, cave streams form nearly hori¬
zontal tubes as water flows to a discharge outlet,
which in many cases is a spring at or near the regional
river. When surface rivers remain at the same eleva¬
tion for long periods of time, called stillstands, cave
streams develop large, horizontal passages at grade
with the rivers (Fig. 1). When any event (such as a
drop in sea level or tectonic uplift) causes the regional
rivers to suddenly lower, or incise, the sudden differ¬
ence in elevation creates disequilibria between the
cave stream and the newly lowered base level. The
increased hydraulic gradient causes the cave stream
(now in the vadose zone) to preferentially downcut
narrow canyons in the floor of the horizontal conduit,
abandoning the upper level to keep grade with the
FIGURE 1 Generalized representation of a multilevel cave and
its relationship with the regional base level shows the extensive
horizontal conduits that form during long periods of base-level
stability. These large passages are quickly abandoned in favor of
lower levels during periods of accelerated river incision.
Encyclopedia of Caves.
© 2012 Elsevier Inc. All rights reserved.
MULTILEVEL CAVES AND LANDSCAPE EVOLUTION
529
new base level. Differences between phreatic tubes
and vadose canyons are often striking when exploring
multilevel caves, and the vadose—phreatic transition
clearly indicates a previous position of the regional
base level punctuated by a sudden incision episode.
RIVER INCISION AND
THE RECORD IN CAVES
Karst geomorphologists, who examine the relation¬
ship between cave levels and the landscape, recog¬
nized that as surface rivers incised, the regional base
level lowered, leaving formerly active cave passages
dry and forming new cave passages beneath the older.
Abandoned cave levels were often found to coincide
with the scars of ancient terraces and remnants of
gravel deposits found on the surface of the river val¬
ley. But when did the river incise, abandoning the
cave levels? Rivers are by nature a destructive force
in the landscape, removing many of the depositional
features that mark their passage. Reconstructing the
position of a surface river over geologic time is made
even more difficult by the weathering and erosion of
remnant fluvial deposits and by the lack of appropriate
methods (such as carbon-14) for determining the abso¬
lute age of unconsolidated gravel and sand.
In sharp contrast to surface streams, multilevel
caves contain vast piles of sediments deposited by
underground streams (Fig. 2). Subterranean capture of
surface streams at sink points directs gravel, sand, and
silt from surface highlands into caves, where sediment
is transported along developing horizontal conduits to
output springs. Banks of sediments are deposited in
conduits in the same manner as sediments in surface
streams, complete with flood deposits, channel cut-
and-fill deposits, and other fluvial features. When an
underground stream abandons a conduit in favor of a
lower level, sediments left behind may remain undis¬
turbed for millions of years, unaffected by either ero¬
sion or weathering. Because sediments in abandoned
conduits were deposited by moving water, they repre¬
sent the last time the conduit was an active part of
the local hydrologic system. Conduit elevation and
the timing of sediment burial serve as proxies for the
paleoelevation of the water table over geologic time.
But how long ago was the sediment washed into the
cave? This important question may now be answered
with a new method of absolute age dating that com¬
pares the radioactive decay of atoms called cosmogenic
nuclides in quartz-rich sediment. The following exam¬
ples illustrate how karst geomorphologists are using
burial dates from cave sediments to interpret the Plio-
Pleistocene history of river incision, tectonic uplift, and
erosion rates in various parts of the world.
FIGURE 2 In-place fluvial sediments (Bone Cave, Tennessee) pre¬
serve the record of a stream once occupying the abandoned conduit.
Cosmogenic nuclides are measured to determine how long ago the
cave stream was active and to constrain the timing of the incision event
that abandoned the conduit. Photograph taken by the author.
LANDSCAPE INTERPRETATION USING
MULTILEVEL CAVES
Water Table Positions at
Mammoth Cave, Kentucky
The development of cave levels in the Mammoth
Cave System of Kentucky (Fig. 3) is now firmly linked
to changes in the elevation of the Green River (a mod¬
ern tributary of the Ohio River), with passage mor¬
phology and sediment age determined by cosmogenic
nuclides permitting a reconstruction of regional base
level over the past 3 million years (myr) and a correla¬
tion of passage development and abandonment with
regional and global events. Several periods of river sta¬
bility punctuated by brief periods of accelerated river
incision allowed the formation and abandonment of
extensive horizontal passages at levels A and B. The
most recent river stillstand ended 1.39 million years
ago (Ma), with the abandonment of level C. Remnant
terraces in the Green River gorge corresponding with
cave levels may now be assigned ages, which correlate
ENCYCLOPEDIA OF CAVES
530
M
Elevation in meters above sea level; no horizontal scale
FIGURE 3 Generalized elevations of cave levels in Mammoth
Cave, Kentucky, and the age of sediments deposited in each level.
Cave levels correlate with remnant terraces and fluvial deposits
along the Green River valley walls. Passage morphology and the
age of sediments in Mammoth Cave constrain river stillstands and
incision events during the past 3 myr.
with major ice advances in the Ohio River basin dur¬
ing the Plio-Pleistocene. For example, the abandon¬
ment of Mammoth Cave's level C at ~1.4 Ma was
initiated by the blocking of major north-flowing rivers
by an ice sheet that forced reorganization of North
American drainage east of the Mississippi River, creat¬
ing the modern Ohio River at ~1.5 Ma.
Cosmogenic nuclides in cave sediments may also be
used to determine rates of surface processes while the
sediments were deposited underground. The dates of
abandonment at Mammoth Cave generate an incision
rate of ~30 m myr -1 for the Green River. Erosion rates
indicate that the sandstone-capped uplands surround¬
ing the Green River gorge have maintained a steady
rate of ~2—7 m myr despite the accelerated river inci¬
sion rate. Although surface valleys have deepened con¬
siderably, landscape features of the Appalachian
Plateaus would have been recognizable more than 3 Ma.
For speleologists, the ages of levels B and C abandon¬
ment demonstrate that level C passages at Mammoth
Cave needed at least 500,000 years to form, defining a
minimum number of years for extensive conduits to
develop in that particular limestone unit.
Incision Pulses on the Cumberland River,
Tennessee
Whereas Mammoth Cave preserved a record of
changes in the water table at one location, abandon¬
ment of multilevel caves along the valley walls of
tributaries to the Cumberland River in Tennessee
tracks the migration of incision pulses, or knickpoints,
along this modern tributary of the Ohio River.
Knickpoints are sudden changes in the gradient of a
river's long profile, such as rapids or waterfalls, and
may form as a response to a sudden drop in the
water table. A knickpoint will migrate in the upstream
FIGURE 4 Abandonment of multilevel caves along tributaries of
the Cumberland River, Tennessee, track the migration of a knick¬
point originating on the Cumberland River (arrow) at time fi prior to
2 Ma and moving up tributaries, reaching caves at times f 2 , t 3 , and f 4 .
Differences in stream power account for the simultaneous abandon¬
ment of caves on separate tributaries.
direction at a rate directly related to the size of the
river's total drainage area, with pulses moving faster
on a larger river than its tributaries. As a knickpoint
migrates up a larger river, knickpoints are initiated on
the smaller tributaries (much like dominoes). The ages
of sediments found in large, abandoned passages
across the Cumberland River basin (Fig. 4) at similar
elevations above the modern river level show that
caves were abandoned sequentially beginning around
1.8 Ma as a knickpoint generated by incision of the
Cumberland River moved up the Caney Fork and
beyond the cave outlets. This incision pulse is older
than the reorganization of the Ohio River basin, indi¬
cating that sea level lowstands at either 2.5 or 2.1 Ma
are likely candidates for triggering this earlier incision
event.
The simultaneous abandonment of passages on two
tributaries of different lengths and drainage areas
(the Caney Fork and the East Fork Obey rivers) illus¬
trates how the same incision event may initiate knick¬
points that migrate upstream of varying stream power,
affecting two different caves at the same time. Incision
rates for tributarie of the Cumberland River are
also ~30 m myr -1 , and the erosion rates for
the sandstone-capped Cumberland Plateau are a stea¬
dy 2—7 m myr -1 , corroborating those found at the
Green River.
Additionally, when knickpoint propagation up flu-
viokarst tributaries of the Cumberland River is mod¬
eled according to the stream power law E = kQ m S n and
tested using the dated incision events, the ratio m / n is
consistent with published parameters but the values
for m and n are much higher than those reported in
previous field studies.
ENCYCLOPEDIA OF CAVES
MULU CAVES, MALAYSIA
531
Incision Rates in the Sierra Nevada, California
Measurement of cosmogenic nuclides in cave sedi¬
ments and bedrock surfaces, along with studies of
landscape morphology, highlight the history of land¬
scape development in the southern Sierra Nevada,
California, over the past 5 myr. Cave sediments record
moderate to slow river incision rates between ~5 and
3 Ma (<70 m myr -1 ), acceleration between 3 and
1.5 Ma (on the order of 300 m myr -1 ), and a decrease
since (on the order of 20 m myr -1 ). The acceleration of
river incision is interpreted to be a response to a dis¬
crete tectonic event as opposed to global climate
change. Topographic evidence and dated cave posi¬
tions include the development of narrow inner gorges
beginning at ~3 Ma that incised into existing river can¬
yons. Tributary streams have not adjusted to the incis¬
ing of the inner gorges, indicating that the response to
Pliocene uplift has not yet migrated to the upland
surfaces. Exposure dating of bedrock shows these
surfaces to have eroded slowly over the past 3 myr
at rates on the order of 10 m myr -1 , whereas adjacent
rivers incised rapidly. Thus, climate change is not
enough to explain uplift, and a tectonic event is
favored.
Erosion Rate Increase in the Southern Rocky
Mountains, Colorado
The measurements of cosmogenic nuclides in cave
sediment, bedrock, and modern fluvial bedload record
a significant increase in Pliocene erosion rates in the
Rocky Mountains of southern Colorado. The burial
age of cave sediments in Marble Mountain coupled
with exposure dating of bedrock on the summit above
the cave indicate a sudden erosion rate increase in the
Rocky Mountains between 4.9 and 1.2 Ma. Based on
passage development and morphology of cave levels
in Marble Mountain, the increase likely occurred
shortly before 1.2 Ma, coinciding with the transition
from a warmer, more humid Pliocene climate to a
cooler periglacial Pleistocene climate.
Young Uplift of the Eastern Alps, Austria
The development of multilevel caves in the Central
Styrian Karst within the Mur River basin, Austria,
records a complex history of changes in the water
table. Were these changes due to incision or uplift?
Studies indicate that the Mur River acted as an ante¬
cedent river in an area of recent uplift of the Eastern
Alps, based on the topographic history of the basin,
stream piracy by the Mur River, and Plio-Pleistocene
incision rates in bedrock on the order of 100 m myr -1 .
Further correlation of cave levels, stream terraces, and
other erosional surfaces along the Mur River constrain
the timing of landscape evolution along the Eastern
Alps since 5 Ma.
The relationship between multilevel caves and the
surrounding landscape has opened the door to many
applications in geomorphology due to the ongoing
development of burial and exposure dating of sedi¬
ments and bedrock by measuring cosmogenic nuclides
in quartz. Geomorphologists around the globe are con¬
tinuing to examine multilevel caves in every type of
karst terrain, measure the rates of geomorphic and tec¬
tonic processes, and test long-established models using
these empirical data.
See Also the Following Article
Cosmogenic Isotope Dating of Cave Sediments
Bibliography
Anthony, D. M., & Granger, D. E. (2006). A new chronology for the
age of Appalachian erosional surfaces determined by cosmogenic
nuclides in cave sediments. Earth Surface Processes ami Landforms,
36, 874-887.
Anthony, D. M., & Granger, D. M. (2007). An empirical stream
power formulation for knickpoint retreat in Appalachian Plateau
fluviokarst. Journal of Hydrology, 343, 117—126.
Ford, D. C., & Williams, P. W. (2007). Karst hydrogeology and geomor¬
phology. New York: Wiley.
Granger, D. E., Fabel, D., & Palmer, A. N. (2001). Pliocene-
Pleistocene incision of the Green River, Kentucky, determined
from radioactive decay of cosmogenic 26 A1 and 10 Be in Mammoth
Cave sediments. Geological Society of America Bulletin, 113,
825-836.
Palmer, A. N. (1987). Cave levels and their interpretation. National
Speleological Society Bulletin, 49, 50—66.
Stock, G. M., Anderson, R. S., & Finkel, R. C. (2005). Rates of erosion
and topographic evolution of the Sierra Nevada, California,
inferred from cosmogenic 26 A1 and 10 Be concentrations. Earth
Surface Processes and Landforms, 30, 985—1006.
Wagner, T., Fabel, D., Miebig, M., Hauselmann, P., Sahy, D., Xu,
Sheng, & Stuwe, K. (2010). Young uplift in the non-glaciated parts
of the Eastern Alps. Earth and Planetary Science Letters, 295,
159-169.
White, W. B. (1988). Geomorphology and hydrology of karst terrains.
New York: Oxford University Press.
MULU CAVES, MALAYSIA
Tony Waltham and Joel Despain
Mulu Caves Project and British Cave Research Association;
Gunung Buda Project and Sequoia, Kings Canyon
National Park Service
One of the world's great tropical karst regions lies
below Gunung Mulu (Mount Mulu) in the Malaysian
state of Sarawak on the island of Borneo, off Southeast
Asia. Here, the Miocene Melinau Limestone is exposed
in a range of precipitous massifs, low mountains, and
Encyclopedia of Caves.
©2012 Joel Despain and Elsevier Inc. All rights reserved.
532
M
hills, in a northeast-oriented outcrop, 10 km wide and
35 km long, broken by the floodplains of the Medalam,
Melinau, and Melinau Paku rivers. Within the lime¬
stone lie cave passages of which more than 440 km
have been surveyed to date (Fig. 1). These include the
largest cave room in the world (Sarawak Chamber),
the largest cave passage in the world (Deer Cave), and
perhaps the largest cave in the world (when measured
by total volume of Clearwater Cave, with nearly 190 km
of passages that are mostly very large).
The terrain around Gunung Mulu is spectacular and
dramatic, due to the precipitous rise of the mountains
above the coastal plain, the white limestone cliffs more
than 100 m high around their lower margins, the pro¬
lific dense vegetation, and an almost persistent cloud
cover. The three main limestone mountains, Gunungs
Api, Benarat, and Buda, lie 75 km from the South
China Sea at the edge of the interior highlands of
Sarawak. They lie 4° north of the equator, and receive
an annual average of more than 5000 mm of year-
round rainfall.
The Melinau Limestone is a lenticular body, and
the rock is very massive and white or gray with gener¬
ally less than 1% insoluble material. It is lagoonal in
origin, and was deposited within a reef complex.
Fossils are locally common and include corals, bryozo-
ans, bivalves, gastropods, and algal balls. The massive
beds of the Melinau Limestone dip to the northwest,
perpendicular to the trend of the range, and are folded
with minor faulting. The limestone is overlain con¬
formably by the Setap Shale, which is 4000—5000 m
thick, and is underlain by the Mulu Formation, with
4000—5000 m of shale, sandstone, and orthoquartzite.
Local forest nomads were the first explorers of some
of the caves. The Berawan, Tabuan, and Penan peoples
still make use of the caves, some of which are burial
sites. The caves also house several species of swiftlets,
whose nests are used to make bird's nest soup, a
highly prized Asian delicacy. Harvesting of nests at
Mulu, and elsewhere in Southeast Asia, has not been
on a sustainable scale, leading to a steep decline in the
numbers of these birds, besides significant damage to
some of the caves.
The first significant cave explorations in the rugged
and remote forests of Mulu were made by G. E.
Wilford of the Malaysian Geological Survey in 1961.
He mapped Deer Cave (but both he and the readers
of his report underestimated its size), the first part
of Cave of the Winds (later connected to Clearwater
Cave), and three inlet passages of the Terikan River
Caves (but without venturing into the main under¬
ground river). His was a pioneering work, but it was
only in 1978 that the Mulu caves were again visited,
by British cavers on the Royal Geographic Society
expedition sponsored by the Malaysian government.
This project sought to survey the flora and fauna of
the region and to study the geology and hydrology of
the area in the newly designated Gunung Mulu
National Park.
In 1978, the caves were almost perceived as an expedi¬
tion sideline, but they proved to be truly spectacular.
Consequently, the British cavers followed up with major
expeditions in 1980 and 1984, and then with visits that
have become annual since the National Park became
more accessible; the Mulu Caves Project continues to
make new discoveries. Parallel explorations by American
cavers revealed many more caves in and around Gunung
Buda during four expeditions between 1994 and 2003.
CAVES IN THE SOUTHERN HILLS
Separated from Gunung Api by the Melinau Paku
valley, the southern hills contain Deer Cave (Gua
Payau in Malay). This could be dismissed as a frag¬
ment of an ancient trunk cave that has been left in a
limestone hill, now isolated from the Api massif. But it
is the largest cave passage in the world, reaching
168 m wide and 125 m high along its southern half
(Fig. 2). The very large passage size makes its geomor¬
phology difficult to appreciate; phreatic origins have
been lost in progressive roof collapse, and it appears
that much of the passage was cut as a giant vadose
canyon by the Melinau Paku River, which took this
underground route for a very long time. The giant pas¬
sage is almost entirely within reach of daylight for its
kilometer course between the stream entrance in the
Garden of Eden doline and the dry exit at the southern
end. North of the Garden of Eden, Green Cave is its
continuation only on a slightly smaller scale. A net¬
work of smaller cave passages lies within the hills par¬
allel to Deer Cave.
The Garden of Eden is an almost enclosed chunk of
rainforest within a massive doline on the edge of the
limestone outcrop. It may be described as a very large
but well-degraded tiankeng, with vertical cliffs still
forming its margins on three sides. The concept of
development by collapse of a very large cave chamber
is reasonable when it is compared to the nearby
Sarawak Chamber, which is not much smaller, though
any large chamber may always have been open through
an inlet from the shale outcrop on its southeast side.
CAVES IN GUNUNG API
Mt. Api is riddled with caves, and nearly 190 km of
the passages are now linked to form the Clearwater
Cave System (or Gua Air Jernih in its Malay translation),
named after the clear karst water which pours from its
ENCYCLOPEDIA OF CAVES
MULU CAVES, MALAYSIA
533
Bukit Assam Cave
Green ////"'-
Cathedral 'V
P Cave^uf /
Sarawak
Snail Shell Cave
Tardis Cave
Terikan River
V Caves
Deliverance
Cave
CCobweb
Cave <
>Benarat
Caverns
White rock
Cave
Blackrock i
Cave M
' cave passages
major surface river
contours (50m interval)
'Api Chamber
Hidden Valley
Cobra
CaveC
Clearwater .
I Cave s M
ASTER based terrain model with 50m contours.
ASTER GDEM is a product of METI and NASA.
http://asterweb.jpl.nasa.gov/gdem.asp
Sarawak Chamber
Cave of the Winds.
if Green Cave
Garden of Eden
-AGunung Buda
National Park
4, limestone hills
‘(Gunung Mulu
.-‘National Park
FIGURE 1 Outline map of the known caves within the limestone mountains adjacent to Gunung Mulu.
ENCYCLOPEDIA OF CAVES
534
M
FIGURE 2 The very large passage in Deer Cave looking out
toward the main entrance at its southern end, which is nearly 600 m
from the camera. Photograph taken by Tony Waltham.
resurgence into the muddy Melinau River. The main
Clearwater River Cave, upstream from its resurgence, is
a magnificent canyon passage, 30 m high and wide,
whose clean walls, cut and undercut in pale limestone,
are the epitome of underground sculpture. This one
cave river drains most of Api, and has been followed
beyond six short sumps. The Firecracker Streamway in
Blackrock Cave and the Whiterock River are two further
long sections of large and gently graded river passages,
all connected by high levels into the sprawling
Clearwater system; Whiterock reaches to within a few
meters of the Melinau Gorge, where it collects water
from sinking distributaries.
The river passages are only the latest phase of the
cave development, as a maze of huge abandoned cave
tunnels forms the high levels of Clearwater. Underfit
streams drain down through some of these, but the
main passages are otherwise inactive, and many are
truncated at exits hidden in the impenetrable forest. At
the southern end. Revival is the main old trunk route
above the Clearwater River; it is mostly 30 m high and
wide and lies about 100 m above the modern stream¬
way. Farther north, the Big Issue and Borderline are
the main trunk routes, 200 m above the Blackrock and
Whiterock streamways, but there are multiple high
levels through most of the system. The enormous Api
Chamber (300 m long and 200 m wide) formed where
parallel trunk passages coalesced by collapse, and an
even higher level at the resurgence end includes an
opening into the forested floor of the Secret Garden
tiankeng. All the main cave passages are aligned on
the strike of the dipping limestone, stacked obliquely
up just a few bedding planes, and covering over 400 m
of vertical range. The high levels also continue beyond
the resurgence to link to Cave of the Winds, a stream¬
way carrying water through from the Melinau Paku.
Lagan's Cave has another segment of large old trunk
passage at an intermediate level, but there is no con¬
nection to Clearwater as yet.
Parallel to and southeast of Clearwater, two sepa¬
rate cave systems drain from the huge blind gorge of
Hidden Valley through to the Melinau Paku lowland.
The Cobra-Bridge-Cloud Cave provides a convoluted
through route. Gua Nasib Bagus can only be entered at
its resurgence, and its high-level route from Prediction
Cave is choked with sediment. Midway through the
mountain, Sarawak Chamber is the world's largest
underground void. It is 700 m long by 300—400 m
wide, and its roof arches 100 m over the ramp of fallen
blocks that forms its sloping floor. The chamber was
formed where the river from Hidden Valley slipped
sideways, down the steep bedding, and created the
giant undercut; this has subsequently been modified as
the roof has collapsed upward to form a stable arch.
The river has since found a new entrance passage far¬
ther downdip, and now just drains through the
boulders at the foot of the chamber, leaving the main
part of the chamber and Prediction Cave dry.
Sarawak Chamber almost defies comprehension.
The three cavers who explored the cave in 1980 did not
know that they were in a chamber; they could only see
one wall, which they followed to make their survey.
Only when they had gone most of the way around a
huge loop did they realize that they might be in a vast
cavern, instead of in a very large, curving passage; they
calculated a compass bearing and then climbed over
the boulders across the middle of the chamber to regain
their entry point. Midway, they could see no walls, no
roof, nothing—the sense of agoraphobia was acute.
The main trunk passages in the Clearwater Cave
System have formed where large flows of water have
poured off the sandstone slopes of Gunung Mulu and
sunk into the limestone in the Melinau Gorge; these
have then flowed underground, to emerge farther south
where the limestone mountains drain onto the adjacent
plains at lower altitudes. The Gorge has long been at
the upper apex of a huge gravel fan that has spread out
onto the lowlands, and this fan has continually evolved
and reformed at lower levels as both Gunung Mulu
and the plains were lowered by surface denudation.
The successive falls of the fan, the plains, and the water
table within Gunung Api can be recognized in the fall¬
ing sequence of trunk cave passages, almost along the
strike of the dipping beds. Clearer indicators of the past
water levels are the deep wall notches which were
undercut by the cave rivers whenever they flowed over
stable gravel fills (Fig. 3). Abandoned notches are
among the most distinctive features of Mulu caves; gen¬
erally 2—3 m high and cut 4—5 m into the vertical walls,
they provide cavers with terraced walkways around
deep lakes or large boulder piles in both the old high
levels and the modern river passage.
ENCYCLOPEDIA OF CAVES
MULU CAVES, MALAYSIA
535
The graded trunk passages, wall notches, and gravel
profiles within Clearwater have been correlated with
the gravel terraces of the Melinau Plain. The clastic
sediments of both the cave and the plain are essentially
FIGURE 3 Clearwater River Passage, with a conspicuous wall
notch at 2—4 m above the present water level. Photo courtesy of Jerry
Wooldridge.
the lower ends of enormous gravel fans supplied from
the slopes of Gunung Mulu and poured through the
breaches in the limestone escarpment. Each terrace
was formed by aggradation during wetter interglacial
periods, followed by incision during drier glacial peri¬
ods. More than 25 km of passages with traceable wall
notches have been mapped within Clearwater Cave,
and these fall into a sequence of 20 levels, spread over
more than 300 m of altitude. Uranium series dating of
calcite on sediments within the lower, more recent,
passages indicates a mean rate of base-level decline of
0.19 m per 1000 years. The sequence correlates well
with all the interglacial peaks on the oxygen isotope
record, and implies that the gravel aggradation was
controlled by climatic variations and not by episodic
uplift. Preliminary data from Al/Be dating of clastic
sediments in some upper levels indicate ages of well
over 2 myr, which fit with extrapolation of the erosion
rates inferred from the younger stalagmite ages. The
constant background rate of base-level lowering is
interpreted as the rate of isostatic uplift in response to
regional denudation. The limestone blocks have risen
almost uneroded on this uplift, while the adjacent
plains, and the noncarbonate slopes, have been
denuded at rates close to that of the uplift. The lime¬
stone mountains of Mulu have taken around 10 myr to
evolve to their present topographic relief.
CAVES IN GUNUNG BENARAT
Massive cave systems within Gunung Benarat largely
follow the style of those in Gunung Api. From the high
cliffs of the Melinau Gorge (Fig. 4), large abandoned
trunk passages head northeast along the strike of the
steeply dipping beds; four main levels lie at altitudes
FIGURE 4 The Melinau Gorge, looking upstream toward the cloud-covered slopes of Gunung Mulu; the cliffs along the left side are bro¬
ken by the dark cave entrances 400 m above the river that are the truncated ends of old trunk passages in Benarat Caverns. Photograph taken by
Tony Waltham.
ENCYCLOPEDIA OF CAVES
536
M
roughly 100 m apart. One outlet of these old trunk
routes is through Deliverance Cave where 4 km of pas¬
sage open into the high slopes of the Medalam Valley,
but there is a long section choked by sediments and
unexplored through the center of Benarat as yet.
From the high-level entrances in the Melinau Gorge,
the Benarat Caverns System has just over 50 km of
mapped caves that also link, across the beds, to the
complex maze of passages in Cobweb Cave. Many of
the Cobweb passages are less mature as they still had
switchback profiles, up and down the dipping beds,
when they were abandoned, though Swift Highway
is one high-level trunk with a well-graded profile.
Within Cobweb only short sections of low-level
streamway have been reached, and these head north
toward the Terikan resurgence.
Across the north end of Gunung Benarat, the
Terikan River Caves contain another 32 km of pas¬
sages, with an active cave river largely fed from vari¬
ous sinks along the floodplain margin, and draining
out to a resurgence at the head of the Terikan River, a
major tributary of the Medalam. Above the river pas¬
sage, multiple high levels almost parallel the modern
river, and also include a complex of caves that reach
very close to the northern ends of the Cobweb pas¬
sages in Benarat Caverns.
It appears that the large phreatic trunk passages
through Benarat have, at various times in the past, car¬
ried water in both directions, toward the Melinau and
toward the Medalam, but the geomorphology of the
Benarat caves is still not understood as well as that of
Api. Many of these caves did not mature to graded
profiles, and consequently wall notches are not as
widespread as they are in Api. As yet, there have been
no programs of detailed leveling and sediment dating
to confirm exactly how the Benarat caves correlate
with those in Api. The Terikan caves are altogether dif¬
ferent, as they were formed at lower levels by water
from a separate alluvial fan built by the River
Medalam.
CAVES IN GUNUNG BUDA
Gunung Buda is as rich in caves as the limestone
mountains to its south, except that it lacks their very
large ancient trunk passages. The most extensive sys¬
tem is Green Cathedral Cave, with 27 km of mapped
passages forming a very complex series of old phreatic
drains on multiple levels. Some passages switch back
up and down to follow steeply dipping bedding
planes and available joints, while others have achieved
more level profiles that were graded to the contempo¬
rary Medalam River. Currently, streams flow from
sandstone sinks to the east into the limestone and then
drain broadly along the strike toward outlets near the
Medalam. To the north, Gua Bukit Assam has 8 km of
passages with several small streams and a complex
of high levels. Cave hydrology is complex in this part
of Buda, with streams forming on adjacent sandstone
and within small limestone basins, before draining
into the caves. In contrast, passages along the southern
edge of Buda (Fig. 5) were probably formed by distrib¬
utaries of the Medalam; a stream passage in Snail Shell
Cave carries water from the Medalam away to the
north, possibly to Lower Turtle Cave. These caves
have even more local relief where they are aligned
directly across the limestone strike; the large phreatic
ramp in Snail Shell Cave has been followed over a
vertical range of 470 m. Buda's high-level caves are
currently being truncated by surface erosion, creating
the many entrances in the forest.
The lower levels of the Buda caves contain signifi¬
cant lengths of active streamways. Most of these are
flood-prone due to sandstone sinkhole catchments and
FIGURE 5 Giant pinnacles etched into the passage floor by eons
of percolation water in Tardis Cave, in the southern tip of Gunung
Buda. Photo courtesy of Dave Bunnell.
ENCYCLOPEDIA OF CAVES
MULU CAVES, MALAYSIA
537
also pulses of rainwater that funnel straight down the
fissures from the forest above. In contrast. Lower
Turtle Cave contains long lakes that vary little in level
or flow within an old phreatic tube.
Buda contains a total of nearly 70 km of known
caves, and various outlying limestone blocks contain
another 11 km of mapped passage, mostly active
stream passages at low level; these include the 5 km of
passages in Spirit River Cave away to the northwest of
the main limestone ridge. Buda continues the trend set
by Api and Benarat that every chunk of limestone in
Sarawak's rainforest is riddled with, large, beautiful
and spectacular caves that are exceptionally extensive.
GUNUNG MULU NATIONAL PARK
The caves are not the only outstanding feature of
Mulu. Gunung Api is one of the world's most spectac¬
ular chunks of limestone karst. Except for the vertical
cliffs of its margins, its entire surface is fretted into
dramatic pinnacle karst. Widely spaced vertical joints
have been etched out by dissolution, and are now
almost choked by vegetation and organic soils. The
remnant blocks of limestone have been carved by
direct rainfall into razor-sharp ridges. Most of these
pinnacles are just a few meters high, and the terrain is
seriously difficult to cross because it consists entirely
of steep pinnacles and tapering fissures choked with
unstable plant debris. High on Api, pinnacles that are
more than 30 m tall rise above the forest canopy to cre¬
ate an amazing landscape, and most of the ground has
never been trodden by humans.
The British Royal Geographic Society and other
researchers before and since have documented an amaz¬
ing diversity of plants and animals in the area. The large
range in elevations, various soil and rock substrates, and
tropical climate have produced what may be the world's
most diverse assemblage of tree species, an amazing
variety of herbaceous plants, 109 species of palms, and
a total of 3500 vascular plants in 17 vegetation zones.
More than 270 bird species including all Bornean horn-
bills, broadbills, and barbets are known from Mulu, and
primates found in the area include monkeys, prosim¬
ians, and the Bornean gibbon. Other important wildlife
includes dozens of snakes, civets, many bats, and more
than 200 species of cave-adapted invertebrates.
Cave biology work was undertaken by both British
and American expeditions. Four dominant terrestrial
habitats (and bio-communities) are recognized within
the Mulu caves: bat guano, bird guano, the entrance
transition zone, and the deep cave community. Each
has unique species of crickets, spiders, beetles, cock¬
roaches, and, in some communities, isopods and milli¬
pedes. Some species are found to occur throughout the
caves, whereas others are restricted to specific habitats.
Swiftlets fly far into many of the larger caves. More
than two million bats roost on the roof of Deer Cave,
far above any predators. They feed outside the cave,
but droppings from their roost accumulate to form
great banks of guano on the cave floor. These are both
home and food supply for millions of beetles, earwigs,
cockroaches, millipedes, and crickets, together with
large, predatory spiders and centipedes and an active
population of racer snakes. Members of the Mulu cave
community include snakes found deep within the
caves that predate primarily on birds but also on bats,
several species of crabs including some that lack pig¬
ment, and at least two species of large centipedes.
The Gunung Mulu National Park was established
primarily to conserve forever a substantial section of
the Borneo rainforest—which was ever diminishing
and under continued threat from Sarawak's powerful
logging industry. To save the forest, the park needed
economic viability through tourism, and the caves
were seen as potential visitor attractions. The fabulous
cave discoveries made the initial plan work out very
well, and the Mulu Park now offers the dual attraction
of the forest environment and the great caves.
The Gunung Mulu National Park was formally
established by the Sarawak state government in 1974,
with 52,864 ha of forest protected, along with its spec¬
tacular caves and its wonderful array of plants and
animals. It was inscribed as a World Heritage Site by
the United Nations in 2000. The Park includes the
sandstone mountain of Gunung Mulu rising to 2377 m,
the rugged limestone mountains along its western
flank, and the adjacent terraced floodplains that reach
to as low as 28 m above sea level. It was opened in
1985, and now boasts a visitor center, forest board¬
walks, tourist cave facilities, a large resort hotel, and a
small adjacent airport, while the high mountains are
preserved as almost inaccessible wilderness. Gunung
Buda was established as a separate but adjacent
National Park in 2001. It protects another 6235 ha of
limestone uplands, swampy floodplains, forested hills,
and rivers, but is kept as a wilderness with neither
facilities nor easy access.
Bibliography
Brook, D. B., & Waltham, A. C. (Eds.), (1978). Caves of Mulu. London:
The Royal Geographic Society.
Eavis, A. J. (Ed.), (1981). Caves of Mulu '80. London: The Royal
Geographic Society.
Eavis, A. J. (Ed.), (1985). Caves of Mulu '84. London: The Royal
Geographic Society.
Farrant, A. R., Smart, P. L., Whittaker, F. F., & Tarling, D. H. (1995).
Long-term Quaternary uplift rates inferred from limestone caves
in Sarawak, Malaysia. Geology, 23, 357—360.
ENCYCLOPEDIA OF CAVES
538
M
Hacker, B. (Ed.), (1997). Caves of Gunung Buda. Huntsville, AL:
National Speleological Society.
Hacker, B. (Ed.), (2000). Caves of Gunung Buda 1997. Huntsville, AL:
National Speleological Society.
Kirby, M. (Ed.), (2009). Mulu caves 2009. Lancaster: Mulu Caves
Project.
Meredith, M., Wooldridge, J., & Lyon, B. (1992). Giant caves of Borneo.
Kuala Lumpur: Tropical Press Sdn. Bhd.
Proffitt, M., & Mosenfelder, J. L. (Eds.), (2003). Caves of Gunung Buda
2000. Huntsville, AL: National Speleological Society.
MYRIAPODS
David C. Culver and William A. Shear '
American University, ^Hampden-Sydney College
INTRODUCTION
Myriapods are wingless terrestrial arthropods with at
least nine pairs of walking legs and a body not distinctly
divided into a thorax and abdomen (Minelli and
Golovatch, 2001). Based largely on molecular evidence,
myriapods are now regarded as a monophyletic group
related to crustaceans and insects. The most familiar
myriapods are centipedes and millipedes. Of the four
classes of myriapods, three have definitively troglobiotic
species. Only the Pauropoda seem to have no troglo-
bionts (Harvey et al., 2000). A variety of cave myriapods
is shown in Figure 1. Millipedes are often the most abun¬
dant component of terrestrial cave communities at least
in terms of biomass. Generally speaking, the myriapods
have been less thoroughly studied than other terrestrial
arthropod groups (insects and arachnids), with the
exception of the mites.
PAUROPODA
Pauropods are the smallest myriapods, rarely more
than 1 mm long, eyeless, generally lacking in pigment.
Nearly all species have 9 pairs of legs, but a few have
10 or 11. Their bodies are relatively short and stocky,
and they have unique 3-branched antennae. Only two
species have been observed feeding; both of them bit
into fungal hyphae and sucked out the contents. They
generally occur in very humid, cool spots, in leaf litter,
and soil fissures. Given their general habits, habitats,
and morphology, it would seem likely that they should
occur in caves. The absence of troglobiontic species may
be attributed to the general troglomorphic appearance of
all members of the group, especially eyelessness and
lack of pigment, which would make it difficult to set tro-
globionts aside from the others. Large numbers of paur¬
opods have been reported from a few caves, such as
Holloch in Switzerland and several caves in southern
France. Pauropods have also been found in mines. It is
somewhat surprising that they have not been found to
be more widespread and common in caves. This may be
due to the fact that they have often been overlooked due
to their small size, or that they have food or habitat
requirements that are rarely met in caves.
FIGURE 1 Illustrations of troglobiotic myriapods: (A) centipede—
Lithobiomorpha, Lithobius anophthalmus; (B) millipede—Chordeumatida,
undescribed species from China; (C) millipede—Glomerida, Trachysphaera
orientalis; (D) millipede—Julida, Antrocoreeana arcuata; (E) millipede—
Polydesmida, Epanerchodus fontium.
SYMPHYLA
Symphylans are all less than 1 cm long, and have the
appearance of small, white centipedes. All species are
eyeless, generally lack pigment, and have one pair
of elongate multisegmented antennae, 12 pairs of legs,
and a pair of posterior unsegmented appendages called
spinnerets. The number of tergites, or dorsal plates, cover¬
ing the body is variable and often not correlated with the
number of legs. Most symphylans feed on dead and
decaying vegetation, but a few species are pests of crops.
They are widespread throughout the world in litter and
soil. Like pauropods, symphylans are often collected in
caves and some species are known only from caves.
However, because relatively little surface collecting has
been done and the species so far described from caves
show few modifications for cave life, it is not known if
any of them are really troglobionts. Scolpendrellopsis
pretneri Juberthie—Jupeau and Scutigerella hauseri Scheller
are candidate troglobionts from the Postojna-Planina
cave system in Slovenia, the most diverse cave commu¬
nity in general.
Encyclopedia of Caves.
© 2012 Elsevier Inc. All rights reserved.
MYRIAPODS
539
CHILOPODA
Centipedes exhibit a great range of body lengths
from a few millimeters to 30 cm. All have 15 or more
pairs of legs. All centipedes are predatory, mostly on
insects. The first trunk segment bears a pair of spe¬
cialized claws that contain a poison gland; the claws
are used in the capture of prey. Some centipedes also
have the last pair of legs modified as pincers to aid in
prey handling. One curiosity of the centipedes is that
the number of leg-bearing segments is always odd,
ranging from 15 to 191. Centipedes are typically forest
litter-dwelling predators although they have also colo¬
nized desert habitats. There are five orders of centi¬
pedes with about 3300 described species and several
times that number of undescribed species. Worldwide,
there are slightly more than 50 described troglobiotic
species (Negrea and Minelli, 1994).
The smallest order is the Craterstigmatomorpha
with 15 pairs of legs and a posterior end shaped like a
clam shell. There are only two species, known from
the soil in Tasmania and New Zealand. Neither has
been found in caves.
Scutigeromorpha is a primarily tropical order charac¬
terized by 15 pairs of extremely long legs and large com¬
pound eyes, the only centipede order with compound
eyes. They are found in caves in Asia, Africa, Australia,
and South America. Scutiem coleoptrata is a European
and North African species that has been spread around
the world by man, and is often found in caves in desert
regions, though it more commonly occurs in houses and
other buildings. Some of the Asian troglophiles are
extraordinarily large with leg spans up to 15 cm. None
of the species is troglobiontic.
Scolopendromorpha is a widespread order charac¬
terized by 21 or 23 pairs of legs. The largest known
centipedes, up to 30 cm in length, belong to this group.
Members of this order are either eyeless or with a few
simple ocelli. Individuals of some of the very large
surface-dwelling species are often found in tropical
caves. One species has been filmed snatching bats out
of the air as they fly by. Two troglobiotic species
('Thalkethops grallatrix Crabill from New Mexico and
Neivportia leptotarsis Negrea, Matic, and Fundora from
Cuba) show morphological modifications for cave
life—depigmentation and elongation of appendages.
A handful of other species from the Mediterranean
and Cuba are known only from caves but show no
morphological modifications for cave life. Species such
as these may also occur in non-cave habitats but have
not yet been found there, or they may recently have
been isolated in caves.
Geophilomorpha is the largest centipede order in
number of species, common on all continents except
Antarctica and in nearly all climates. They are long.
thin centipedes with between 35 and 191 pairs of legs
and without eyes. The antennae are relatively short
and always consist of 14 articles. Some species are
capable of producing cyanide as a chemical defense,
and still others give off a sticky, luminous secretion.
Nearly all species of geophilomorphs are found in
moist habitats in leaf litter and in soil, sometimes a
meter or more beneath the surface. In spite of the simi¬
larities of this habitat to caves, geophilomorphs are
quite uncommon in caves. This is probably because
geophilomorphs are adapted to small rather than large
cavities, are strongly thigmotactic, and have rarely
made the transition to large cavities (caves). This
explanation is, of course, incomplete because many
other groups have made this transition. Although sev¬
eral troglophilic species have been reported in Cuba,
there are no known troglobionts.
It is with the Lithobiomorpha that one finds the
greatest number of chilopod troglobionts. They have
15 pairs of stout legs, relatively short antennae, and at
most a few ocelli. Of the 850 or so described species,
nearly 50 are troglobionts. Some troglobiotic species,
such as Lithobius matulicii, from caves in the Popovo
Polje of Bosnia and Herzegovina, have more than
100 segments in their antennae (Fig. 2)! Caves in the
Pyrenees Mountains of France and Spain are a hotspot
of lithobiomorph diversity with more than 20 species.
Europe as a whole has approximately 40 species. By
contrast, in the United States, only two species of
Typhlobius, one from Virginia and one from California,
and Nampabius turbator Crabill have been described
from caves. However, several additional undescribed
species from Texas, Colorado, Idaho, and California are
known. Curiously, the undescribed Rocky Mountain
species belong to a family that is much more dominant
in the temperate parts of the Southern Hemisphere,
but rare in the Northern Hemisphere. It is likely that the
FIGURE 2 Illustration of the cave centipede Lithobius matulicii from
Vjetrenica Jama in Bosnia and Herzegovina. Drawing by Slavko Polak.
ENCYCLOPEDIA OF CAVES
540
M
disparity in described species is due to the historical
lack of centipede specialists in North America. A scatter¬
ing of troglobionts have been described from elsewhere,
including the Near East, Japan, and Korea.
DIPLOPODA
These millipedes range in size from a few millimeters
to 35 cm, and are by far the most common of the myria¬
pod orders. The first trunk segment is legless, followed
by three or four segments with one pair of appendages.
The remaining segments each have two pairs of legs.
Millipedes are the "leggiest" animals, and one species
from California may have more than 700. They are
generally elongate, and the body may or may not be flat¬
tened. The exoskeleton is hardened in all but one small
group, due to the presence of calcium salts. This require¬
ment for calcium predisposes them to abundance in
karst regions. They typically have groups of simple
eyes, but all millipedes in the abundant order
Polydesmida lack eyes. The mouth parts of many milli¬
pedes are adapted for cutting and chewing tough mate¬
rial, such as wood or dead leaves. This may explain in
part the remarkable numerical dominance of millipedes
in many terrestrial cave communities both in the tem¬
perate zone and in the tropics. They are often the domi¬
nant detritivores in a cave. It is not unusual to see
dozens of millipedes, both troglophilic and troglobiotic,
in caves. A single bait with cheese or rotten meat will
often attract 100 or more millipedes. Most millipedes
have potent chemical defenses that protect them against
predators, but these defenses are missing in the order
Chordeumatida, the dominant cave millipedes of the
Northern Hemisphere (Shear, 1999).
The taxonomic situation with millipedes has been
likened to that occurring in entomology at the middle
of the nineteenth century (Hoffman, 1999). About
11,000 species have been described and probably
70,000 species await description. At least several hun¬
dred troglobionts have been described. Centers of cave
millipede richness are in the Dinaric region of Bosnia
and Herzegovina, Croatia, and Slovenia, as well as in
the Pyrenees of France and Spain. In North America,
particular centers of cave millipede richness are in
northeast Alabama and adjacent Tennessee, southern
Indiana, western Kentucky, and the region along the
border between Virginia and West Virginia, extending
into Tennessee (Fig. 3). Many species have been
described from caves in Mexico and Central America,
but in other tropical regions the cave millipede fauna
is much less well known. Currently, 16 orders of
millipedes are recognized. The Glomerida, or pill milli¬
pedes, can roll up into an almost perfect sphere.
They are common troglobionts in Europe, Mexico, and
Central America (Maruies, 1994).
FIGURE 3 Map of troglobiotic millipede species richness in caves
in the eastern United States. Each hexagon has an area of 1000 km 2 .
TABLE 1 Number of Species of Troglobiotic Millipedes in
U.S. Caves. Modified from Shear (1969, 2008, 2010)
Order
Number
of Families
Number
of Genera
Number
of Species
Callipodida
2
2
3
Chordeumatida
5
15
63
Julida
1
2
2
Polydesmida
1
5
16
Spiros treptida
1
1
1
Polydesmida is the largest millipede order in num¬
bers of species. They are usually easy to recognize by
the wing-like lateral extensions on the dorsal side of
the animal, though these are absent or inconspicuous
in some species. All species in the order are eyeless.
Many troglobiont species have been described from
Europe, Japan, and Mexico, and they are the dominant
invertebrate detritivores in many Mexican caves. The
polydesmid Oxidus gracilis is a Southeast Asian species
that has been spread by man to warm regions all over
the world, and may occur in huge numbers in caves.
The impact of this species on native troglobionts has
not been studied. Relatively fewer troglobiont polydes-
mids have been described from U.S. caves, mostly in
the southwest (Table 1); these species tend to be small
and inconspicuous.
The order Julida is restricted to the Northern
Hemisphere. Julids are long-bodied, cylindrical, and
usually black or gray in color. They are the dominant
millipede order in Europe and there are at least two
troglobiont species in North America as well. Some
troglobiont European julids in the genus Trogloiulus
have mouth parts modified to comb bacteria from
surfaces and to filter food particles from water. As a
result, they are amphibious, and often collected in cave
pools and streams.
ENCYCLOPEDIA OF CAVES
MYRIAPODS
541
Diplopoda Pseudotremia
Species Name
▲ Pseudotremia cavemarum Cope, 1869
■ Pseudotremia fulgida Loomis, 1943
■{> Pseudotremia lusciosa (Loomis, 1939)
O Pseudotremia nodosa Loomis, 1939
# Pseudotremia tuberculata Loomis, 1939
4 Pseudotremia valga Loomis, 1943
FIGURE 4 Map of troglobiotic Pseudotremia species in Virginia
and West Virginia caves. Three additional species (see Shear, 2008),
not shown, are each known from a single cave.
The most important millipede order in north tem¬
perate caves is the Chordeumatida. Most surface¬
dwelling species are adapted to high humidity and
cool temperatures and are found in leaf litter and in
the soil. Given similarities of these habitats to the cave
environment, it is not at all surprising that there are
so many species found in caves. In the tropics, few
chordeumatids are found in caves, but those that do
occur are often members of families and genera that
occur on the surface farther north. Warming postgla¬
cial climates may have extinguished surface popula¬
tions, leaving relicts behind in the cooler conditions
of the caves. In the United States alone, there may be
more than 100 species, about a third of them unde¬
scribed (Table 1). Two genera that include the majority
of species in eastern North America are Pseudotremia
and Scoterpes; recent studies have greatly increased the
numbers of recognized species in both genera. Most of
the ranges of these troglobionts are very small (Fig. 4),
and some supposedly widespread described species
may actually be a complex of genetically distinct but
morphologically very similar species. Examination of
the DNA of populations that are indistinguishable on
the basis of anatomy have shown that even caves quite
close to each other may contain genetically isolated
populations that could deserve species rank.
Bibliography
Harvey, M. S., Shear, W. A., & Hoch, H. (2000). Onychophora,
Arachnida, Myriapods and Insecta. In H. Wilkens, D. C. Culver &
W. F. Humphreys (Eds.), Subterranean ecosystems (pp. 79—94).
Amsterdam: Elsevier.
Hoffman, R. L. (1999). Checklist of the millipeds of North and middle
America. Virginia Museum of Natural History Special Publication, 8,
1-584.
Mauries, J.-P. (1994). Diplopoda. In C. Juberthie & V. Decu (Eds.),
Encyclopedia biospeologica I (pp. 255—262). Moulis, France: Societe
de Biospeologie.
Minelli, A., & Golovatch, S. I. (2001). Myriapods. In S. A. Levin (Ed.),
Encyclopedia of biodiversity, Vol. 4 (pp. 291—303). San Diego:
Academic Press.
Negrea, S., & Minelli, A. (1994). Chilopoda. In C. Juberthie & V. Decu
(Eds.), Encyclopedia biospeologica I (pp. 249—254). Moulis, France:
Societe de Biospeologie.
Shear, W. A. (1969). A synopsis of the cave millipedes of the United
States, with an illustrated key to genera. Psyche, 76, 126—143.
Shear, W. A. (1999). Millipede American Scientist, 87, 232—240.
Shear, W. A. (2008). Cave millipedes of the United States. VII. New
species and records of the genus Pseudotremia Cope. I. Species from
West Virginia, USA (Diplopoda, Chordeumatida, Cleidogonidae).
Zootaxa, 1764, 53—65.
Shear, W. A. (2010). The milliped family Trichopetalidae, Part 2:
The genera Trichopetalum, Zygonopus and Scoterpes (Diplopoda,
Choredeumatida, Cleidogonoidea). Zootaxa, 2385, 1—62.
ENCYCLOPEDIA OF CAVES
NATURAL SELECTION
Peter Trontelj
University of Ljubljana, Slovenia
Natural selection is a major evolutionary force, the
only one that consistently leads to adaptation. The pro¬
cess of natural selection is central to Darwinian adap¬
tive evolution that in turn is widely believed to have
shaped much of the diversity of life on Earth. A suc¬
cinct definition is that of differential survival and
reproduction. Those individuals that manage to sur¬
vive and reproduce most prolifically will contribute
the highest number of offspring to the next generation,
and with them their own variants of heritable traits. In
the course of several generations, this process will lead
to an increase in frequency for some variants and to
the demise of others, which, in a nutshell, means that
the population is evolving by natural selection. The
key issue is the difference in the net reproductive
success between individuals that possess certain
heritable variants of traits and those who do not.
Survival, mating success, fecundity, longevity, and
other factors contribute to it, but ultimately it can
be reduced to the number of offspring, technically
termed fitness. When a population is exposed to a new
environment—like at the invasion of a cave—certain
variants might prove advantageous over those favored
in the old environment, and their bearers will produce
more offspring. After some generations, their share in
the population will have increased. Consequently, the
average fitness of the members of that population
will have increased, and the population as a whole
will become better adapted to the new environment.
The traits contributing to higher survival and repro¬
duction are called adaptations.
Cave animals often differ strikingly from their close
surface relatives in phenotype {i.e., the totality of biologi¬
cal form and function), but much less so in the
heritable information in their genes, or genotype. At the
same time they tend to resemble other cave species.
even if not closely related. This phenomenon is known
as convergent evolution. Features like eyelessness, depig¬
mentation, and elongated appendages are among the
most common cave-related traits, or troglomorphies.
Together with convergent changes in physiology, life
history, and behavior (see articles in this volume on
behavioral and morphological adaptations), they form a
rich source of data demonstrating the power of natural
selection. Repeated evolutionary convergence in the
same kind of environment but at different places, even
continents, can be better explained by natural selection
than by random genetic drift simply because the latter is
expected to produce random changes (Endler, 1986).
HISTORICAL DEVELOPMENT
The first one to mention the above argument in a
slightly modified form was none other than Darwin
(1859): "The principle which determines the general
character of the fauna and flora of oceanic islands,
namely, that the inhabitants, when not identically the
same, yet are plainly related to the inhabitants of that
region whence colonists could most readily have been
derived—the colonists having been subsequently mod¬
ified and better fitted to their new homes—is of the
widest application throughout nature. ... We see
this same principle in the blind animals inhabiting
the caves of America and of Europe." He attributed
the increase in the length of the antennae or palpi in
cave arthropods to natural selection, as a compensa¬
tion for blindness. The loss of eyes and pigment he
explained by disuse, a concept now considered obso¬
lete. In the decades to follow Darwin's discourse on
evolution in caves, the notion of natural selection shap¬
ing the bizarre appearance of cave animals was
increasingly losing ground. The debate centered on
regressed and lost organs and structures, chiefly blind¬
ness and skin pigment. Speleobiologists working in
North America (Arthur M. Banta, Carl H. Eigenmann,
Encyclopedia of Caves.
543
© 2012 Elsevier Inc. All rights reserved.
544
N
Alpheus S. Packard) and Europe (Rene G. Jeannel,
Armand Vire), although still being evolutionists, put for¬
ward nonselectionist and Neo-Lamarckian theories. They
were criticized by the Romanian naturalist Emile
Racovitza in his famous Essai sur les problemes biospeleolo-
giques (1907). He suggested that both the cave environ¬
ment and competition between and within species
exercised selection on cave-dwelling animals. He dis¬
agreed also with Darwin, who was hesitant in ascribing
a strong role to these two factors in caves. With regard to
the mechanisms of natural selection, Racovitza seemed to
adhere to non-Darwinian explanations about the inheri¬
tance of traits acquired and lost by the impact of environ¬
ment. Non-Darwinist views have retained a strong
impact on the discipline until the 1960s. The last major
proponent of the nonselectionist view of subterranean
biology was the French biospeleologist Albert Vandel.
At about the same time, American speleobiologists
(Thomas C. Barr, Kenneth Christiansen, and Thomas. L.
Poulson) started a new chapter in the understanding of
the adaptive and functional significance of the morphol¬
ogy, physiology, life history, and behavior of cave ani¬
mals. They used laboratory experiments, quantitative
and statistical methods, and comparative approaches to
demonstrate the adaptive features of aquatic and terres¬
trial cave life. However, while their work brought an
end to metaphysical explanations of evolution in caves,
it was still focused on adaptive explanations, not the
processes that gave rise to them. The difficult job of
showing that natural selection actually takes place in
caves, and of explaining its mechanisms, was under¬
taken by what we may call the second generation of the
American selectionist school of biospeleology, repre¬
sented by David C. Culver, Daniel W. Fong, Ross T.
Jones, and Thomas C. Kane.
In the 1940s, the German zoologist and geneticist
Curt Kosswig provided an explanation for the regres¬
sion of structures that had apparently lost their biologi¬
cal function in the subterranean environment, known
as the neutral mutation theory. This was probably the
first evolutionary explanation for cave animals that was
completely free of metaphysical speculations. The the¬
ory actually accepts the importance of natural selection
in that it predicts what happens when it no longer acts,
like in the case of sight in total darkness. Mutations
that obstruct sight are normally removed by purifying
selection, but have a chance to spread in a cave popula¬
tion by random genetic drift. The process of random
genetic drift has later became accepted as a major evo¬
lutionary force at the level of genes and proteins, for¬
malized in Motoo Kimura's neutral theory of molecular
evolution in 1968. The question today is not whether
Kosswig's neutral mutation theory is correct, but how
important it is compared to natural selection in shaping
the subterranean phenotypes (Protas et al., 2007).
Finally, it should be noted that although evolution
in caves seemingly provides ideal examples of natural
selection, cave fauna never became a really important
model for studying natural selection, as did, for exam¬
ple, fruit flies and the Galapagos finches. Moreover,
the actual processes of natural selection in caves were
explicitly addressed by very few studies, and, in gen¬
eral, remain poorly understood.
THE ACTION OF NATURAL SELECTION
IN CAVES
Most of our understanding of subterranean biology
rests on the assumption that troglobionts are in some
way functionally adapted to the subterranean environ¬
ment. However, by hypothesizing about the possible
functional role and adaptive value of troglomorphic
traits we are not studying the evolutionary process. To
directly measure and demonstrate selection in subter¬
ranean environments is not easy. It requires an accu¬
rate knowledge about the variation of the examined
trait(s), a functional explanation of the biological per¬
formance of different variants, evidence of their herita-
bility, and measurement of the actual difference in
fitness (survival and reproduction). The only work so
far to explicitly address all of the above points is a
large scale monographic account on the amphipod
crustacean Gammarus minus (family Gammaridae) by
Culver et al. (1995).
This aquatic amphipod inhabits surface springs and
caves in a wide area in the eastern and central United
States, usually not showing any morphological differ¬
ences between habitats. Morphologically distinct, trog¬
lomorphic populations have evolved only in the long
caves in the karst areas of Greenbrier Valley, West
Virginia, and Wards Cove, Virginia. Several cave popu¬
lations originated from separate cave invasions and
acquired the troglomorphic phenotype independently
from one another. They are genetically isolated from
nearby surface populations. Their cave-related traits—
prolonged antennae, reduced eyes and skin pigment,
modified brain morphology with increased olfactory
and decreased optic ganglia—are less expressed than in
many other cave crustaceans, but still distinct enough
to be measured. That makes them ideal for a compari¬
son of natural selection between cave and surface envir¬
onments. Perhaps the biggest difficulty of such studies
is to accurately measure fitness of different phenotypes.
It is neither possible to follow individual amphipods
throughout their life in their natural habitat nor to
count their offspring. Therefore, Culver and coworkers
measured two out of several components of the overall
fitness related to reproductive success: frequency of
amplexus (a precopula phase of tight physical contact
ENCYCLOPEDIA OF CAVES
NATURAL SELECTION
545
between a male and female amphipod) and number of
eggs carried by the female in her brood pouch.
Amplexing animals and those with more eggs were
considered to have higher fitness. Using these data,
they estimated selection gradients for three traits that
are often associated with the transition to subterranean
environment: body size (represented by head length),
antennal length, and eye size. They found convincing
evidence that natural selection in caves favors larger
animals, longer antennae, and reduced eyes. This is in
line with expectations, because larger animals are, in
the absence of predators, stronger competitors, energeti¬
cally more efficient, and produce more offspring.
Likewise, longer antennae enhance the sensory abilities
in a lightless environment. The possible advantage of
superficially reduced eyes is less clear, and the effect
might be caused by genetic correlation with another,
unmeasured trait. In surface springs, larger size of all
three traits was selected for. It is harder to explain why
longer antennae were selected for in this environment.
As important as identifying the traits under selec¬
tion is the question about the mode, direction, and
strength of selection. The view of caves as extreme
environments implies monotonous directional selec¬
tion toward a fully adapted troglobiont (Fig. 1A), one
that minimizes variation. Contrary to these predictions,
cave populations often show remarkable amounts of
morphological variability and genetic polymorphism
that would better fit a model of balancing selection.
The only ones to explore this problem in caves were
Culver et al. (1994) using their model organism
Gammarus minus. They found that directed selection
for larger body, longer antennae, and smaller eyes
indeed strongly prevailed in caves. A third mode of
selection, pulling trait values apart and thereby dis¬
rupting the variability (Fig. IB) might be involved in
colonization and speciation, and is discussed below.
The strength of selection is closely linked to the
time required for a new cave invader to evolve its
adaptive troglomorphic traits. Culver et al. (1995) esti¬
mated that a wild population would need some
100,000 years to become eyeless and otherwise troglo¬
morphic. This estimate agrees well with the facts that
most molecular clock datings gave cave species that
much time or more, and that very few troglomorphic
species are found in areas where the subterranean hab¬
itat has only been available since the last glacial retreat
about 10,000 years ago. On the other hand, this seems
rather slow in the light of several new reports about
both vertebrates and invertebrates adapting to new
environments and speciating within tens or hundreds
of generations. Also, strong directional selection and a
consequent fast evolution are to be expected under the
sharp ecological gradient from the surface to the sub¬
terranean environment.
Less More
Measured degree of troglomorphism
FIGURE 1 Modes of natural selection in caves. (A) Directional
selection: When isolated from the surface ancestor, for example, by its
extinction, a new cave species undergoes directional selection lead¬
ing to adaptation to the new environment. (B) Divergent selection:
When the surface ancestor remains in contact with the cave popula¬
tion, divergent selection drives trait values of the cave and the sur¬
face form apart. This kind of selection can lead to ecological
speciation.
A few cases might nevertheless be pointing to a
very fast evolution in caves. One is the troglomorphic
form of the aquatic isopod Asellus aquaticus (family
Asellidae). Several local, independently evolved cave
populations and species are known from the Dinaric
Karst in Slovenia and Italy and from Dobruja Plateau
in Romania. Multivariate morphometric comparisons
between historical and recent samples suggest that one
of them might have invaded Postojna Cave and
evolved a troglomorphic appearance in no more than
80 years (Prevorcnik et al., 2010).
The same species in the same cave demonstrates
how natural selection interacts with gene flow to form
a cline (Fig. 2; Sket, 1985). A population of nontroglo-
morphic, normally eyed surface phenotypes of Asellus
aquaticus lives in the surface freshwaters outside
Postojna Cave including the sinking river passing
through the cave. Deeper in the cave, a troglomorphic
population of depigmented and eyeless individuals
inhabits the same sinking river. Individuals from the
surface migrate into the cave following the flow direc¬
tion. Within the first kilometer from the entrance, the
nontroglomorphic phenotype strongly predominates.
From approximately kilometer two, troglomorphic
ENCYCLOPEDIA OF CAVES
546
N
FIGURE 2 Phenotypic dines along the sinking
Pivka River in the Postojna-Planina Cave System,
Slovenia. The steep dine in 1934—1935 is indica¬
tive of strong selection against the surface form.
Forty years later, the dine was much flatter, prob¬
ably because of increased nutrient input by
organic pollution of the Pivka River, changing the
selective regime in favor of the surface pheno¬
type. Adapted from Sket (1985), simplified. Actual
counts are represented only at the starting, break-
c 0 1 2 3 4 5 6 7 8 910 ing, and end points of the curves.
=3
CT
Subterranean river course in kilometers
Asellus begin to increase in number together with mor¬
phologically intermediate animals. Gradually, their fre¬
quency increases until at about kilometer four the
surface phenotype becomes very rare. The width of the
cline is determined by the rate of immigration relative to
selection acting against immigrants. Interestingly, in the
first half of the twentieth century C. Kosswig observed a
much narrower cline with the surface phenotype becom¬
ing rare already within the first kilometer. As it is
unlikely that the mobility of the animals would have
changed during that time, we can assume a change in
selection pressure. This was induced by organic pollu¬
tion introducing nutrients and thus making the subterra¬
nean river more hospitable to the surface animals.
THE SELECTIVE REGIME IN CAVES
The selective regime, or selective environment, com¬
prises those environmental and organismic factors that
determine how natural selection will act on character
variation. It is important to have an idea about the key
selective factors of the cave environment if we wish to
know whether a troglomorphic trait evolved under the
impact of natural selection. The usually mentioned criti¬
cal factors in caves are complete and constant darkness
and food scarcity. Other features, such as absence of pre¬
dators and competitors, constancy of the environment,
and high humidity, are more likely to loosen the selec¬
tive pressure than to intensify it. Studies on North
American amblyopsid cavefish postulated strong selec¬
tive constrains by food scarceness in a permanently
lightless aquatic environment. The flattening of the cline
between cave and surface Asellus phenotypes following
increased availability of food (Fig. 2) suggests a similar
selective regime for aquatic crustaceans. Conversely, the
selective regime for cave populations of Gammarus minus
mentioned above was postulated to be determined by
darkness alone, as the cave streams have considerable
organic input. Culver et al. (2010) proposed that different
kinds of aquatic subterranean environments (deep
groundwater, cave streams, epikarst, superficial subterra¬
nean habitats) differ in the availability of food and space,
and that these differences should influence the selective
regimes in those habitats. Using this prediction they tried
to explain the morphological variability within the North
American subterranean amphipod genus Stygobromus
(family Crangonyctidae). Contrary to expectations, they
could find no support for the hypothesis that differences
in antennal elongation reflect differences in food and
nutrient supply. Given that all Stygobromus species are
eyeless and depigmented, they concluded that absence
of light was the single key selective factor responsible for
general troglomorphic appearance.
Few examples demonstrate that other features of
the cave environment can contribute to the selective
regime. One is the permanently wet substratum on
which surface collembolans would have trouble mov¬
ing, and in response to which troglobiotic species have
evolved a special foot structure and locomotory behav¬
ior. Another one is a special cave microhabitat consist¬
ing of films of water moving down cave walls, known
as cave hygropetric (Sket, 2004). The water contains
organic particles, but special adaptations on legs and
ENCYCLOPEDIA OF CAVES
NATURAL SELECTION
547
mouth parts are required to exploit them. Several lep-
todirine cave beetles have evolved such adaptations
convergently. This shows that the cave environment
not only poses selective constraints to its colonizers
but also offers ecological opportunities, for example, in
the form of new trophic niches.
This idea was examined on the Western Palearctic
amphipod genus Niphargus (family Niphargiadae).
The high morphological variation among coexisting
Niphargus species can be explained by their adapta¬
tions to microhabitats within caves (Trontelj et al.,
unpublished). Up to nine species inhabit a single cave.
Large-bodied species with long antennae and legs are
consistently associated with phreatic cave waters, stag¬
nant or slowly flowing and with an abundance of
available space. Large but slender species with short
antennae and legs show a preference for cave streams
with high flow velocity. Small-bodied species are asso¬
ciated with either of the two small pore microhabitats:
the epikarst and the cave interstitial. The limiting fac¬
tor in these two habitats appeared to be body diameter
rather than body length. This demonstrates that selec¬
tive regimes in caves can be more complex than simple
directional selection for compensation of eyesight and
energetic efficiency. What at a first glance seems to be
different stages of adaptation might in reality be adap¬
tations to different niches within a cave. Finally, the
coexistence of several closely related species means
that biotic interactions such as competition and preda¬
tion add to the selective regime.
THE ROLE OF NATURAL SELECTION IN
COLONIZATION AND SPECIATION
Speciation—the origin of new species—often occurs
by means of natural selection. One particularly impor¬
tant way for new cave species to evolve is by divergent
natural selection operating between the new cave inva¬
ders and the part of the population that remains in the
original surface environment. Speleobiologists have
been aware of this mode of colonization for about
thirty years thanks to the work of Francis G. Howarth
on Flawaiian cave arthropods. Flistorically, in subterra¬
nean biology, the distinction between active and pas¬
sive colonization and colonization in the presence or
absence of surface ancestors was considered more
important than the direct role of selection in this pro¬
cess. Less attention has therefore been paid to the
question of how selection leads to speciation than to
the question of whether a particular cave fauna had
originated by adaptive shift or not. On the other hand,
speciation by divergent natural selection, known also
as ecological speciation, has been in the wider focus
of speciation research for many years. Generally,
ecological speciation refers to the evolution of repro¬
ductive isolation between populations or subsets of a
single population by adaptation to different environ¬
ments or ecological niches.
Natural selection can be directly involved when
new cave species are budding off from a surface ances¬
tor. There is an important distinction between the
mode of selection involved here and the more familiar
directional selection, although both will eventually
result in the same troglomorphic adaptations such as
longer legs and antennae, smaller eyes, lower metabo¬
lism, or a longer life span. This distinction occurs
when the cave invaders remain in contact with the
ancestral population, for example, through repeated
immigration from the surface. Selection now favors
one extreme of a trait in the surface environment and
the other one in the cave (Fig. IB). Theory predicts that
the population subdivided between two different
environments will adapt locally according to the differ¬
ences in selective regimes. Gene flow between both
environments may still persist, but selection will act
against it. Traits favored in one environment will be
disadvantageous in the other. Intermediate, heterozy¬
gous individuals will have lower fitness in both envir¬
onments. Speciation with ongoing gene flow can only
take place under strong divergent selection that results
in some form of reproductive isolation.
The reproductive barriers leading to speciation in
caves have not been thoroughly studied, but a handful
of studies suggest that they may be quite diverse. The
most obvious is isolation by habitat where the non-
adapted surface species simply cannot penetrate deep
enough in sufficient numbers to mate and maintain
the contact to the cave population. A case of strong hab¬
itat isolation was reported for locally adapted popula¬
tions of the Atlantic molly (Poecilia mexicana, family
Poeciliidae) in the Cueva del Azufre system in southern
Mexico (Plath et al., 2010). In addition to the lightless
cave environment, toxic hydrogen sulfide is another
important selective factor there. Adaptation to divergent
environmental conditions and selection against immi¬
grants from ecologically different populations drives
the genetic differentiation in this system. An exception¬
ally strong flood with the potential to intermix animals
from all parts of the system occurred in 2007. Plath and
coworkers used this event as a natural perturbation
experiment. They compared the genetic structure of
populations from different environments before the
flood to the structure encountered two months after the
flood. They found that the catastrophic flood displaced
individuals among compartments of the same type of
environment, but the structure between environments
remained unchanged. What this means is that natural
selection has eliminated—probably by mortality—all
involuntary immigrants from their non-native habitats.
ENCYCLOPEDIA OF CAVES
548
N
This example might be extreme, but it shows that isola¬
tion by habitat is enough to keep the cave and the sur¬
face population from admixing even if there are no
physical barriers to migration. Similarly, the habitat bar¬
rier probably plays a role in keeping some cave popula¬
tions of Asellus aquaticus in the Postojna-Planina Cave
System separate from the surrounding surface popula¬
tions. In captivity, they interbreed successfully, but in
the wild there is very little or no genetic exchange
between the surface and cave populations (Verovnik
et ah, 2003; Protas et ah, 2011).
Another type of reproductive barrier was suggested
by Niemiller et ah (2008) to be responsible for the
ecological speciation of cave salamanders of the genus
Gyrinophilus (family Plethodontidae). The cave forms
are paedomorphic like many troglobiotic salamanders
(see the article "Salamanders" in this volume).
Paedomorphosis, and the permanent aquatic lifestyle it
is associated with, is believed to be an adaptation to
the cave environment, where food and space are more
abundant in the aquatic than in the terrestrial part.
As a side effect of this adaptive divergence, paedo¬
morphosis may contribute to reproductive isolation.
Paedomorphs must court and mate in water while
metamorphosed Gyrinophilus have only been observed
mating on land.
CONCLUSIONS AND PROSPECTS
The evolving cave fauna offers powerful models for
the study of natural selection. Although researchers
have been aware of this potential ever since Darwin,
few have made use of it. Biospeleology has been
historically burdened by non-Darwinian explanations
of evolution that have left little space for natural
selection. Much effort has therefore been put into
demonstrating the adaptive value of troglomorphic
characteristics, leaving the vast field of the actual evo¬
lutionary processes virtually untouched. This has
changed since 1990 when natural selection was exten¬
sively studied on the amphipod Gammarus minus. We
know since then that reduced organs—for example,
smaller eyes—are selected for in caves, but selected
against outside. We know that selection for other trog¬
lomorphic traits, such as longer appendages and larger
body size, is directional. Individuals with mixed traits
probably result from ongoing colonization processes or
incomplete reproductive isolation of the cave popula¬
tion, and are selected against, rather than being the
result of stabilizing selection. Food scarcity and dark¬
ness seem to be the main common factors shaping the
selective regime in caves, but interactions between spe¬
cies might be important in caves with communities
of ecologically similar or closely related species. The
coexistence of several species of niphargid amphipods
is possible because they are adapted to microhabitats
within a cave, which is likely the result of divergent
natural selection. Natural selection is important also
for the colonization and speciation under the adaptive
shift model, more generally known as ecological specia¬
tion. Divergent selection between the surface and the
cave environment can produce local adaptations that
bring about reproductive isolation. Speciation can hap¬
pen while individuals from the surface continue to
immigrate if selection acts against the surface pheno¬
type inside the cave and vice versa. Reproductive isola¬
tion can also evolve as a by-product of adaptation to
the cave environment, for example, by paedomorpho¬
sis in cave salamanders.
Cave animals are traditional examples of evolution,
but are young models in natural selection research.
Many questions remain to be answered or even
addressed. Most of our knowledge comes from aquatic
cave animals such as cavefishes, amphipods, isopods,
and salamanders; the validity of models and concepts
needs to be verified on the terrestrials. The strength
of selection and the time needed for a troglomorphic
species to evolve—anything between 100 and 100,000
years—are still under debate. So is the role of selection
in the origin of cave species; here the debate has joined
the wider field of speciation research, so examples
from caves are likely to attract more general interest.
Almost nothing is known about how natural selection
shapes the diversity of cave species within caves,
that is, the process of adaptive radiation following the
initial colonization. Here, the interesting questions
include evolutionary tradeoffs between adaptations to
various facets of the cave environment, strength and
forms of competition and other interactions, and speci¬
ation within the subterranean realm without the aid of
sharp dine to the surface.
No other scientific debate has marked biospeleol¬
ogy like the debate about the role of neutral mutations
and selection in the evolution of reductive traits. The
dispute is likely never to be solved in favor of one or
the other side, not only because both processes are
probably involved but also because the crux of the
problem is often not presented clearly enough. If the
question is, are neutral mutations involved in eye and
pigment degeneration, the answer has to be a clear
"yes." And if the question is, are degenerated or
reduced eyes adaptive in the cave environment, the
answer might be a cautious "probably not in itself."
However, if one asks what processes are involved in
the developmental background that phenotypically
manifests itself as reduced eyes and body pigment,
the answer will most likely include natural selection.
Through this debate, the most advanced scientific
studies in biospeleology have been motivated, ranging
ENCYCLOPEDIA OF CAVES
NEUTRAL MUTATIONS
549
from extensive laboratory breeding and classical
genetic experiments to developmental genetic and
genomic research. It also points the way to the future
of evolutionary speleobiology, where genomics meth¬
ods will be used to identify what genes are under nat¬
ural selection, and how the changes at these genes
repeatedly yield the morphological, physiological,
behavioral, and other adaptations to cave life.
See Also the Following Articles
Neutral Mutations
Morphological Adaptations
Behavioral Adaptations
Bibliography
Culver, D. C., Holsinger, J. R., Christman, M. C., & Pipan, T. (2010).
Morphological differences among eyeless amphipods in the
genus Stygobromus dwelling in different subterranean habitats.
Journal of Crustacean Biology, 30, 68— 74.
Culver, D. C., Jernigan, R. W., O'Connell, J., & Kane, T. C. (1994).
The geometry of natural selection in cave and spring populations
of the amphipod Gammarus minus Say (Crustacea: Amphipoda).
Biological Journal of the Linnean Society, 52, 49—67.
Culver, D. C., Kane, T. C., & Fong, D. W. (1995). Adaptation and natu¬
ral selection in caves: The evolution of Gammarus minus.
Cambridge, MA: Harvard University Press.
Darwin, C. (1859). On the origin of species by means of natural selection,
or the preservation of favoured races in the struggle of life. London:
John Murray.
Endler, J. A. (1986). Natural selection in the wild. Princeton, NJ:
Princeton University Press.
Kimura, M. (1968). Evolutionary rate at the molecular level. Nature,
217, 624-626.
Plath, M., Hermann, B., Schroder, C., Riesch, R., Tobler, M., Garcia
De Leon, F. J., Schlupp, I., & Tiedemann, R. (2010). Locally
adapted fish populations maintain small-scale genetic differentia¬
tion despite perturbation by a catastrophic flood event. BMC
Evolutionary Biology, 10, 256.
Prevorcnik, S., Trontelj, P., & Sket, B. (2010). Rapid re-invasion and
evolution following the mysterious disappearance of Racovitza's
Asellus aquaticus cavernicolus (Crustacea: Isopoda: Asellidae). In
A. Moskric & P. Trontelj (Eds.), Organizing Committee of the 20th
International Conference on Subterranean Biology, Abstract book
(p. 172). Postojna.
Protas, M., Conrad, M., Gross, J. B., Tabin, C., & Borowsky, R. (2007).
Regressive evolution in the Mexican cave tetra. Astyanax mexica-
nus. Current Biology, 17, 452—454.
Protas, M. E., Trontelj, P., & Patel, N. H. (2011). Genetic basis of eye and
pigment loss in the cave crustacean, Asellus aquaticus. Proceedings of
the National Academy of Science (U.S.A.), 108(14), 5702—5707.
Racovitza, E. G. (1907). Essai sur les problemes biospeologiques.
Archives due Zoologie Experimental et Generate, 6, 371—488.
Sket, B. (1985). Why all cave animals do not look alike: A discussion
of the adaptive value of reduction processes. National Speleological
Society Bulletin, 47, 78—85.
Sket, B. (2004). The cave hygropetric—A little known habitat and its
inhabitants. Archiv fur Hydrobiologie, 160, 413—425.
Verovnik, R., Sket, B., Prevorcnik, S., & Trontelj, P. (2003). Random
amplified polymorphic DNA diversity among surface and subter¬
ranean populations of Asellus aquaticus (Crustacea: Isopoda).
Genetica, 119, 155—165.
NEUTRAL MUTATIONS
Horst Wilkens
Zoologisches lnstitut und Zoologisches Museum,
University of Hamburg
The paradigm of selection dominates all modern inter¬
pretations of evolutionary processes and provides their
most efficient explanation. However, this is in contrast
with the conditions characteristic of the basal process
of molecular evolution. Here mutation pressure and
drift play the major role (Kimura, 1983), whereas as a
rule selection only gets involved secondarily, at the
phenotypic level. The question is whether these basic
principles of molecular evolution, random mutations,
and drift may also manifest in phenotypic evolution
and can, in some cases, be traced here.
The reduction and rudimentation of structures is
very common throughout nature. It can be observed in
wingless birds and insects such as ostriches and cara-
bid beetles. Also baleen whales, which have not only
reduced their hind legs but also the teeth, provide
spectacular examples. The most conspicuous and curi¬
ous phenotypes are found in cave-dwelling animals.
Eye loss and paleness of the so-called troglobites have
stimulated the thinking of scientists from the beginning
of research. Even Darwin was aware of the bizarre phe¬
notypic appearance of cave animals, which seemed to
him "wrecks of ancient life." Cave animals are a central
object for the study of the causes of regressive evolu¬
tionary processes, because most reduced traits may be
no longer influenced by persisting biological functions.
As a unique model system in research of regressive
evolution the characid Astyanax fasciatus (Cuvier, 1819)
( = A mexicanus ; Filippi, 1853) from Mexico has served
since the middle of the twentieth century (Wilkens,
1988, 2010; Jeffery, 2001). This widely distributed sur¬
face fish is the sister form of more than 20 troglobitic
populations distributed in different caves within a
rather restricted karst area in northeast Mexico
(Mitchell et al., 1977). Unlike the majority of surface
species entering the hypogean realm, the large-eyed
surface Astyanax is not a true troglophile, which, like
catfish, are usually characterized by extremely negative
scotophilia and adaptations to life under light-poor
conditions (Langecker, 2000). Specimens of A. fasciatus
did not colonize caves voluntarily, but got trapped in
the underground without the possibility of escape
(Wilkens, 1988). They were able to survive in darkness
because of preadaptations such as chemically triggered
spawning behavior and because they could orientate
in darkness on behalf of a well-developed lateral line
sense (Wilkens, 2010). Whereas traits such as olfaction,
taste, lateral line, yolk content, and fat storage were
constructively improved by directional selection in
Encyclopedia of Caves.
©2012 Published by Elsevier Inc.
550
N
darkness, those having become biologically function¬
less (e.g., dorsal light reaction, circadian rhythm,
shoaling, optically triggered aggressive behavior, pig¬
mentation, eyes, and so on) were subjected to reduc¬
tion (Langecker, 2000; Wilkens, 1988, 2010).
Surface and cave forms of Astyanax are closely
related and thus morphological, physiological, and
behavioral comparisons can be performed to analyze
their adaptations to the cave habitat. As surface and
cave forms are interfertile, it was possible to study the
genetic basis of regressive as well as constructive traits
by crossing analyses. The material examined was
derived from nine cave populations (Sotano de la
Molino, Cueva de El Pachon, Sotano de la Tinaja,
Cueva de los Sabinos, Sotano de Yerbaniz, Sotano de
las Piedras, Cueva de la Curva, Cueva Chica, and
Cueva del Rio Subterraneo = Micos cave).
One of the most conspicuous traits characterizing
cave animals is the loss of coloration. In cave Astyanax
this relies on several different factors. Albinism, the
complete loss of melanin, is known from three popula¬
tions (Molino, Pachon, Yerbaniz). It is based on one
recessive allelic gene, for which all three populations
are homozygous. Molecular studies in two cave popu¬
lations, Pachon and Molino, showed that albinism in
them can be attributed to different deletions in the
OcaZ pigment gene, responsible for loss of the corre¬
sponding protein OCA2 (Protas et ah, 2007). Thus
the inability to convert L-tyrosinase to L-dopa and
melanin (Jeffery, 2001) seems to rely on the before-
mentioned mutation and is convergently achieved.
With the exception of the Molino and the Micos cave-
fish, all populations mentioned above are homozygous
for the recessive "brown gene," due to which the
melanophore color cells as well as the eye pigment
epithelium only show partially reduced melanin con¬
tent. This so-called brown phenotype has arisen inde¬
pendently in geographically separate caves and is
mediated by different mutations in the same "brown"
gene (Gross et al., 2009).
Also the amount of guanine in the scales, which is
camouflaging the fish in the water by its silvery shine
under normal light conditions, is slightly reduced in all
cave populations. However, in two of them, different
recessive genes are found, which reduce the guanine
content of the scales in part (Molino) or completely
(Piedras). The Piedras population was found to be het¬
erozygous for this gene. Another regressive mutation
(yellow gene) causes bright yellow body color. This
probably results from the disturbed ability to decom¬
pose carotenoids ingested with the food. For this, a
recessive gene is responsible, which was only found in
the Piedras population (Culver and Wilkens, 2000).
Besides by mere monogenic color mutations, reduc¬
tion of the dark melanin pigmentation of Astyanax
cavefish is additionally caused by a decrease of the
number of melanophores, of which only about 20% are
left. Inheritance of this feature is independent from
that of the albino and the brown genes, a fact that is
often not considered in studies. Crossings have reve¬
aled a polygenic basis for this (Wilkens, 1988; Culver
and Wilkens, 2000). Like the color mutations, this fea¬
ture is already manifested at early embryogenesis, dur¬
ing which fewer precursor cells, the melanoblasts, are
developed in the neural crest in the cave forms.
Besides loss of pigmentation, the reduced eyes of
cave animals have generated the greatest interest of
scientists. In the Astyanax cavefish the eye rudiments
show varying degrees of differentiation and sizes. In
correlation to this, they are sunk below the body sur¬
face at differing extents. The degree of differentiation
is assumed to be dependent on the phylogenetic age of
the cave populations. However, all of them are smaller
in size and show less structural differentiation. In
adults of the most reduced cavefish, neither lenses nor
visual cells are developed. Only those retinal layers
not involved in the formation of the visual cells can be
found in the eyes best differentiated. In the most
reduced ones, the sclera just encloses undifferentiated
remnants of nervous, pigment, and chorioid tissues
(Wilkens, 1988).
In crossing experiments between surface and cave¬
fish, no genes could be analyzed to be responsible for
specific eye structures such as lens or retina. It was
only found that size and degree of structural differenti¬
ation are in principle correlated and that the larger an
eye is, the better its single structures are developed.
The genetic factors influencing eye size were called eye
genes. Crossing experiments and quantitative trait loci
(QTL) studies revealed that at least six of them were
responsible for eye reduction in the cavefish (Wilkens,
1988; Protas et al., 2007). These eye genes seem to have
mutated negatively. The defect eye genes are assumed
to be randomly distributed in the different cave popu¬
lations providing a partially diverging genetic basis.
They seem to be at least in part heterozygous. This can
be concluded from the fact that the hybrid offspring
between different cave populations, may develop
larger and correlated to this, better differentiated eyes
than the parental generation (Wilkens, 1988, 2010).
The genetic character of the eye genes, and which
they are, is still being studied. Mapping of developmen¬
tal control genes such as sonic hedgehog (shh), twiggy
winkle hedgehog (tzvhh), and Pax6 revealed that no eye
QTLs are located near these loci. This result makes it
unlikely that mutations in any of these control genes are
directly responsible for eye regression (Protas et al.,
2007). Gene expression and sequencing data showed
that destructive loss-of-function mutations have proba¬
bly not or only in few cases occurred in the cavefish eye
ENCYCLOPEDIA OF CAVES
NEUTRAL MUTATIONS
551
gene cascade, including those structural genes that func¬
tion at the bottom of the regulatory cascades (Jeffery,
2001). This finding is also supported by the observation
that all the different-sized eyes developed in hybrids
between cave and surface forms show all the structures
characteristic of an eye such as lens, lens muscle, pupil,
and retinal layers (Wilkens, 1988, 2010). Only the size of
these structures generally varies in correlation with
overall eye size. This observation furthermore shows
that the genes involved in eye development have not
mutated and can potentially be expressed. For example,
a red-like opsin gene is transitorily transcribed in
Piedras cavefish during early ontogeny (Langecker,
2000) and the a-A-crystalline gene, which is responsible
for the development of a proper lens in surface fish,
shows no destructive mutations, but is never activated
in strongly reduced cavefish (Wilkens, 2010).
Thus, it can be assumed that eye regression in
Astyanax seems to be mainly because of down regulation
by the expanded hh gene expression, which was sug¬
gested to have a causal role in eye development of the
cavefish (Yamamoto et ah, 2004). The question of what
causes down regulation of hh genes as yet remains
unsolved. However, deeper insight into eye genetics
can be derived from crossings between different cave
populations: whereas hybrids between surface and cave
populations contain eyes of all possible intermediate
sizes between the parental forms, the crossings between
different cavefish populations very often may develop
slightly larger and better differentiated eyes than the
parental cavefish. Usually, rudimentary lenses are devel¬
oped in adult cavefish hybrids (Wilkens, 1988, 2010). An
exceptional situation is found in the F2-cross between
the Pachon and the Molino cavefish populations, in
which "back-to-surface eyes" develop. Among a large
number of specimens with small eyes that, as in the
parental forms and in their FI-cross, are sunken into
the orbital cavity, several of the adult specimens have
externally visible eyes with wide pupils. Histological
sections of the eyes from these latter specimens show
the presence of large transparent crystalline lenses as
well as retinas containing all optical layers with morpho¬
logically intact visual cells including outer segments,
pigment epithelium, optic nerve, and lens muscle
(Wilkens, 2010). Keeping in mind that expanded expres¬
sion of the hh genes may have a causal role in eye
reduction of the parental cavefish, the development
of back-to-surface eyes could be explained by the sec¬
ondary restriction of expression and secondary down-
regulation of hh genes by as yet unidentified genes.
In contrast to the unmutated hh genes, it is likely that
these as yet unidentified genes show loss-of-function
mutations. In back-to-surface eyes they could be ex¬
pressed again because of the complementary restitution
of their function. The complementary restitution then
would be brought about by the recombination of differ¬
ent loss-of-function genes in cross hybrids, which derive
from different geographically remote cavefish popula¬
tions that have evolved independently. As shown
above, the phenomenon of varying loss-of-function
mutations occurring in the same gene in geographically
remote cave populations, but nonetheless causing the
same phenotypic effect in them, has also been found for
the albino gene as well as for the brown gene (Protas
et ah, 2007; Gross et ah, 2009).
The as yet unidentified genes regulating hh expres¬
sion are suggested to be equivalent to those eye genes
that were found by crossing experiments to be respon¬
sible for eye size (Wilkens, 1988). As mentioned above
they were quantitative in character and were originally
called "eye genes" because it was not possible to iden¬
tify specific genes, such as ones responsible for the
lens or visual cells, with the crossing experiments.
Histological studies have shown that in the eye two
developmentally independent subunits, lens- and ret¬
ina-dependent structures, exist. Furthermore, it was
found that the cavefish growth curve at early ontogeny
is characterized by two important subsequent develop¬
mental stages contributing to eye regression (Fig. 1).
Therefore it was suggested that the eye genes divide
into two groups: the first group express upstream of
the hh genes during the first step of eye regression.
These genes included in the first group would deter¬
mine the size of the primordial eye cup and of the lens
through regulation of hh expression.
However, whereas the size and the degree of differ¬
entiation of the lens is determined during the first step
of eye development, that of the retina is induced both
during this first step and additionally during the sub¬
sequent second step. Thus a second group of eye genes
is suggested to act during this second step of eye
development, regulating structural genes such as those
responsible for the visual cell outer segment formation.
Hh genes could again be involved as a regulator in the
development of the retina, because it is expressed sev¬
eral times during eye development (Wilkens, 2010).
By stating that "It is an important fact that rudimen¬
tary organs, such as teeth in the upper jaws of whales
and ruminants, can often be detected in the embryo,
but afterwards wholly disappear" and that "It is also a
universal rule, that a rudimentary part or organ is of
greater size relative to the adjoining parts in the
embryo, than in the adult," Darwin recognized the
process of ontogenetic regression as a characteristic of
regressive evolution (Darwin, 1859, p. 430).
Whereas biologically functionless traits such as
body color in cave animals are reduced from the very
beginning of ontogeny, the eye is subjected to the
aforementioned process of ontogenetic reduction.
Cavefish are primarily developing smaller but
ENCYCLOPEDIA OF CAVES
552
N
0 Eye in mm
FIGURE 1 Growth of eyes in Astyanax surface and cave forms.
structurally almost complete eyes which, for example,
contain lenses and visual cells. These are getting
reduced during later stages of development and only
get completely lost in adult fish. This process can be
explained as caused by developmental functions: for
example, the eyeball provides by induction the correct
formation of the head skeleton. It is this internal
selective force that will eliminate all regressive eye
structural mutations and which might heavily disturb
eye- and, as a consequence, head-building. Disruptions
in the formation and evagination of the optic primor-
dia can result in defects in both ocular and craniofacial
development. Only those mutations that do not totally
disturb eye formation, as, for example, many artificially
induced mutations in the zebrafish ( Danio rerio) or the
medaka ( Oryzias latipes) do, are not subjected to elimi¬
nation by selection in cavefish in which regression can
be studied under "natural conditions." The existence of
internal selection is demonstrated by the variability of
cavefish eye size during early ontogeny, which is not
higher than in the Astyanax surface fish until a certain
size is reached. Only during the subsequent growth
stages the characteristic variability of cavefish eye size
develops and enhances, because internal selection
has ceased (Fig. 1). Principally similar observations
were made in cave-living catfish ( Rhamdia spp.,
Pimelodidae) and the cyprinid cavefish Garra barreimiae
during ontogeny. Summarizing, it can be concluded
that only part of the genes involved in eye formation
has a neutral character, whereas others are still sub¬
jected to internal developmental selection—a character¬
istic of ontogenetic regression, which, in the case of the
cavefish eye, can be coined the eye paradox.
The enigmatic disappearance of traits in cave ani¬
mals has stimulated a large series of different explana¬
tions. It is still under dispute, and time and again old
theories of eye regression are revived once more and
new ones are put forward.
One of these is the deleterious risk hypothesis, which
suggests that an exposed eye was a deleterious risk in
ENCYCLOPEDIA OF CAVES
NEUTRAL MUTATIONS
553
darkness. This hypothesis seems unlikely, because
cave animals usually derive from troglophile ancestors,
which are already adapted to a partial lifestyle in
darkness. Such species are mostly night-active and still
have small eyes, because selection is acting against
the complete loss of eyes. However, they have already
improved senses such as taste, olfaction, or orientation,
which enable them to live in complete darkness,
too. Furthermore the deleterious risk hypothesis is not
apt to explain the reduction of all other biologically
functionless traits getting lost in darkness.
Also the energy economy hypothesis has long been
considered a driving force of eye reduction. Recently,
it was speculated that the high energetic costs of visual
cell metabolism could be a strong selective force in eye
regression. This assumption was based on the finding
that neural energy costs of fly visual cells do not lower
in darkness when kept in a light—dark 12:12 cycle.
These results partially concur with findings of retinal
protein synthesis in toads, in which it was found
that protein synthesis did not change during the nor¬
mal light—dark cycle. However, major differences in
protein synthesis were evident when animals were
stressed with continuous darkness for several days.
Under these conditions, a 50% lowering of retinal pro¬
tein synthesis, which can be considered an indicator of
retinal metabolism, took place (Wilkens, 2010).
It was furthermore found in the fish retina that
light responsiveness of cone horizontal cells is even
suppressed in prolonged darkness. These results show
that the effects of darkness on the vertebrate retina are
quite complex and it is questionable whether conclu¬
sions can be drawn from studies in fly visual cells. To
date, no measurements of energy costs of true cave
animal retinas have been performed. As the entire
ontogenetic development of cave species takes place in
complete darkness and thus proceeds without any
light stimulus, one can argue that this cannot be com¬
pared at all with animals kept in a normal light—dark
12:12 cycle even when measured in darkness. This
assumption is substantiated by histological analysis of
retinas of cave-living fish sampled under the exclusion
of light in their original habitat. The fish bred in dark¬
ness showed considerable destructive malformations
of the retina, whereas their offspring, hatched and kept
at 12:12 light conditions, developed a completely intact
and functional retina (Wilkens, 2010). Therefore, physi¬
ological deficiencies caused by retinal destruction can
be expected in the original cavefish. To date, it is ques¬
tionable whether energy economy is responsible for
cavefish eye reduction.
The hypothesis of energy economy as a selective factor
in eye regression is furthermore improbable because
behavioral traits, the display of which depends on vision.
are reduced although the ancestral colonizer did no
longer perform them in darkness and hence they were
not subjected to selection. For example, in Astyanax
cavefish the visually triggered aggressive behavior as
well as the dorsal light reaction get reduced in the
cavefish, nonetheless (Langecker, 2000; Wilkens, 2010).
Because of the failure to detect specific selective
forces responsible for the regression of functionless
traits, selection was suggested to be mediated by plei-
otropy. The pleiotropy hypothesis was recently revived
again (Yamamoto et al., 2009). They tried to demon¬
strate it by conditional overexpression of an injected
shh transgene at specific times in development of
embryonic Astyanax, and found that taste-bud amplifi¬
cation and eye degeneration are sensitive to shh over¬
expression during early developmental periods. From
this observation, it was concluded that pleiotropy
mediated developmental tradeoffs between the regres¬
sive eye and oral/pharyngeal constructive cave traits
such as jaw size and taste-bud number as well as feed¬
ing behavior. However, no reliable experimental proof
of this inverse relationship was provided. First, it was
attempted to verify it by generating an F3 generation,
which is an uncertain tool of genetic crossing analysis,
because whereas each individual in an F2 is an unbi¬
ased sample of alleles in the original FI, the F3 is a
biased sample which may actually create a spurious
correlation. Furthermore, the differences in taste-bud
numbers of surface and cavefish become significantly
notable at much later stages than those studied. Also
the assumption of Yamamoto et al. (2009) that the
wider cavefish mouth alone accounts for the substan¬
tially elevated numbers of taste buds in cavefish was
not confirmed. Last but not least, these findings are
also in contrast to QTL (Protas et ah, 2007, 2008)
and morphological studies, which did not reveal
correlations between eye size and taste-bud number
(Wilkens, 2010).
Independent inheritance of regressive and construc¬
tive traits was also shown for the lateral line sense,
which compensates the loss of vision in darkness by
enlargement of the size and the number of sensory
structures sensitive to the slightest water movements,
the neuromasts. Counting of the neuromasts on the
lateral head side of F2 crossings between surface and
cavefish did not reveal any statistical correlation
between the number of taste buds and eye size. The
distribution of neuromasts is even not expanded into
those areas of the cavefish head surface, where the eye
was originally located.
The independent assortment of constructive and
regressive traits seems to have general validity as was
revealed by studies of other constructive and regres¬
sive traits in F2-crosses between Astyanax surface and
ENCYCLOPEDIA OF CAVES
554
N
cavefish and is corroborated by QTL analyses (Protas
et ah, 2007). Quite generally the improvement of cave
adaptive traits cannot be explained as a tradeoff, pleio-
tropically mediated by the regression of biologically
functionless traits such as the eye or melanin pigmen¬
tation (Wilkens, 2010). It is suggested that this inde¬
pendence is even favored by selection because it
guarantees evolutionary flexibility, which is an advan¬
tage when adapting to a new environment, as cavefish
do (Wilkens, 2010).
Darwin (1859, p. 432) observed that "An organ
when rendered useless, may well be variable, for its
variations cannot be checked by natural selection." The
loss of the biological function of those structures for
which light is a transporter of information in cave ani¬
mals was recognized as the key event of their future
development by the German biologist Curt Kosswig
(Kosswig and Kosswig, 1940). He stated that because
of this, stabilizing selection would no longer act on
such features. As deleterious mutations became neutral
under these instances they were no longer eliminated
but accumulated randomly. In the initial phase of
reduction, variability of biologically functionless traits
arises. For the first time this was experimentally
described by him in detail for cave-living isopods of
the genus Asellus from the Dinarian karst. He showed
the high variability of the ocelli these species are con-
vergently developing in three different cave popula¬
tions—one of them later named A. kosswigi in honor of
his studies. He found that the crystalline cones may be
smaller, completely reduced, or with respect to their
number lower or higher compared to the surface sister
species A. aquaticus. These phenomena also occur on
the individual level, generating left—right asymmetry.
Similar variability was found for the eyes of other
arthropods such as the cave mysid Heteromysoides cotti,
which additionally even shows variability between the
ommatidia of the same eye.
Also in cavefish this variability was observed at the
individual and population levels as well as between
populations (Wilkens, 1988, 2001). For example, the
eye size of the phylogenetically young Micos cave
population of Astyanax shows half the range of an
F2-cross between surface and strongly reduced cave¬
fish (Wilkens, 1988). In the regressive vertebrate fish
eye, too, considerable differences in histology and size
can be observed between the left and the right side eye
rudiment of one single specimen (Wilkens, 2001).
Furthermore, in cases in which several cave popula¬
tions have convergently evolved in geographically
remote caves, as in Astyanax cavefish or in the Rhamdia
cave catfish, mean eye size varies among different
populations.
The process of reduction, as it is most conspicuously
becoming apparent by the variability developed, is
perpetuated and finally completed by the continuous
accumulation of deleterious mutations in the course of
time simply because these mutations are no longer
eliminated by selection. The neutral imitation theory
(NMT) provides the most parsimonious explanation of
regression (Wilkens, 2010). In accordance with the gen¬
erally unquestioned role of selection in constructive
evolution, the neutral mutation theory does not contra¬
dict the modern synthetic theory of evolution, but is
embedded in all its rules except for stabilizing selec¬
tion. It is by no means Lamarckism in disguise, as
repeatedly was speculated, the idea that an organism
can pass on characteristics that it acquired during its
lifetime to its offspring. Furthermore, the neutral muta¬
tion theory is until present under dispute because loss
of selection is implied to be responsible for troglobitic
evolution in general and it is not considered that a dif¬
ference is made in this theory between constructive
troglomorphic traits driven by selection and regressive
ones exclusively explained by the above-mentioned
theory.
As shown above, strong arguments corroborating
the NMT are provided by the reduction of a series of
behavioral traits that are no longer performed after the
lightless cave environment has been colonized. For
example, the optically orientated aggressive behavior
of Astyanax is getting reduced although not performed
by the surface fish in darkness. The same was
found for the dorsal light reaction, which helps the
surface fish to camouflage under light conditions.
Unexpectedly even negative scotophilia, a behavior
enabling troglophiles to avoid light, is lost. This behav¬
ior is no longer needed, because the hypogean realm is
usually so vast that there is no danger of the majority
of a cave population to be in contact with the lighted
surface environment.
The conditions of absolute darkness in caves can be
compared with phenomena arising in domesticated
animals, when loss of selection is provided by human
care. Breeders allow traits such as body color and
shape to no longer be subjected to selection and so var¬
iability will arise, by which they are enabled to select
for specific features. Similarly, several traits in humans
such as the eyes are subjected to a loss of selection,
too, because spectacles are used to neutralize the
developing variability caused by mutations responsible
for short- and wide-sighted phenotypes and many
other defects.
Summarizing, the neutral mutation theory and the
characteristic variability developed in functionless
regressive traits in cave animals would be one of
ENCYCLOPEDIA OF CAVES
NIPHARGUS: A MODEL SYSTEM FOR EVOLUTION AND ECOLOGY
555
the rare cases in which random mutations are not
eliminated by natural selection acting to preserve the
functional capability of a trait. Loss of selection allows
random mutations as a basic evolutionary principle to
manifest in the phenotype.
See Also the Following Articles
Adaptive Shifts
Astyanax mexicanus: A Model System for Evolution
and Adaptation
Natural Selection
Bibliography
Culver, D. C., & Wilkens, H. (2000). Critical review of the relevant
theories of the evolution of subterranean animals. In H. Wilkens,
D. C. Culver & W. F. Humphreys (Eds.), Ecosystems of the world:
Subterranean ecosystems (pp. 381—389). Amsterdam: Elsevier.
Darwin, C. (1859). On the origin of species (Reprint). New York:
Gramercy Books.
Gross, J. B., Borowsky, R., & Tabin, C. J. (2009). A novel role for
Mclr in the parallel evolution of depigmentation in independent
populations of the cave fish, Astyanax mexicanus. PLoS Genetics,
5(1), el000326.
Jeffery, R. J. (2001). Cavefish as a model system in evolutionary
developmental biology. In Developmental biology (pp. 1 — 12).
New York: Academic Press.
Kimura, M. (1983). The neutral theory of evolution. Cambridge, U.K.:
Cambridge University Press.
Kosswig, C., & Kosswig, L. (1940). Die Variability bei Asellus aquati-
cus, unter besonderer Beriicksichtigung der Variability in isolier-
ten unter- und oberirdischen Populationen. Revue de la Paculte des
Sciences de VUniversite d'lnstanbul, 5, 1—55 (in German).
Langecker, T. G. (2000). The effects of continuous darkness on cave
ecology and cavernicolous evolution. In H. Wilkens, D. C. Culver
& W. F. Humphreys (Eds.), Ecosystems of the zvorld: Subterranean
ecosystems (pp. 135—158). Amsterdam: Elsevier.
Mitchell, R. W., Russell, W. H., & Elliott, W. R. (1977). Mexican
eyeless characin fishes, genus Astyanax: Environment, distribution
and evolution. Special Publications of the Museum of Texas
Technological University, 12. Lubbock, TX: Texas Tech Press.
Protas, M. E., Conrad, M., Gross, J. B., Tabin, C., & Borowsky, R.
(2007). Regressive evolution in the Mexican cave tetra, Astyanax
mexicanus. Current Biology, 17(5), 452—454.
Protas, M., Tabansky, I., Conrad, M., Gross, J. B., Vidal, O., & Tabin,
C. J. (2008). Multi-trait evolution in a cave fish, Astyanax mexica¬
nus. Evolutionary Developments, 10(2), 196—209.
Wilkens, H. (1988). Evolution and genetics of epigean and
cave Astyanax fasciatus (Characidae, Pisces). Support for the neu¬
tral mutation theory. In M. K. Hecht & B. Wallace (Eds.),
Evolutionary/ biology (pp. 271—367). New York: Plenum Publishing
Corporation.
Wilkens, H. (2001). Convergent adaptations to cave life in the
Rhamdia laticauda catfish group (Pimelodidae, Teleostei). In E. K.
Balon (Ed.), Environmental biology of fishes (pp. 251—261).
Dordrecht: Kluwer Academic Publishers.
Wilkens, H. (2010). Genes, modules and the evolution of cave fish.
Heredity, 105(5), 413-422.
Yamamoto, Y., Stock, D.W., & Jeffery, W.R. (2004). Hedgehog signal¬
ing controls eye degeneration in blind cavefish. Nature, 431,
844-847.
Yamamoto, Y., Byerly, M. S., Jackman, W. R., & Jeffery, W. R. (2009).
Pleiotropic functions of embryonic sonic hedgehog link jaw and
taste bud amplification with eye loss during cavefish evolution.
Developmental Biology, 330(1), 200—211.
NIPHA RQUS: A MODEL SYSTEM FOR
EVOLUTION AND ECOLOGY
Cene Fiser
University of Ljubljana, Slovenia
INTRODUCTION
Evolutionary ecology depends on model systems.
To understand evolutionary processes in subterranean
environments over a broad geographic scale, amphi-
pod genus Niphargus (family Niphargidae) seems to be
the most appropriate model for focused analyses. With
over 300 species it is the largest genus of freshwater
amphipod in the world. It is distributed across most of
Europe, with the exception of the Pyrenean Peninsula
and Northern Europe, and few species are known
from the Middle East (Karaman and Ruffo, 1986).
The estimated age of three fossilized specimens dates
between 30 and 50 Ma BP, and imply synchronous
evolution of the genus and geological formation of the
continent (e.g., Jazdzewski and Kupryjanowicz, 2010).
All species within the genus, including fossils, are
eyeless and without pigmentation.
Niphargus is a nightmare for taxonomists. This fact is
best illustrated by the five international meetings held
between 1969 and 1980 devoted to problematic taxonomy
of the genus. Indeed, intraspecific variability, often cou¬
pled with sexual dimorphism and allometric growth,
may be large compared to interspecific differences.
Diagnostic details can easily be overlooked. Furthermore,
samples counting a few morphologically unique indivi¬
duals often remain taxonomically unevaluated (e.g., Fiser
et al, 2009).
On the other hand, the high variability of Niphargus
makes the genus an attractive object for evolutionary
ecology if analyzed in spatial and temporal contexts.
Until recently, no relevant phylogenetic hypothesis was
available; therefore evolutionary hypothesis testing was
not possible. Comparative studies are in their early
phase and this article mainly reviews the potential of
Niphargus of being a model system. It consists of three
parts: (1) a review of the properties of Niphargus
Encyclopedia of Caves.
©2012 Published by Elsevier Inc.
556
N
considered to be of key importance for comparative
studies; (2) studies describing interactions within and
between species; (3) a reconsideration of some of these
elements within an ecological or evolutionary context.
PART ONE: KEY PROPERTIES
Morphological Variation
The overall morphological variation of Niphargus
exceeds the variation of most amphipod genera or
even families (Fiser et al., 2009). Summarized here are
only functionally important morphological characters,
or traits otherwise relevant for ecological research.
Morphological terminology is clarified in Figure 1 .
Body size is the most obvious difference between
species (Fig. 2). Body length spans between 2 mm and
35 mm.
Body shape varies from stout to slender, almost ver¬
miform. It determines hydrodynamic properties of an
amphipod body. Paired, ventrally extended coxal and
epimeral plates form a longitudinal ventral channel.
Pleopod action generates water currents, which flow
through the channel. The water brings oxygen, food
particles, and chemical cues. Currents can be used for
jet propulsion in stretched animals while in flexed ani¬
mals their sum equals zero. Body shape and strength
of self-producing currents therefore strongly depend
on shape of coxal plates and bases of pereopods V—VII
(Figs. 1 and 2).
Maxillae I are the most variable appendages in
the mouthpart apparatus. The number of spines on its
FIGURE 1 Functional model of an amphipod. Arrows indicate water
currents created by pleopod action. Abbreviations are al, all—antenna I
and II; gl, gll—gnathopod I and II; pV, pVI, pVII—pereopods V, VI, and
VH; bV, bVI, bVII—basis V, VI, and VII; ul, uHI—uropods I and IE;
cxI, cxll, cxIII, cxIV—coxal plates I, II, III, and IV; e—epimeral plate (only
the first one indicated).
outer lobes and density of denticles along these spines
perhaps indicate different modes of feeding (Fig. 3).
Gnathopods are the two anterior-most appendages
on a trunk. Two distal articles are modified for feeding
and grooming. Their variable size and shape suggest
differences in feeding biology (Figs. 2 and 3).
Antennae and pereopods are of different lengths.
Length of the first pair of antennae, for example,
measures only one third of the body length in some
species, but exceeds the body length by factor 1.3 in
others. Similarly variable in length are pereopods
V-VII (Fig. 2).
Sexual dimorphism is common in several species.
Differences between sexes include body size (males
larger), body shape (males slenderer), and elongation
of distal articles of male's uropods I and III (Fig. 4).
Allometric growth, that is, nonlinear change of pro¬
portions in two body measures during growth (Fig. 4), is
another well-known phenomenon in Niphargus. Details
for most of the species are not known; however, a list of
possible allometries has been made. It includes append¬
age lengths, shapes of coxal plates, and shapes of
gnathopods; allometric growth is a proximate mecha¬
nism of sexual dimorphism (Fiser et al., 2008a).
Developmental patterns were studied as temporal
sequences of post-embryonic developmental events,
that is, outgrowth of particular seta or sudden shifts
in growth (Fiser et al., 2008a; Fig. 4). To date, only
one analysis of developmental events was made on
a limited data set. It tentatively implies high inde¬
pendence among individual developmental events,
meaning that selection can act on every individual
trait without affecting other morphological traits (low
pleiotropy).
Physiological Tolerance to Abiotic Factors
in Relation to Distributional Data
Niphargus is found in virtually all types of subterra¬
nean waters or waters connected with subterranean
environments such as epikarst, hypothelminorheic,
interstitial, sinking rivers, phreatic habitats, and
springs. The overall eurytopic nature of Niphargus can
be inferred from laboratory and field data. In most
cases it is not clear whether individual species tolerate
or select for some abiotic conditions. However, on a
large scale, when compared to surface crustaceans,
niphargids appear to be more tolerant to extreme
values for most physical and chemical parameters.
This tolerance apparently varies between Niphargus
species. Here are reviewed known interactions
between abiotic parameters and Niphargus species,
while consequences for local biodiversity are discussed
in the last section.
ENCYCLOPEDIA OF CAVES
N1PHARGUS: A MODEL SYSTEM FOR EVOLUTION AND ECOLOGY
557
LA
3
oo
T3
ID
n
id'
l/l
O
3
3
o
(D
oo
i/i
QJ
“0
to
fD
oo
TJ
id
n
id'
1/1
C
T3
r-h
o
UJ
cn
3
3
FIGURE 2 Morphological diversity of Niphargus: size, shape, and proportions. Species are not drawn to scale: size proportions are indicated
by scale bars. Note that size minimally contributes to diversity of interstitial species, which is in stark contrast to the cave species.
Light elicits a stress response in some species, as
inferred from measurements of the respiratory electron
system and oxygen consumption (Simcic and Brancelj,
2007). Contrary to laboratory evidence, many species
are found in springs. Some of them presumably
migrate back to darker regions at dawn, having spent
the night feeding at springs (Mathieu and Turquin,
1992), but some species constitute permanent popula¬
tions several hundred meters away from resurgences
on the surface (Fiser et al., 2006). It is unclear to what
extent the latter tolerate and to what extent they avoid
light by usage of dark microrefugees.
Resistance to starvation depends on high fat contents
which are 3% and 6% of wet weight in N. rhenorhodanesis
(hyporheic, more food) and N. virei (cave, less food),
respectively. Both niphargid species can live without
food for more than 200 days. For comparison, the surface
amphipod Gammarus fossarum and the isopod Asellus
aquaticus do not survive a starvation period of one month
(Hiippop, 2000). Similarly, relying on measurements of
the respiratory electron system and oxygen consump¬
tion, two other species, N. stygius and N. krameri, seem to
be extremely resistant to starvation compared to surface
amphipods (Simcic and Brancelj, 2007). Proximate causes
for resistance to starvation include different enzymatic
activity and energy-saving behavior such as decreased
locomotory and respiratory activity (Hiippop, 2000). At
least N. rhenorhodanesis can aestivate when water dries
up (Mathieu and Turquin, 1992).
Response to hypoxia was studied in N. virei and
N. rhenorhodanesis comparatively to surface amphipods
(Gammarus fossarum ) or isopods ( Asellus aquaticus).
Both hypogean species resist hypoxia significantly
longer than surface species. Response to hypoxia
ENCYCLOPEDIA OF CAVES
558
FIGURE 3 Diversity of feeding structures. (A) Maxilla I varies mainly in number and denticulation of spines on the outer lobes (upper
part); these spines are in some species greatly modified, such as in Niphargobates orophobnta (lower part). (B) Gnathopods are used for feeding
and grooming. The shape and size of gnathopods I and II greatly vary.
(A) (B)
FIGURE 4 Allometry and sexually dimorphic
traits. (A) Growth of sexually dimorphic uropod
III; nonlinear growth (black lines) evidently better
describes growth of the article than linear growth
(gray line) ( after Fiser et al., 2 008a). (B) Sexually
dimorphic traits include lengths of inner ramus of
uropod I, distal article of uropod III, and inner
ramus of uropod III (arrows).
seems to be mediated by different enzyme activities
and immediate decrease in locomotory and respiratory
activities. Although N. virei seems to have lower oxy¬
gen consumption and overall lower locomotory and
respiratory activities than N. rhenorhodanesis, the latter
in post-hypoxic recovery more efficiently assimilates
metabolites accumulated during hypoxia (Hiippop,
2000). Field evidence implies that at least N. sphagnico-
lus survives in hypoxic waters.
Temperatures in subterranean environments are
presumably stabile and subterranean species are pre¬
sumably stenothermal. Many data support this view.
Niphargus virei is intolerant to cold temperatures in
the laboratory, which agrees with its distribution
that is limited to presumable borderlines of Alpine
Pleistocene glaciers (Folquier et al., 2008). Furthermore,
N. thermalis has been reported only from thermal
waters. In a cold-adapted mountain complex N. tatrensis,
ENCYCLOPEDIA OF CAVES
NIPHARGUS: A MODEL SYSTEM FOR EVOLUTION AND ECOLOGY
559
post-Pleistocene warming presumably caused extinction
of lowland populations and fragmentation of ancestral
populations. By contrast, N. rhenorhodanesis shows
remarkable tolerance to low temperatures in the labora¬
tory. Unexpectedly and difficult to explain, the tolerance
for cold in this species varies on population level
(Colson-Proch et al., 2009).
Tolerance to water chemistry is less well known.
Measurements of the respiratory electron system and
oxygen consumption imply high tolerance for acid
environment in N. sphagnicolus (Simcic and Brancelj,
2007), which agrees with its distributional data (Fiser
et al., 2010). Quite a few species live in anchihaline
water, where coexisting species apparently distribute
along a salinity gradient. Some species survive in
mildly to highly sulfide waters and some do not
(Flot et al., 2010). Moreover, few species were found
exclusively in toxic hydrogen sulfide-rich aquifers,
although their preference for sulfide remains unstudied
(Dattagupta et al., 2009). Water conductivity may be
related to water hardiness or to pollution. Niphargus
sphagnicolus, for example, was found in a pond where
conductivity increased because of pollution. By contrast,
allopatric distribution of four species in the western
Dinarics corresponds to water conductivity and water
hardiness though causal links between physiology and
distribution need to be addressed in the laboratory (Fiser
et al., 2006). Last but not least, N. montellianus showed
increased resistance to high concentrations of heavy
metals compared to surface Gammarus species (Coppellotti
Krupa and Guidolin, 2003).
Endemism and Poor Migratory Abilities
Most of niphargid species are known only from few
localities, many of these only from their type localities.
The overall impression is that the degree of endemism
is high, but few species have been reported from large
areas. Individual distributional patterns may be biased
because of either incomplete sampling or vague taxon¬
omy coupled with misidentifications. Trontelj et al.,
(2009) addressed the second issue using molecular
phylogeny. Indeed several species names in fact cover
a number of species, which are not necessarily closely
related. Moreover, the authors proposed that the lon¬
gest diameter of a range only exceptionally exceeds
200 km. The range-size hypothesis was tested by
Zagmajster et al. (unpublished) using a dataset of 43
species that could have been unambiguously identified
and originated from locally well-explored areas. This
analysis indeed showed that most of the species have
small ranges, although species with large ranges do
exist. Three species were collected in a single spot,
about 60% of species had the maximum extent of
the range less than 100 km, and about 78% of the
species less than 200 km. Extent of range in nine spe¬
cies (21%) exceeded 200 km, reaching up to 600 km.
Therefore, it seems plausible to conclude that the
majority of species have poor migratory capabilities,
a finding that might importantly explain some evolu¬
tionary patterns.
PART TWO: INTERACTIONS WITHIN
AND BETWEENSPECIES^
Intraspecific Interactions: Cannibalism
A key intraspecific interaction seems to be cannibal¬
ism, which affects microdistribution and breeding biol¬
ogy. Observations of N. timavi in the cave laboratory
unambiguously confirmed that large individuals prey
on small conspecifics, even on their own offspring
(Lustrik et al., 2011). These observations, although
conducted in Petri dishes where juveniles cannot hide,
may indicate that niphargids are not selective pre¬
dators. This also means that juveniles would benefit if
they evolved avoidance behavior. Following this idea,
Lustrik et al. (2011) observed the distribution of juve¬
niles in aquaria filled with layers of differently sized
glass pebbles. Juveniles preferred layers of midsized
pebbles, but roamed also through large sized pebbles.
As adults were released to aquaria, juveniles withdrew
to the midsized layer, evidently inaccessible to adults.
Therefore, cannibalism probably best explains
results from field studies using artificial substrates.
Juveniles of N. rhenorhodanesis are abundant only in
the absence of larger (>7 mm) individuals, and rapidly
disappear when larger individuals colonize the space
(Mathieu and Turquin, 1992 and references therein).
Similarly, in the field juvenile N. timavi invaded fine-
sand boxes while adults preferred gravel and leaves
(Lustrik et al., 2011).
Cannibalism in combination with water regime may
regulate breeding biology of species. Juveniles in both
cave N. virei and interstitial N. rhenorhodanesis hatch
when chances to be eaten by conspecifics are small. In
cave N. virei, juveniles hatch at the onset of floods,
when drifting accelerates dispersion of juveniles. In
interstitial N. rhenorhodanesis, hatching corresponds
to the onset of a dry period, when habitat partially
dries up and adult individuals aestivate (Mathieu and
Turquin, 1992).
Interspecific Interactions: Predation
and Competition
Interactions between different species of Niphargus
are poorly understood. Perhaps the most informative
are results of transplant experiments, in which
ENCYCLOPEDIA OF CAVES
560
N
individuals of N. virei were introduced into a cave
populated by N. rhenorhodanesis. The introduced spe¬
cies exterminated the native one in less than 20 years,
implying that competition and predation are important
interactions that determine local and regional niphar-
gid diversity (after Folquier et al., 2008). On the other
hand, niphargid species frequently coexist, and con¬
stitute communities of up to nine species (Table 1).
A prerequisite for temporal stable communities is
minimization of harmful interactions (competition and
predation). Niche separation can be expected. A few
cases are reviewed in the final section.
Although various subterranean amphipods may
coexist with Niphargus species, neither interactions
nor niche separation was studied at this taxonomic
level. More attention was paid to interactions be¬
tween subterranean Niphargus and surface amphipods.
Observations in the Postojna-Planina Cave System in
Slovenia implied that competitive strength of surface
species increases when food is not a limiting factor.
This was extended to the hypothesis that competition
influences distribution of troglobionts and prevents
their spreading into surface waters (Sket, 1981). Recent
work partially supports this hypothesis. Annual obser¬
vations of selected sites in spring areas revealed high
tempor-spatial variation in the microdistribution of
G. fossarum and N. timavi. Boxes with artificial sub¬
strates showed that the two species differ in habitat
preferences. While G. fossarum preferred decaying
leaves, N. timavi invaded leaves and gravel in similar
shares. Laboratory observations suggest that the
spread of N. timavi into surface waters is restricted
by a combination of predation and cannibalism, but
not by interspecific competition (Lustrik et al., 2011,
and references therein).
TABLE 1 Structure of Niphargus Community in Vjetrenica and
the Postojna-Planina Cave System
Vjetrenica
Habitat
PPCS
N. boskovici
Niphargobates orophobata
N. factor
epikarst
N. cvijici
N. wolfi
-
cave streams
N. stygius
N. spoeckeri
N. zavalanus
cave interstital
N. dobati
N. balcanicus
N. vjetrenicensis
phratic, deep lakes
N. hercegovinensis
N. orcinus
N. trullipes
N. kolombatovici
Symbiosis with Chemoautotrophic Bacteria
Dattagupta et al. (2009) report on the discovery of
abundant filamentous bacteria on the exoskeletons
of Niphargus species within the sulfide-rich Frasassi
cave complex in central Italy. At least two out of four
Niphargus species living in that complex are infected
with filamentous bacteria. Using 16S rDNA sequenc¬
ing and fluorescence in situ hybridization (FISH), they
showed that amphipods throughout the cave complex
are colonized by a single type of bacteria in the
sulfur-oxidizing clade Thiothrix. The epibiont phylo-
type is distinct from Thiothrix phylotypes that form
conspicuous biofilms in the cave streams and pools
inhabited by niphargids. Using a combination of 13C
labeling, FISH, and secondary ion mass spectrometry,
they show that the epibiotic Thiothrix are autotrophic,
establishing the first known example of a nonmarine
chemoautotroph-animal symbiosis.
PART THREE: COMPARATIVE
REVIEW OF KEY ELEMENTS
Phylogeny Reveals Morphological
Diversity and Stasis
Introduction of molecular data largely modified our
view of the evolution of morphology of Niphargus.
Allozyme studies and sequencing of mitochondrial
COI and 12S, and nuclear 28S and ITS DNA (Trontelj
et al., 2009; Fiser et al., 2008b; Flot et al., 2010) yielded
several important findings.
First, all members of the family Niphargidae diver¬
sified from the common ancestor Niphargopsis, and
at least two other niphargid genera ( Niphargobates,
Pontoniphargus) are nested within Niphargus, losing
justification for their separate generic status (Fiser
et al., 2008b; Flot pers. comm; unpublished data).
Second, several cryptic lineages, which may be
tentatively equated with cryptic species, have been
discovered. Convergent selection and poor migratory
abilities are key factors that explain morphological simi¬
larity among some niphargid species. Morphological
similarity may be the result of fragmentation of ancestral
species range followed by speciation in the absence of
morphological change. Alternatively, strong selection
operating on entire suits of morphological characters
in unrelated species results in morphologically indistin¬
guishable species (Trontelj et al, 2009).
Third, closely related species may be morphologically
unexpectedly different (Fig. 5). A tentative conclusion
is that not only convergent but also divergent selection
operates in subterranean environments (Fiser et al,
2008b).
ENCYCLOPEDIA OF CAVES
NIPHARGUS: A MODEL SYSTEM FOR EVOLUTION AND ECOLOGY
561
N.vinodolensis
N.hadzii
N.puteanus spoeckeri *
WMMWUMrf g — inA
lius p
Iius novemestanus f gabrovceci
lius novomestanus
N. orcinus longi
N. croaticus pachytelson *
i oo r N. steueri subtypicus
' N.dolichopus
N. tridentin us
N.lessiniensis
o 83j- N. pupetta syn. Karamaniella
0 sP—
0 51 I— N. kochianus labacensis *
j- N.strouhali alpinus
P- N kochianus longidactylus *
lerrans _
bajuvancus grandii
•novici bajuvaricus *
AJaouJgxctogaj/
N. kochianus wolfi *
N. tauri carniolicus *
\l. tauri cf koblars ka
orcinus virei *
N.delamarei
N.laisi
N.scopicauda n.n.
N. tatrensis tatrensis
N. aquilex schellenbergi *
i do r N.kochianus kochianus
1 kochianus dimorphopus
N jovanovici kieferi *
0.1
Bogidiella semidenticulata
L Bogidiella albertimagni
i oo - •» balcanicus syn. Stygodites
N.polymorphus
orcinus vjetrenicensis *
N. fact or
lunaris
N.hvarensis cf.mlini
N.hvarensis
N.boskovici
N. puteanus krameri *
Istygius karamani
o 98 - N. orcinus orcinus
°- 7 A >— N. krameri f. spinulifemur *
■ N. krameri timavi
■ Niphargopsis caspary
0 91 r N.puteanus
i N.stygius kenki
1 -QQ- j- N sanctinaumi
N.lourensis
i.OO j N.dolenianensis
N.sphagnicolus
N. tatrensis hrabei *
us rhen(
— Ma uerbach
- N. stygius costozzae *
1.001-
— Urothoe brevicomis
■ Fuegophoxus abjectus
Synurella ambulans
— Crangonyx subterraneus
- Crangonyx forbesi
— Gammarus fossarum
Eriopisa elongata
FIGURE 5 Phylogenetic tree based on fragments of 12S and 28S DNA (after Fiser et al. : 2008b). Colors indicate one of the traditional divi¬
sions of the genus based on morphology. There is only little congruence between morphology and phylogenetic relationships.
ENCYCLOPEDIA OF CAVES
562
N
Communities: Niche Separation
and Functional Morphology
Up to nine Niphargus species may coexist in one
locality. Communities such as these are priceless
natural laboratories for evolutionary and ecological
research. As expected, coexisting species differ in their
ecological niches along chemical and physical gradi¬
ents. In addition, it appears that microhabitats defined
by availability of space and water currents harbor mor¬
phologically distinct species.
Hypothelminorheic (seep-dwelling) N. sphagnicolus
and N. slovenicus live in sympatry. More careful examina¬
tion of distributional maps reveals that the two species
may be collected in close proximity, but never together
(Fig. 6). It appeared that localities invaded by either of
the species differ in oxygen saturation, conductivity, and
pH. pH is a key factor that determines which species
prosper in a particular locality (Fiser et ah, 2010).
Both chemical and physical gradients may be
involved in niche separation, as is the case in the sulfide-
rich Frasassi cave complex in central Italy. N. ictus
and N. frasassianus live in sulfidic water. In localities
where both species live together they occupy different
microhabitats. Niphargus frasassianus is restricted to fast¬
flowing portions of the stream, whereas N. ictus inhabits
stagnant parts that are deeper than 20 cm. Two other
species seem to be entrapped in nonsulfidic lakes in dif¬
ferent parts of the cave complex. Niphargus montanarius
was collected from the nonsulfidic lake II Bugianardo.
A fourth, yet undescribed, species was found in a remote
locality of Lago Primo, a stratified lake with oxygenated
water on top and reducing sulfidic waters at depths
greater than 3.5 meters (Flot et ah, 2010).
Most caves have less distinct, if any, chemical gradi¬
ents. Nevertheless, presence of water currents and
availability of space define at least four physically
distinct microhabitats: epikarst, interstitial, streams,
and lakes. Niche separation, at least in part, includes
occupancy of different microhabitats. For instance,
in the Postojna-Planina Cave System Niphargobates
orophobata and N. zvolfi were found in drips, N. stygius
and N. spoeckeri in sinking rivers and ponds along
shores of the river, N. dobati in cave interstitial, and
N. orcinus in phreatic. Moreover, a comprehensive
analysis of seven cave communities identified four dis¬
tinct morphological types (unpublished). Phreatic spe¬
cies are of two size classes with elongated appendages.
Species from streams are large with short appendages.
Interstitial species cannot be distinguished from epi-
karstic species. Both are small and of various lengths
(Fig. 2). In any case, the communities provide some
evidence for divergent selection.
A Case of Adaptation: Female Body Size
Whereas small pore size in interstitial and epikarst
habitats may explain the evolution of small species,
the question remains: What is the adaptive value of
large body size in food-poor phreatic waters? Phreatic
niphargids are not sexually dimorphic, and ovigerous
females (females carrying eggs) often exceed the
size of 20 mm. By contrast, species from streams and
springs are sexually dimorphic, where females rarely
reach 18 mm in body length (males can reach 30 mm).
In a recent study, egg volume, brood volume, and
reproductive investment (i.e., what proportion of body
volume represents brood volume) were estimated
for ovigerous females of seven phreatic species and
six species living in springs. Large phreatic females
have larger eggs but lower reproductive investment
into a single brood compared to species from springs.
Moreover, the correlations between body size and egg
volume and body size and reproductive investment
are significantly positive and negative, respectively.
Large body size therefore presumably increases both
female and offspring fitness. Egg size in general posi¬
tively correlates with juvenile size, and larger juveniles
presumably more easily cope with environmental chal¬
lenges, such as avoidance from predators, increased
resistance to starvation, and exploitation of a broader
range of food. Since phreatic species have a prolonged
life span (Mathieu and Turquin, 1992), it is likely that
they have increased numbers of broods. Fitness of the
female—the product of number of broods and average
number of survived offspring per brood—therefore
depends on her own survival, which modifies her
allocation strategy. Reduction of reproductive effort
into one brood—a phenomenon common to species
with multiple broods—increases her chances to sur¬
vive until the next brood. Increase in body size also
affects other traits. Larger eggs, for instance, require
larger marsupium for two reasons. Not only that more
space is needed, but that larger eggs may require
increased ventilation due to decreased surface-to-
volume ratio. This explains the relatively deeper coxal
plates in phreatic species compared to species from
springs.
CONCLUSIONS
This short overview illustrates that the genus
Niphargus is extremely variable considering morphologi¬
cal, developmental, physiological, and life-history traits.
This variability, when analyzed in temporal and spatial
contexts, may be a precious source of information about
ENCYCLOPEDIA OF CAVES
NIPHARGUS: A MODEL SYSTEM FOR EVOLUTION AND ECOLOGY
563
- 1,00 0,00 1,00 2,00
regression factor score 1
Niphargus slovenicus Niphargus sphagnicolus
FIGURE 6 Niche separation of two hypothelminorheic species (red circles: N. sphagnicolus; blue diamonds: N. slovenicus). (A) Distribution
of the two species reveals no geographic segregation; white color indicates alluvial plain at altitudes between 290 and 390 m, and most sam¬
ples were collected below 500 m altitude (after Fiser et al., 2010). (B) Principal component analysis of 11 functional characters shows that the
two studied species cannot be unambiguously discriminated (the first two principal components explain 98.7% of variance); this may indicate
that hydrological conditions in hypothelminorheic species do not allow morphological differentiation of the species. (C) Evidence that distri¬
bution of the two species depends on acidity of water. Note that N. sphagnicolus tolerates a large range of pH, possibly indicating that this spe¬
cies frequently lives in suboptimal acidic conditions.
ENCYCLOPEDIA OF CAVES
564
N
processes such as speciation, specialization, adaptation,
competition, and migration on different geographic
scales (from single locality to Western Palearctics) and
different taxonomic levels (from population up to a
family).
See Also the Following Articles
Asellus aquaticus: A Model System for Historical
Biogeography
Gammarus minus: A Model System for the Study of
Adaptation to the Cave Environment
Invasion, Active versus Passive
Vicariance and Dispersalist Biogeography
Bibliography
Colson-Proch, C., Renault, D., Gravot, A., Douday, C. J., & Hervant, F.
(2009). Do current environmental conditions explain physiological
and metabolic responses of subterranean crustaceans to cold? The
Journal of Experimental Biology, 212(12), 1859—1868.
Coppellotti Krupa, O., & Guidolin, L. (2003). Responses of Niphargus
montellianus and Gammarus balcanicus (Crustacea, Amphipoda)
from karst waters to heavy metal exposure. Journal de Physique IV,
107, 323-326.
Dattagupta, S., Schaperdoth, I., Montanari, A., Mariani, S., Kita, N.,
Walley, J. W., et al. (2009). A novel symbiosis between chemoau-
totrophic bacteria and a freshwater cave amphipod. The ISME
Journal, 3(8), 935-943.
Fiser, C., Bininda-Emonds, O. P. R., Blejec, A., & Sket, B. (2008a). Can
heterochrony help explain the high morphological diversity
within the genus Niphargus (Crustacea: Amphipoda)? Organisms,
Diversity and Evolution, 8(2), 146—162.
Fiser, C., Konec, M., Kobe, Z., Osanic, M., Gruden, P., & Potocnik, H.
(2010). Conservation problems with hypothelminorheic Niphargtis
species (Amphipoda: Niphargidae). Aquatic Conservation: Marine
and Freshwater Ecosystems, 20(5), 602—604.
Fiser, C., Sket, B., & Stoch, F. (2006). Distribution of four narrowly
endemic Niphargus species (Crustacea: Amphipoda) in the west¬
ern Dinaric region with description of a new species. Zoologischer
Anzeiger, 245(2), 77—94.
Fiser, C., Sket, B., & Trontelj, P. (2008b). A phylogenetic perspective
on 160 years of troubled taxonomy of Niphargus (Crustacea:
Amphipoda). Zoologica Scripta, 37(6), 665—680.
Fiser, C., Trontelj, P., Lustrik, R., & Sket, B. (2009). Toward a unified
taxonomy of Niphargus (Crustacea: Amphipoda): A review of
morphological variability. Zootaxa, 2061, 1—22.
Flot, J.-F., Woerheide, G., & Dattagupta, S. (2010). Unsuspected diver¬
sity of Niphargus amphipods in the chemoautotrophic cave eco¬
system of Frasassi, central Italy. BMC Evolutionary Biology, 10, 171.
Folquier, A., Malard, F., Lefebure, T., Douady, C. J., & Gibert, J.
(2008). The imprint of Quaternary glaciers on the present-day dis¬
tribution of the obligate groundwater amphipod Niphargus virei
(Niphargidae). Journal of Biogeography, 35(3), 552—564.
Hiippop, K. (2000). How do cave animals cope with the food scarcity
in caves? In H. Wilkens, D. C. Culver & W. F. Humphreys (Eds.),
Subterranean ecosystems (pp. 159—188). Amsterdam: Elsevier.
Jazdzewski, K., & Kupryjanowicz, J. (2010). One more fossil
Niphargid (Malacostraca: Amphipoda) from Baltic amber. Journal
of Crustaceana Biology, 30(3), 413—416.
Karaman, G. S., & Ruffo, S. (1986). Amphipoda: Niphargus —Group
(Niphargidae sensu Bousfield, 1982). In L. Botosaneau (Ed.),
Stygofauna mundi (pp. 514—534). Leiden: E. J. Brill/Dr. Backhuys.
Lustrik, R., Turjak, M., Kralj-Fiser, S., & Fiser, C. (2011). Coexistence
of surface and cave amphipods in an ecotone environment
(spring area). Contributions to Zoology, 80(2), 133—141.
Mathieu, J., & Turquin, M. J. (1992). Biological processes at the
population level. II. Aquatic populations Niphargus (stygobiont
amphipod) case. In A. I. Camacho (Ed.), The natural history of
biospeleology (pp. 264—293). Madrid: Monografias Museo Nacional
de Ciencas Naturales.
Simcic, T., & Brancelj, A. (2007). The effect of light on oxygen
consumption in two amphipod crustaceans—The hypogean
Niphargus stygius and the epigean Gammarus fossarum. Marine and
Freshwater Behaviour and Physiology, 40, 141—150.
Sket, B. (1981). Distribution, ecological character, and phylogenetic
importance of Niphargus valachicus. Bioloski Vestnik, 29(1), 87—103.
Trontelj, P., Douady, C., Fiser, C., Gibert, J., Goricki, S., Lefebure, T.,
et al. (2009). A molecular test for hidden biodiversity in ground-
water: How large are the ranges of macro-stygobionts? Freshwater
Biology, 54(4), 727- 744.
NITRATE CONTAMINATION
IN KARST GROUNDWATER
Brian Q. Katz
U.S. Geological Survey, Tallahassee, Florida
BACKGROUND INFORMATION
In karst terrain, as well as in other hydrogeologic
systems, numerous interrelated factors account for
nitrate contamination of groundwater. These factors
can be grouped into two main categories: land use
(including nitrate sources, land application rates and
timing of fertilizers and manure, and waste disposal
practices) and hydrogeologic characteristics (such
as the degree of connectivity between the aquifer
and surface solution features, soil drainage charac¬
teristics, the timing and amount of recharge, degree
of aquifer confinement, shallow and deep flow
patterns, and type of water-supply well). Other con¬
ditions that further promote nitrate contamination in
karst aquifer systems are the typically oxygenated
conditions that contribute to the stability of nitrate
(preventing its transformation to reduced nitrogen
species), thin soil cover overlying many karst aqui¬
fers, and often careless waste-disposal practices.
Also, in many areas, an epikarst zone directly above
the aquifer can provide temporary storage for nitrate
and other contaminants, which can be released from
this zone into the aquifer by recharge from storm
events. Losses or decreases in nitrate concentrations
within a karst aquifer system can result from natu¬
rally occurring processes, such as denitrification
(microbially mediated transformation of nitrate to
Encyclopedia of Caves.
©2012 Published by Elsevier Inc.
NITRATE CONTAMINATION IN KARST GROUNDWATER
565
reduced nitrogen species) and dilution (e.g., result¬
ing from interactions between surface water and
groundwater).
ENVIRONMENTAL CONCERNS
AND NITROGEN CYCLING
IN GROUNDWATER
Several human and ecological health concerns are
associated with elevated nitrate concentrations in
groundwater. Infants under 6 months of age who ingest
nitrate in drinking water are susceptible to methemoglo¬
binemia, which can lead to reduced blood oxygen levels
and can result in death. For these health concerns, the
U.S. Environmental Protection Agency established a
maximum contaminant level (MCL) for nitrate of 10 mg
L _1 , as nitrogen (N), for drinking water. Another study
found an increased risk of non-Hodgkins lymphoma
associated with nitrate concentrations of 4 mg L” 1 or
more in rural drinking-water supplies. Groundwater
contributes to base flow in many streams and rivers
and elevated nitrate concentrations from groundwater
can stimulate the growth of nuisance aquatic vegetation
in surface waters. Currently, there are no guidelines for
nitrate concentrations in streams to limit the growth of
nuisance aquatic vegetation.
Nitrate is one of several species of nitrogen (N) that
can exist in the subsurface. The presence of other
N species in a karst aquifer depends on oxidation-
reduction conditions and microbial reactions in the
subsurface. N-cycling processes are complex. For
example, oxidation of reduced N species (ammonium,
organic N, nitrite) to nitrate or transformation of
nitrate to reduced N species generally takes place in
the soil or unsaturated zone where organic material
and microbial communities are present. These pro¬
cesses are controlled by the presence or absence
of oxygen and other terminal electron acceptors (e.g.,
sulfate, ferric iron, organic carbon). These processes
need to be understood to properly assess the fate of
nitrate in karst aquifers. It is beyond the scope of this
chapter to discuss these processes in detail. Numerous
reviews, texts, and other publications have described
biologically mediated transformations, storage pools,
and nitrate removal.
LAND-USE FACTORS
Activities associated with various land uses can
substantially affect nitrate concentrations in ground-
water in karst systems. Most cases of elevated nitrate
concentrations in karst aquifers are related to agricul¬
tural activities. Numerous studies have demonstrated
that nitrate concentrations in groundwater are higher
beneath agricultural areas compared to forested areas
and are related to the percentage of agricultural land
used near wells or springs.
Nonpoint or diffuse sources account for nitrate
contamination of groundwater on a regional scale in
many agricultural areas throughout the world. Elevated
nitrate concentrations in many of these areas often are
associated with contamination from other agricultural
chemicals, such as herbicides and other pesticides.
Although several retrospective studies have reported
lower nitrate concentrations in groundwater from the
southeastern United States compared to other regions,
elevated nitrate concentrations exceed the nitrate MCL
for drinking water in many parts of the karstic Upper
Floridan aquifer in areas of extensive agricultural land
use. The application of manure also has contributed
to nitrate contamination of groundwater as indicated
by statistically significant relations between nitrate con¬
centration in groundwater and the amount of manure
applied to soils overlying carbonate aquifers.
Nitrate contamination of groundwater also can
originate from localized agricultural point sources,
such as concentrated animal feeding areas, poorly
designed waste storage areas, and disposal of dead
animals in sinkholes. Also, other documented cases
of nitrate contamination of groundwater are related to
the spreading of organic wastes over agricultural
lands, irrigation of pasture land with cheese factory
effluent, and spray irrigation of wastewater from a
vegetable- and meat-canning facility.
There have been several documented cases of nitrate
contamination of groundwater in karst areas that are
associated with land uses other than agriculture. In rural
areas, contamination from septic tanks and cesspools is
fairly common. Effluent from these systems can reach
the water table after undergoing little filtration or attenu¬
ation due to the typically thin soil cover in many karst
areas. Also, nitrate contamination of groundwater has
resulted from diversion of untreated sewage into nearby
sinkholes or swallow holes. Leaky sewers have contrib¬
uted to nitrate contamination in some urban areas. In
many rural landscapes, the dumping of household
refuse, construction materials, and dead livestock has
resulted in groundwater contamination problems.
Landfills and open dumps in limestone outcrops in the
recharge area of a well field have contributed to nitrate
contamination of groundwater in China. In an Indiana
karst system, nitrate contamination of groundwater was
attributed to by-products from ordnance manufacturing
and demolition processes at a military institution.
Atmospheric deposition (wetfall and dryfall) is
potentially another source for elevated nitrate con¬
centrations in karst groundwater, although very few
studies have addressed this issue. Seasonal variability
ENCYCLOPEDIA OF CAVES
566
N
of nitrogen (ammonium and nitrate) in precipitation
most likely is related to the application of fertilizers in
the spring, but other potential atmospheric sources of
nitrate, including industrial waste gases and fossil-fuel
combustion, could be responsible. The influence of
atmospheric sources has been inferred from isotopically
enriched values of fc ls O of nitrate found in winter
and summer spring-water samples, when nitrate in
groundwater was relatively low and its concentration
in precipitation was high.
Most studies have focused on the collection of water
from wells or springs in karst aquifers; however, a few
studies have collected samples of cave drip water as a
novel way of characterizing the chemistry of recharge
water moving through the unsaturated zone into caves
and groundwater. Caves near and below agricultural
land had higher nitrate concentrations in drip water
than did groundwater in forested areas. Concentrations
of nitrate were about an order of magnitude higher in
autogenic recharge waters at a cave located beneath
an agriculturally intensive landscape than in recharge
at caves beneath less-intensive land uses. This study
also found that nitrate varied seasonally with a flush
of nutrients during late summer and autumn that corre¬
sponded to trends in surface rivers and streams.
HYDROGEOLOGIC FACTORS
Based on an analysis of retrospective data (1972—92)
from sites around the United States, the U.S. Geological
Survey (USGS) reported that nitrate concentrations were
highest in karst bedrock areas compared to other hydro¬
geologic settings, such as sand and gravel, alluvium, and
other bedrock types. Differences reflected the intensive
agricultural activity on land overlying the carbonate
settings, but also were related to the hydraulic connec¬
tions to the surface and solution features typical of frac¬
tured carbonate rocks.
Similar findings were reported for other karst areas in
the midwestern United States. Based on data from more
than 6000 water samples collected during 1977—80 from
22 counties in Iowa, nitrate-N concentrations from all
well depths were higher in a karst region compared to
samples from shallow and deep bedrock. Even though
nitrate concentrations decreased with depth in all
regions, elevated nitrate concentrations in the karst
region extended to 30 m greater depth than in the con¬
fined deep bedrock region.
The degree of confinement of the aquifer can also affect
nitrate concentrations in groundwater. Several studies
have shown that nitrate concentrations in limestone aqui¬
fers are substantially higher in water from unconfined
aquifers than from confined aquifers. However, in some
areas, a rapid response of groundwater to elevated nitrate
concentrations after fertilizer applications were attributed
to the rapid movement of fertilizer-derived nitrate into
macropores.
The density of karst features, such as sinkholes,
is another factor that affects nitrate contamination of
groundwater. Based on water samples from seven
carbonate aquifers in the eastern United States, nitrate
concentrations were significantly higher in agricultural
areas with a moderate to high density of sinkholes
(1—25 sinkholes per 100 km 2 ) compared to areas with
low sinkhole density (<1 sinkhole per 100 km 2 ).
Nitrate concentrations in groundwater also are
related to the type and depth of a sampled well,
although very few studies have specifically addressed
this issue in karst areas. Based on more than 5600 sam¬
ples collected throughout the U.S. in both karst and
nonkarst settings, the USGS reported that median
nitrate-N concentrations ranged from 0.2 mg L~ in
public supply wells to 2.4 mg L for irrigation and
stock wells. Water samples from about 16% of the gen¬
erally shallow irrigation and stock wells exceeded the
nitrate MCL, compared to only about 1% of samples
from public supply wells, which are typically open to
much deeper parts of an aquifer.
NITRATE CONTAMINATION OF SPRINGS
Springs represent discharge points of groundwater
basins and are typically the dominant source of water
for surface streams in karst areas. Because springs inte¬
grate water temporally, spatially, and vertically from
an aquifer, they offer an effective way of assessing
sources and time scales of nitrate contamination of
groundwater in karst areas. In many areas, nitrate
concentrations in spring waters have increased sub¬
stantially over the past 40—50 years and in most cases,
the increasing trends were related to increased
amounts of fertilizer application. Nitrate concentra¬
tions in spring waters (and nitrate flux) are related to
the extent of cultivated zones at the surface.
In several studies, nitrate concentrations were found
to be higher in springs than in water from nearby wells.
For example, in the Ozarks, higher nitrate concentra¬
tions in springs were attributed to groundwater that
generally follows shallow flow paths and issues from
conduits and, therefore, is more susceptible to contami¬
nation from the surface. Elevated nitrate concentrations
are found in shallow systems, as well as in springs
draining deep groundwater systems (up to 100 m
depth) in areas of intensive water exploitation and fertil¬
izer use.
Stable nitrogen isotopes (6 15 N) in hydrologic studies
have been particularly effective in assessing the
sources and fate of nitrate. In studies of karst springs.
ENCYCLOPEDIA OF CAVES
NITRATE CONTAMINATION IN KARST GROUNDWATER
567
these and other naturally occurring tracers have been
used effectively to test hypotheses about flow-system
characteristics and dominant geochemical processes.
For example, based on measurements of 6 ls O and 6 15 N
of nitrate in spring waters in the southwest Illinois
sinkhole plain, significant denitrification occurred
within the soil zone, epikarst, and shallow karst hydro¬
geologic system and accounted for the loss of nitrate
in spring waters from spring to winter. A multitracer
approach that included isotopic and other chemical
tracers was used to show that nitrate concentrations
in Florida spring waters were related to average
groundwater residence times determined from age¬
dating techniques. Springs with lower flow rates had
short residence times (around 10 years) and relatively
high nitrate-N concentrations (20—35 mg L -1 ), whereas
those with higher flow rates had longer residence times
(20—30 years) and lower nitrate concentrations (0.2 to
5 mg L '). These differences were attributed to the
higher contribution of young water (recharged less than
7 years) from shallow parts of the Upper Floridan aqui¬
fer in spring waters with low flows (0.028—0.28 m 3 s -1 ),
compared to a higher contribution of deep older water
(recharged 2—3 decades ago) that accounted for lower
nitrate concentrations in spring waters with high flow
rates (>2.8 m 3 s -1 ).
TEMPORAL VARIABILITY OL NITRATE
Temporal variations in nitrate concentrations in
groundwater and springs reflect the interrelation of
several important hydrogeologic, geochemical, land use,
and climatic factors. The degree of temporal variation is
related to the degree of karstification, mantle thickness,
and the proportion of conduit to diffuse flow. For exam¬
ple, nitrate concentrations in groundwater generally
increase following seasonal applications of fertilizers
and subsequent recharge of the groundwater system.
Large fluctuations in nitrate concentrations in springs
have been observed during high discharge events
related to annual variations in rainfall. In some areas,
seasonally high water tables may result in groundwater
that is anoxic and if conditions persist, loss of nitrate
through denitrification may occur (see below).
PROCESSES THAT RESULT IN LOSSES
OF NITRATE IN KARST
GROUNDWATER
Although many karst aquifers tend to be oxic,
nitrification-denitrification processes can occur within
the overlying soil zone, epikarst, and aquifer (rock)
matrix. In the relatively immobile pore waters in the
rock matrix of the Chalk aquifer in Britain, denitrifica¬
tion accounted for the decrease in nitrate concentra¬
tions downgradient in the direction of flow. Other
studies have documented nitrate losses in karst
groundwater as a result of denitrification.
Dilution of nitrate can result from mixing of surface
water and groundwater during following periods of high
rainfall in karstic systems. Dilution by runoff transported
through fast-flow conduits also has accounted for large
decreases in nitrate concentrations in spring waters dur¬
ing high-rainfall periods. In a study in Iowa, wetter than
average conditions following a drought period resulted in
a substantial increase in nitrate concentrations as a result
of the increased water volume moving through the soils
and the leaching of nitrate left over from the drought.
PROTECTING GROUNDWATER
RESOURCES
Understanding the factors and processes that affect
the occurrence, movement, and fate of nitrate is essen¬
tial for protecting groundwater quality and preventing
further nitrate contamination in karst areas. Although
the delineation of groundwater contributing areas and
(or) protection zones for springs and wells in karst
areas remains difficult and challenging, recharge areas
are protected in many locales—especially springs or
wells used for water supply. Considerable work is
being done to evaluate management strategies for
reducing nitrate contamination from point and non¬
point sources in karst systems. The time lag between a
reduction in land application of fertilizer or manure
spreading can vary greatly, depending on site-specific
soil characteristics, climatic conditions, and recharge
rates. Given the large year-to-year variations in nitrate
concentrations at Big Spring, in Iowa, it has been
difficult to identify any improvements in water quality
resulting from reductions in nitrogen that have
occurred incrementally over the past decade. In about
one-third of the studied basin, decreases in nitrogen
inputs have been overshadowed by large variations in
annual recharge. As the worldwide demand for water
increases, much work remains to be done to character¬
ize both the residence time of groundwater in varied
and complex karst systems, and the travel time from a
recharge area to a point of spring discharge or with¬
drawal of groundwater for public consumption.
Bibliography
DiGnazio, F. J., Krothe, N. C., Baedke, S. J., & Spalding, R. F. (1998).
6 15 N of nitrate derived from explosive sources in a karst aquifer
beneath the Ammunition Burning Ground, Crane Naval Surface
Warfare Center, Indiana, U.S.A. journal of Hydrology, 206, 164—175.
ENCYCLOPEDIA OF CAVES
568
N
Drew, D., & Hotzl, H. (Eds.), (1999). Karst hydrogeology and human
activities: Impacts, consequences, and implications. International contri¬
butions to hydrogeology. Rotterdam, Netherlands: A. A. Balkema.
Dubrovsky, N. M., Burow, K. R., Clark, G. M., Gronberg, J. M.,
Hamilton, P. A., Hitt, K. J., et al. (2010). The quality of our
nation's waters—Nutrients in the nation's streams and ground-
water, 1992—2004. U.S. Geological Survey Circular, 1350, 1—174.
Hallberg, G. R., & Keeney, D. R. (1993). Nitrate. In W. M. Alley (Ed.),
Regional ground-water quality. New York: Van Nostrand Reinhold.
Johnson, C. J., Bonrud, P. A., Dosch, T. L., Kilness, A. W., Senger, K. A.,
Busch, D. C., et al. (1987). Fatal outcome of methemoglobinemia in an
infant. Journal of the American Medical Association, 257(20), 2796—2797.
Katz, B. G., Bohlke, J. K., & Hornsby, H. D. (2001). Timescales for
nitrate contamination of spring waters, northern Florida, U.S.A.
Chemical Geologi/, 179(1— 4), 167—186.
Kendall, C., & Aravena, R. (2000). Nitrate isotopes in groundwater
systems. In P. G. Cook & A. L. Herczeg (Eds.), Environmental
tracers in subsurface hydrology. Boston: Kluwer Academic Press.
Lindsey, B. D., Berndt, M. P., Katz, B. G., Ardis, A. F., & Skach, K.
(2009). Factors affecting water quality in selected carbonate
aquifers of the United States. U.S. Geological Survey Scientific
Investigations Report 2008-5240.
Lindsey, B. D., Katz, B. G., Berndt, M. P., Ardis, A. F., & Skach, K.
(2010). Relations between sinkhole density and anthropogenic
contaminants in selected carbonate aquifers in the eastern United
States. Environmental Earth Sciences, 60, 1073—1090.
Mueller, D. K., Hamilton, P. A., Helsel, D. R., Hitt, K. J., & Ruddy, B. C.
(1995). Nutrients in ground water of the United States—An analysis
of data through 1992. U.S. Geological Survey Water-Resources
Investigations Report 95-4031.
Panno, S. V., Hackley, K. C., Hwang, H. H., & Kelly, W. R. (2001).
Determination of the sources of nitrate contamination in karst
springs using isotopic and chemical indicators. Chemical Geology,
179(1-4), 113-128.
Peterson, E. W., Davis, R. K., Brahana, J. V., & Orndorff, H. A. (2002).
Movement of nitrate through regolith covered karst terrane,
northwest Arkansas. Journal of Hydrology, 256, 35—47.
Plagnes, V., & Bakalowicz, M. (2001). The protection of karst water
resources: The example of the Larzac karst plateau (south of
France). Environmental Geology, 40, 349—358.
Scanlon, B. R. (1990). Relationships between groundwater contamination
and major-ion chemistry in a karst aquifer. Journal of Hydrology, 119,
271-291.
Spalding, R. F., & Exner, M. E. (1993). Occurrence of nitrate in ground-
water—A review. Journal of Environmental Quality, 22(3), 392—402.
Wells, E. R., & Krothe, N. C. (1989). Seasonal fluctuation in 6 15 N
of groundwater nitrate in a mantled karst aquifer due to macropore
transport of fertilizer nitrate. Journal of Hydrology, 112, 191—201.
Zhu, X. Y., Xu, S. H., Zhu, J. J., Zhou, N. Q., & Wu, C. Y. (1997).
Study on the contamination of fracture-karst water in Boshan dis¬
trict, China. Ground Water, 35(3), 538—545.
NULLARBOR CAVES, AUSTRALIA
Julia M. James, A nnalisa K. Contos, and Craig M. Barnes
University of Sydney, Australia
INTRODUCTION
In Australia, the karst land, referred to as the Nullarbor,
covers an area in excess of 200,000 km 2 , although in
proportion to its area the karst is poor in known caves.
Despite being referred to by their collective name, the
Nullarbor Caves, many of the individual caves are
world class in their own right. They have played an
important role in elucidating Australian prehistory and
they have been the focus of many intrepid expedi¬
tions—first to find the caves and then to explore them
both in the dry waterless passages and in the extensive
flooded passages.
The very existence of caves beneath a desert is
remarkable, and because of this they have individually
and as a group contributed significantly to the under¬
standing of speleogenesis in arid environments. These
caves are renowned for their rare secondary minerals,
and the great variety of novel cave decorations has pro¬
vided valuable insight into past surface and climatic
conditions. The desiccating environment in these caves
has also preserved biological materials in such pristine
condition that they are considered outstanding repo¬
sitories of paleoenvironmental information. Despite
being nutrient poor, the dry passages of the Nullarbor
Caves host a range of terrestrial invertebrates including
several unique troglobitic species. Colonies of microbes
have produced abundant biofilms in the saline cave
waters and these contain many new species of bacteria.
PHYSICAL SETTING
The Nullarbor Caves are found on one of the largest
continuous exposed karst lands in the world in the
southwest part of the Australian continent (Fig. 1).
These caves are developed in four limestone strata: the
Nullarbor (early-middle Miocene), the Mullamullang
member of the Nullarbor limestone, Abrakurrie, and
Wilson Bluff (middle-late Eocene). Bounded to the
north by the Colville Sandstone, the limestone strata
are unconformable, almost horizontally bedded and
dip to the south-southeast (Fig. 2). Uplift of the lime¬
stones probably occurred at the end of the Lower
Miocene. Relatively weak jointing and the absence of
folding or major tilting suggest that the region has
remained relatively stable since uplift. Particularly
important to cave development is that all limestone
strata are both porous and permeable.
The Nullarbor is a plain that is almost completely
flat. The plain in the south is sparsely vegetated and in
the north it is treeless. At present, the tree line that
delineates the edge of the treeless plain is approxi¬
mately 80 km north of Eucla (Fig. 1). Pollen studies of
sediments and gut contents of extinct fauna have
shown that the "tree line" extended to Eucla during
the last glacial, a time when the coastline was 160 km
farther to the south. These studies have also shown
that the vegetation has remained similar throughout
and since the last cold period.
Encyclopedia of Caves.
© 2012 Elsevier Inc. All rights reserved.
NULLARBOR CAVES, AUSTRALIA
569
FIGURE 1 Location map for the Nullarbor Caves. The golden yellow color depicts the limestone area. After Lowry and Jennings (1974).
Nullarbor
limestone
Abrakurrie
limestone ^
Wilson Bluff
limestone
100
0 meters
0 100
Kilometers
x 125 Vertical
exaggeration
Coliville sandstone
FIGURE 2 The geology of the Nullarbor (after Lowry and Jennings, 1974)— the Mullamullang strata are not shown for clarity—together
with a plate showing an outcrop of the limestone strata exposed in the cliffs bordering the Great Australian Bight. Photograph taken by Annalisa
Contos.
ENCYCLOPEDIA OF CAVES
570
N
FIGURE 3 The mouth of Pannikin Plain Cave set
in the flat plain. The cave entrance is 10 m in diameter
and after 8 m leads down through collapse to the main
passage. Photo courtesy of Dirk Stoffles.
FIGURE 4 The main chamber in Abrakurrie.
The chamber is in the white Wilson Bluff lime¬
stone. The flat floor consists of fluvial sediments
mixed with the products of crystal weathering.
The volume of the chamber is ~ 135,000 m 3 .
Photograph taken by Julia James.
The major surface karst features are ridges and cor¬
ridors, with relief of 3—5 m amplitude, and large but
extremely shallow circular depressions called dongas.
Set into the plain and distributed throughout the
Nullarbor are thousands of blowholes, holes in the
ground through which air blows in and out strongly.
More frequent in the southern part of the plain are
hundreds of abrupt, sharp-edged collapse dolines
(Fig. 3). Typical of semiarid to arid karst, the Nullarbor
has poorly developed karren features, although there
are extensive calcretes and the surface of the limestone
is case hardened. Immediately below the indurated
surface, the rock mass is extensively weathered with
many small cavities.
The caves can be entered from either blowholes or
dolines that often descend over breakdown to large
passages that are tens of meters in width and height.
The passages enlarge to form huge chambers (Fig. 4)
with domed ceilings and arch entrances. The most
striking feature of these caves is the spectacular lakes
(Fig. 5) of considerable depth and clarity. These lakes
have helped establish the international reputation of
the Nullarbor Caves. In a few of the known caves, kilo¬
meters of flooded passage leading from the lakes have
been discovered.
Although open to debate, some evidence suggests
that the Nullarbor has been semiarid to arid for some
6 myr. At present, the Nullarbor has a warm semiarid
ENCYCLOPEDIA OF CAVES
NULLARBOR CAVES, AUSTRALIA
571
FIGURE 5 Weebubbie Lake has an area of ~ 7500 m 2 . From the end of the lake a series of large flooded passages continue, reaching a
depth of 45 m. Photo courtesy of Alan Warild.
climate in the south, but it is hot and arid in the north.
Surface temperatures range widely both seasonally and
diurnally as a result of continentality, although winters
are short and cool to mild. The average temperature of
the caves is 18°C. An important meteorological feature
is that potential evaporation exceeds precipitation for
the entire region (Fig. 1). The rainfall pattern is unreli¬
able; along the coast most precipitation falls in winter as
light rain and mist, whereas inland, irregular summer
thunderstorms distribute the rainfall. This annual pat¬
tern is interrupted every few summers by intense rain¬
bearing cyclonic depressions. These events can drop
hundreds of millimeters of rain in an hour, causing con¬
siderable local flooding.
There are no permanent streams on the Nullarbor,
but there are numerous intermittently active water¬
courses. During intense rainfall, water flows into
depressions or collapse dolines and rapidly disappears
underground. In the caves, the streams flow only short
distances before entering lakes or sinking in sediment.
Lakes are encountered at similar depths in the
Nullarbor Caves and they form part of a regional
water table known as the Nullarbor aquifer. The surface
of the aquifer gently slopes toward the Great
Australian Bight and is close to sea level in height. It
has both conduit and diffuse flow characteristics, with
the flow of water in the caves being from north to
south. The caves act as conduits within the unconfined
aquifer whose surface marks the top of the phreatic
zone. The ultimate destination of the cave waters
remains an enigma as there are no large springs issu¬
ing from the coastal cliffs and dunes, and no known
submarine springs in the Great Australian Bight.
PREHISTORY, HISTORY, AND
SPELEO LOGIC AL HISTORY
On the Nullarbor, the nomadic Aborigines inhabited
the fringes of the plain from the earliest times and
entered the caves in a search for flint. Indeed, several
of the caves, such as Koonalda (Fig. 1), have been
mined for flint. Archaeological excavations 80 m below
the surface, on the edge of the twilight zone in
Koonalda, the only cave containing a constant supply
of potable water, have provided evidence of a period
of uninterrupted human "occupation" spanning ap¬
proximately 30,000 years. The stone artifacts found
there show discernible changes in stone implement
technology but there are no grinding stones, indicating
a lack of grasses and cereals and a purely hunting life¬
style. Several caves are decorated with hand stencils or
primitive engravings similar to artwork found in
Paleolithic caves in Europe and dating to Late Glacial
times.
The presence of caves on the Nullarbor was first
recorded in 1866 just prior to European settlement at
Eucla in 1867. The local Aboriginal tribe was the
Miming, which at its peak numbered 200 persons but
by 1867 numbered only 30. Anthropologists have
ENCYCLOPEDIA OF CAVES
572
N
established that at this time the Miming no longer
used the Nullarbor Caves. Even after European settle¬
ment, the population density of the area remained
very low on a world scale; thus, until the early twenti¬
eth century, many of the Nullarbor Caves formed a
pristine underground wilderness, preserved by their
remoteness in an inhospitable landscape. However, the
change of land use from wilderness to sheep and cattle
grazing, coupled with the introduction of other feral
animals, especially the rabbit, has subsequently caused
considerable damage to many of the caves.
The first expeditions specifically to locate and
explore caves took place during the 1930s. On some of
these early expeditions, a light aircraft was used to
locate the caves. With the formation of Australian spe¬
leological societies in the 1950s, major caving expedi¬
tions commenced. These benefited from the newly
available air photographs, which proved invaluable in
locating caves with large dolines and entrances. The
highlight of the 1950s exploration was the discovery
and survey of Mullamullang, which at 10.8 km was the
longest cave in Australia at that time. The flooded pas¬
sages in many of the caves attracted cave divers, with
the first major attempt at underwater exploration
occurring in 1972. The 1980s saw cave exploration
intensify once more and the use of ultralight aircraft
for prospecting. This systematic searching of the sur¬
face and depressions led to many new caves being
found and explored. In the known caves, passages
were extended by dedicated cavers who were pre¬
pared to continue even when the way became con¬
stricted or required excavation. Improvements in cave
diving equipment saw flooded caves explored to
world record lengths. This attack on the Nullarbor
underground has continued into the new millennium;
even in 2002 new caves and flooded maze passages
coated in crystal were discovered. Since the 1950s, spe¬
leological investigations in the Nullarbor Caves have
taken place in parallel with exploration, leading to
their being among the most studied arid caves in the
world.
ORIGIN AND ENLARGEMENT
OF THE CAVES
At first it was anticipated that the number and size
of the Nullarbor Caves would be limited by a lack of
surface water and biogenic carbon dioxide. This
hypothesis was supported by an apparently immature
karst surface, lacking in richness and diversity of land-
forms. In contrast, the enormous caves found indicated
that below ground the karst was mature, while the
large number of blowholes showed that cavities are
numerous beneath the surface, the latter supported by
extensive drilling records. In fact, the Nullarbor is not
lacking in caves but in entrances to them. Old
Homestead (Fig. 1) illustrates this; it has a single
entrance, yet on the surface above the cave at least 30
blowholes too narrow for human passage have been
identified as having an air connection with the voids
below. Within the cave there is evidence of additional
entrances that are now blocked with eolian surface
materials.
As early as the 1970s, scientists hypothesized that
the caves had shallow phreatic origins, developing in
zones of fluctuating water tables. This hypothesis was
readily accepted as an explanation for the smaller,
shallow Nullarbor Caves. For the huge caves, other
investigators invoked more pluvial times in the past to
supply copious aggressive waters for dynamic phreatic
development. The existence of the paleodrainage chan¬
nels found in the north of the plain (Fig. 1) was
regarded as support for this argument. At present,
runoff generated by intense rainfall is the only aggres¬
sive water to enter the caves. Freshwater lenses form
on top of the brackish waters in the cave lakes, allow¬
ing corrosion notches (Fig. 6) to develop on the walls
around the lake.
The brackish waters in the lakes, flooded passages,
and the Nullarbor aquifer are saturated with respect to
calcite, evidence that they are dissolving limestone and
have a role in the development of the caves. They have
large amounts of total dissolved solids, with sodium
ions and chloride dominating. The highest levels of
totally dissolved solids in caves are found in Nurina,
where the water is hypersaline, and the lowest in
Mullamullang (Fig. 1). The generalization that salinity
in the cave waters decreased with distance from the
ocean led to the hypothesis that seawater was back-
flooding into the caves. Stable isotope analysis has
shown that the source of the salt is actually sea spray
and that it is the salt content of the meteoric waters
that decreases with distance from the ocean.
The aggressive meteoric waters are implicated in
many aspects of karst and cave development on the
Nullarbor. The light rains dissolve some limestone on
the surface and in the soil profile but when it evapo¬
rates calcium carbonate (calcite), sodium chloride
(halite), and calcium sulfate-2water (gypsum) crystal¬
lize out on the surface. Subsequent cycles of light rain
and evaporation increase the salt content on surface
and in soil. This concentration of calcium carbonate,
where potential evaporation exceeds rainfall, is a rec¬
ognized mechanism for the formation of calcretes and
case-hardening of limestone. On the arid Nullarbor,
halite and gypsum are also concentrated in this man¬
ner. During periods of intense rain, the soluble salts
concentrated through the above mechanism dissolve.
The resulting vadose waters are aggressive because
ENCYCLOPEDIA OF CAVES
NULLARBOR CAVES, AUSTRALIA
573
FIGURE 6 Tafoni in Cocklebiddy entrance
passage. The tafoni are in lines separated by
harder layers in the limestone. Gypsum and
halite efflorescence can be seen on the curved
surfaces and a fine-powdered deposit of gypsum
and calcite sand can be seen on the ledges below
the tafoni. Above the surface of the lake is a
deeply cut corrosion notch. Photograph taken by
Julia James.
they have increased ionic strength and are thus able to
create cavities immediately below the indurated sur¬
face. These cavities, called the zone of intense phreatic
preparation, were previously interpreted as resulting
from shallow phreatic solution at times of past high
water tables. However, using the concept of renewed
aggressivity from increased soluble salt content, it is
possible to propose that the zone is recent, not ancient,
and vadose not phreatic, in origin.
The vadose waters continue through the limestone,
now saturated with respect to calcite and with a high
ionic strength, until they reach the water table. The
brackish waters of the cave conduits and lakes and of
the Nullarbor aquifer are also saturated with respect to
calcite. When these two waters mix, the mixed water is
aggressive and mixing corrosion takes place at the
mixing surface. The location of the mixing surface in
the karst has varied over time as sea levels were low¬
ered during periods of glacial advance. It is likely that
this process has been active throughout the history of
cave development beneath the Nullarbor, creating
shallow phreatic caves.
Whatever the process of creating the first caves,
when they drain due to the lowering of the water
tables, the major process of enlargement is considered
by all to be crystal weathering. It is the scale and inten¬
sity of crystal weathering in the Nullarbor Caves that
makes them the foremost example in the world of this
process of cavern enlargement. The agents of crystal
weathering are the meteoric salts in the vadose waters.
The Nullarbor Caves are barometric breathers; thus,
airflow in the caves reverses in response to pressure
change. This breathing removes moisture from the
caves and replaces it with desiccating desert air. If
the humidity in the caves is sufficiently reduced, the
vadose waters will start to evaporate until they become
so concentrated that salts they are carrying crystal¬
lize in the surface layers of the limestone. Calcite, the
first mineral to crystallize, can be observed in thin sec¬
tions of the wall rock. Following calcite, gypsum and
then halite crystallize causing granular disintegration
of the wall rock. Gypsum is the more effective weath¬
ering agent because it has greater expansion on crystal¬
lization. This process produces upward-sloping domes
and highly fretted walls. The caves enlarge, sloping
upward due to crystal weathering and subsequent
collapse, ultimately penetrating to the surface. These
collapsed entrances allow aggressive runoff to reach
the caves and contribute to their enlargement.
The collapse and crystal weathering products would
fill a cave if they were not removed. Beneath the
Nullarbor, there are a number of ways for their removal.
Mixing corrosion at the vadose and phreatic interface
will dissolve them. In Murra-el-elevyn, aggressive runoff
waters are contributing to their removal. Solution within
the freshwater lens has etched fallen boulders, leaving
fossil corals standing up to 3 cm proud of their surface.
Eolian processes are the only way the products of crys¬
tal weathering can be removed from the dry cave pas¬
sages. In Mullamullang, crystal weathering has reduced
the bedrock to sand below a dome. The sand, carried by
the strong cave winds, shifts through the cave.
Crystal weathering masks, modifies, and mimics
other geomorphic forms in the caves. Pseudocorrosion
notches form around lakes that never receive runoff.
These are a result of the brackish waters rising through
the surrounding porous rock by capillary action and
evaporation. This produces a band of halite and gyp¬
sum flowers and causes granular disintegration of
the wall rock. Figure 6 shows the multiple rows of
tafoni in the entrance chamber of Cocklebiddy. To the
advocates of dynamic phreatic solution, tafoni were
ENCYCLOPEDIA OF CAVES
574
N
solution scallops but are actually a product of crystal
weathering. They have formed in bands separated by
harder indurated limestones. These harder layers were
generated when the parent coral reefs that produced
the limestones emerged briefly above water.
Crystal weathering is also responsible for the novel
"coffee and cream" sediments found in many of the
caves. The weathering products are made up of two
forms, which appear to flow over one another. Both
the light- and dark-colored materials are made up of
high magnesium calcite with different impurities. The
"cream" contains goethite-dominated iron minerals
and some kaolin. The strikingly contrasting dark
brown coffee has more gypsum, hematite, and manga¬
nese dioxide.
It has been proved that crystal weathering is occur¬
ring at observable rates in the caves under the present
climatic conditions. Thus, it is easy to accept that these
enormous caves could have enlarged continuously
over millions of years at a slow but variable pace dom¬
inated by this geomorphic process.
Throughout this Encyclopedia of Caves article, the
words large, enormous, and huge are used. Yet none of
the Nullarbor Caves would rate a mention in lists of
the deepest and longest caves of the world. Old
Homestead is the longest of the Nullarbor Caves at
25 km. Mullamullang is 135 m deep and is the second
longest cave (>13 km). Weebubbie is equally deep but
its large passage (Fig. 5) is submerged for 45 m of its
overall depth. It is the flooded passages of the
Nullarbor Caves that have attracted international expe¬
ditions in search of world records. Cocklebiddy has
the longest underwater section, 6.5 km through large
passages with only two modest air spaces to the limit
of exploration and has held world records. Pannikin
Plain has the largest flooded passages; they are the
size of the Abrakurrie main chamber, the largest
air chamber in Australia (Fig. 4). Another feature of
the Nullarbor Caves that is internationally recognized
is the strength of the cave winds. These have been
measured at 70 km/h" 1 issuing from the blowhole
entrance of Thampanna.
MINERALS
The Nullarbor Caves are renowned for their second¬
ary minerals. The most unusual are the very rare
organic minerals produced within bat guano. These
are exceptionably soluble and are thus found only in
the driest of caves. Only nine organic minerals have
been reported from caves worldwide; of these, five
were first collected from the Nullarbor Caves.
Halite is rare in limestone caves around the world,
yet beneath the Nullarbor it is abundant and occurs in
FIGURE 7 A halite curtain ~100 mm high. Photo courtesy of Norm
Poulter.
all the known forms of halite speleothems. Of excep¬
tional beauty are the wide varieties of delicate and
erratically shaped halite flowers (Fig. 7). The source of
the halite is the meteoric salt in the vadose water. It
persists only in places in the caves where there is con¬
tinuously low relative humidity. One of the most excit¬
ing halite speleothems found is a 2.7-m-long fallen
column. For its protection this giant, christened Big
Salty, was removed to the Western Australian Museum.
A number of uranium series dates were obtained from
it for growth rate studies and it was concluded that
there had been two previous periods on the Nullarbor,
dry enough for massive halite deposition. The same
cave has five other halite columns, disguised from col¬
lectors by their unattractive appearance.
Gypsum is even more widespread and is found in
many speleothem shapes. Halite is substantially more
soluble than gypsum. This differential solubility is
demonstrated when slow-dripping gypsum straws
develop halite tips that are lost after subsequent rain
increases flow through the straws.
In the Nullarbor Caves calcite speleothems, once
regarded as rare, are widespread, although under the
current arid conditions subaerial calcite deposition is
rare. Conventional calcite deposition by CO 2 degassing
is to be found in Nurina where pools are covered with
calcite rafts that have coalesced to form fragile floors
and have stacked into piles against the cave walls.
Ancient "black" calcite is found in all of the com¬
mon speleothem shapes. This black calcite, which has
a humic compound as the chromophore, is seen both
on the surface and throughout the caves to the water
table. Uranium series dating has shown that its deposi¬
tion ceased more than 350,000 years BP. Flowstones of
black calcite over a meter thick suggest that its deposi¬
tion extended over a long period of time. The black
calcite must have been deposited when the climate
was wetter or at a time when there was a much lower
potential evaporation in order to allow the calcite-
ENCYCLOPEDIA OF CAVES
NULLARBOR CAVES, AUSTRALIA
575
saturated waters to reach the caves. A more extensive
vegetation cover would also have been necessary to
provide both the carbon dioxide and the humic chro-
mophore. Some of the massive speleothems of black
calcite have been reduced to shards by the relentless
attack of the agents of crystal weathering.
Mixing-zone calcites lie submerged in the flooded
passages of the Nullarbor Caves. In Tommy Grahams
(Fig. 1), these speleothems are being deposited below a
halocline at a depth of 22 m. Above the halocline, the
limestone wall rock is being eroded, while below it cal¬
cite is precipitating as a result of "mixing crystalliza¬
tion." Similar calcite deposits are found in many of the
water-filled Nullarbor Caves. Particularly significant is
a calcite flowstone over a kilometer in length and cov¬
ering many meters of the lower passage walls and
floor in the flooded passage of Cocklebiddy. It is the
nature of their deposition that makes these a valuable
source of past environmental data.
Calcite biominerals encapsulated within biofilms
have formed speleothems that have been named
microbial mantles. Six different mantle morphologies
have been observed: icicles (Fig. 8), U-loops, bulbs,
feathers, and thin roof and thick floor sheets. Biofilms
are implicated in both the nucleation and habit modifi¬
cation of the spindle-shaped microcrystals found
within them. The floors below the mantles are either
snow fields of loosely packed white crystalline detritus
or pavements of cemented calcite.
BIOLOGY PAST AND PRESENT
The Nullarbor Caves contain repositories of biologi¬
cal materials of outstanding paleoenvironmental value
preserved in the desiccated sediments. The bone and
subfossil material found covers a range of faunal
FIGURE 8 Microbial mantles, icicle morphology. The gelatinous
biofilm can be seen with its encapsulated calcite crystals reflecting light
from the flash gun. In the top right corner, above the mantles, the lime¬
stone bedrock is etched. Photo courtesy of Peter Rogers.
elements, including extinct megafauna—many modern
species now extinct on the Australian mainland
and species whose range has markedly contracted.
The most exciting findings have been the complete
skeleton of a marsupial lion Thylacaleo and mummi¬
fied Thylacines. The latter species is known as the
Tasmanian tiger and is now extinct, although at the
time of European settlement it was found in Tasmania
and the southeastern mainland. Desiccation resulted in
mummies being found with soft tissues intact; one car¬
cass has been dated at 4600 BP.
The entrance is the only source of nutrients and
energy for the cave ecosystems, excepting those caves
inhabited by the bat species Chalinolobus morio where
modest guano piles add additional nutrients. This lim¬
ited supply of nutrients restricts the numbers and types
of invertebrates. Despite this, the Nullarbor Caves host
a range of terrestrial invertebrates, which includes sev¬
eral remarkable and unique troglobitic species that are
survivors of local evolutionary, environmental pro¬
cesses. These troglobites have preserved a tiny sample
of late Tertiary/Pleistocene surface fauna, now absent
from the region, and give the systems their zoogeo¬
graphic importance. The caves also contain a large num¬
ber of troglophilic insects (mainly cockroaches, beetles,
and crickets) and accidental species.
The almost total absence of aquatic fauna in the
cave waters of the aquifer is surprising. Salinity is not
the cause of this absence as there is one known aquatic
troglobitic species, an amphipod crustacean, in the
hypersaline waters of Nurina. Neither is a lack of
nutrients responsible, for biofilms are abundant in the
cave waters and, like the invertebrate fauna, are
remarkable. So far, microbiological studies have been
restricted to the calcite precipitating microbial mantles
(Fig. 8). They are communities of microorganisms
consisting primarily of Pseudomonas spp. and Pseudoal-
teromonas spp. as well as nitrite oxidizers. Of the 36
phylotypes, 12 could not be identified to a subdivision
level and only 2 showed any relationship to previously
described environmental clones.
The presence of chemotrophic bacteria in the
Nullarbor introduces another variable into the under¬
standing of the cave origins and their enlargement.
They can add and remove carbon dioxide or hydrogen
carbonate from the cave waters, thus dissolving lime¬
stone or precipitating calcite. They are widespread in
the flooded cave passages and thus they may have a
considerable impact.
CONCLUSION
In the world of caves, where in most cases biogenic
carbon dioxide and surface waters provide the
ENCYCLOPEDIA OF CAVES
576
N
essential ingredients for the solution of limestone and
the precipitation of decorations, the more of both, the
bigger and more spectacular the caves. The Nullarbor
Caves provide an example of how caves can originate,
enlarge, and become decorated with speleothems at
times when there is a shortage of both.
Bibliography
Contos, A. K., James, J. M., Heywood, B. R., & Rogers, P. A. W.
(2001). Morphoanalysis of bacterially precipitated subaqueous cal¬
cium carbonate from Weebubbie Cave, Australia. Geomicrobiology
Journal, 18, 331—343.
Davey, A. G., Gray, M. R., Grimes, K. G., Hamilton-Smith, E., James,
J. M., & Spate, A. P. (1992). World Heritage significance of karst and
other landforms in the Nidlarbor region. University of Canberra,
ACT, Australia: Applied Ecology Research Group.
Ford, D. C., & Williams, P. W. (1989). Karst geomorphology and hydrol¬
ogy. London: Chapman & Hall.
Gillieson, D. (1996). Caves processes, development, management. Malden,
MA: Blackwell.
Hill, C. A., & Forti, P. (1997). Cave minerals of the world (2nd ed.).
Huntsville, AL: National Speleological Society.
Holmes, A. J., Tujula, N. A., Holley, N., Contos, A. K., James, J. M.,
Rogers, P. A. W., et al. (2001). Phylogenetic structure of unusual
aquatic microbial formations in Nullarbor Caves, Australia.
Environmental Microbiology, 3, 256—264.
James, J. M. (1992). Corrosion par melange des eaux dans les grottes
de la Plaine de Nullarbor, Australie. In J. N. Salomon & R. Marie
(Eds.), Karst et evolutions climatiques. Bordeaux, France: Presses
Universitaires de Bordeaux.
Long, J., Archer, M., Hannery, T., & Hand, S. (2002). Prehistoric
mammals of Australia and New Guinea: One hundred million years of
evolution. Sydney, Australia: University of New South Wales
Press.
Lowry, D. C., & Jennings, J. N. (1974). The Nullarbor Karst,
Australia. Zeitschrift fur Geomorphologie, 18, 35—81.
ENCYCLOPEDIA OF CAVES
PALEOCLIMATE RECORDS FROM
SPELEOTHEMS
Victor J. Polyak and Rhaivn F. Denniston '
University of New Mexico, ^Cornell College
INTRODUCTION: WHY SPELEOTHEMS
FOR PALEOCLIMATE RESEARCH
Stalagmites contain numerous physical and chemical
characteristics that can be used to reconstruct past cli¬
mate records, although stable isotopes of oxygen and
carbon are the two most frequently used. Visual char¬
acteristics are shown in Figure 1. Oxygen isotope
values, once thought to be reliable temperature indica¬
tors, are now used primarily as indicators of moisture
source or rainfall amount while carbon isotopes
express changes in vegetation, soil, and aridity
(Denniston et al., 1999; Lachniet, 2009). The roles played
by concentrations and isotope values of trace elements
such as strontium, magnesium, and uranium have
future potential, as does grayscale. This article focuses
on published continuous paleoclimate records, most of
which were derived from carbon and oxygen stable
isotopes.
Most stalagmites grow in environments of total
darkness, where temperature and humidity changes
are dampened, and this growth may occur continu¬
ously for thousands and even hundreds of thousands
of years. Water falling as rain or snow finds its way
through the soil and into fractures in the limestone
rock, and eventually into the caves. This water has a
direct link to climate, and oxygen, hydrogen, and car¬
bon that represent that surface environment are trans¬
ferred by the water to the calcite crystals that form
stalagmites and other speleothem types.
Thus far the best records of past climate change are
being generated from fine-scale sampling of powders
for carbon and oxygen stable isotope analyses tied to
high-resolution uranium-series dating of stalagmites.
The records included in this chapter illustrate the
influence these cave-related studies are having on cli¬
mate reconstructions from around the globe (Fig. 2).
IMPORTANCE OF ABSOLUTE
CHRONOLOGY: DATING TECHNIQUES
In order to be useful, all paleoclimate records
must provide for some means of age control, and
speleothems are ideally suited for dating by uranium-
series techniques. Uranium-238 ( 238 U), which is typi¬
cally present in 0.1—1 ppm abundances in speleothem
calcite, decays through a series of relatively short-lived
daughters to lead-206 ( 206 Pb), and U-Pb dating is a cut-
ting-edge component of speleothem-based paleoclimate
work. However, disequilibrium of the 238 U- 234 U- 230 Th
portion of this decay chain is the most commonly
applied method for dating speleothems. In normal,
shallow crustal waters (i.e., well-oxygenated, slightly
alkaline to slightly acidic pH), U is highly soluble while
thorium (Th) is insoluble, and thus cave drip waters
often contain very high U/Th ratios. Crystallization of
speleothem calcium carbonate from such solutions
satisfies one criterion for dating: that nearly all 230 Th
contained within the sample was formed by in situ
decay from 234 U and thus 234 U/ 230 Th ratios may be
used as an absolute chronometer.
The measuring of U and Th isotopic ratios was origi¬
nally performed by alpha spectrometry, which required
large samples and long counting times while offering
only limited temporal resolution (±10% of the age), but
advances in thermal ionization mass spectrometn / (TIMS)
greatly improved the U-Th dating techniques over the
earlier alpha radiation methods, allowing uncertainties
at ±1% of the age. The latest advancement is the appli¬
cation of the multicollector inductively coupled plasma
mass spectrometer (MC-ICPMS) which has led to signifi¬
cant improvements on the measurement of the decay
Encyclopedia of Caves.
577
© 2012 Elsevier Inc. All rights reserved.
578
p
~200-year
growth hiatus
Glacial-aged stalagmites
about one meter tall.
Stalagmite top
at -30,000
years ago.
Stalagmite quit
8000-yr hiatus
growing 20,000
of growth in
years ago.
stalagmite BC2.
Transparent slice of
stalagmite FS4, Fort
Stanton Cave, New
Mexico. Bottom age
is 54,000 years.
Top -65 years
(annual bands) of
stalagmite BC11.
hin section of the top three
cm of stalagmite BC11.
FIGURE 1 Stalagmites are the speleothem of
choice for paleoclimate studies. (A) Typical stalag¬
mites in Fort Stanton Cave, southwestern United
States. (B) Stalagmite BC2 from Carlsbad Cavern
grew during the Late Holocene and exhibits two
growth hiatuses. (C) Stalagmite FS4 grew during
Glacial 1. Growth of this stalagmite terminated
during or immediately after the glacial maximum.
Transparency of this slice shows subtle growth
layers. (D) A thin section showing the gray differ¬
ences in the top 300 years of stalagmite BC11,
Carlsbad Cavern. The top dark area includes the
past 100 years, the period of cave commercializa¬
tion. (E) Magnification of (D) shows the annual
bands in the upper 65 years of growth.
constants (half-lives) of 230 Th, 234 U, and 235 U and which
has increased the precision of isotopic measurements.
These changes, in turn, have led to improved accuracies
of U-series-generated chronologies with absolute ages
having errors in sub-1 % ranges for samples hundreds of
thousand years old (Cheng et ah, 2009).
CONTRIBUTIONS TO
PALEOCLIMATOLOGY
The Early Studies: Paleotemperature
The early paleoclimate studies using stalagmites
were focused on paleotemperature reconstructions
over time. The functionality of speleothems as paleo¬
temperature records is founded on two relationships:
the temperature dependence of the oxygen isotopic
composition of meteoric precipitation (+0.7%o/°C) and the
temperature-dependent oxygen isotopic fractionation that
accompanies mineral crystallization (—0.23%o/°C). Because
the climate of deep, poorly ventilated caves remains steady
throughout the year at approximately mean annual tem¬
perature, and because drip waters often reach thermal
equilibrium with their surroundings during infiltration
into deep caves, the operative relationship linking spe¬
leothem calcite 6 ls O values to mean annual temperature is
approximated by 0.7® c / o C —0.23%o/°C = 0.4%c/°C.
However, the magnitude and direction of the slope
linking meteoric precipitation with temperature intro¬
duces a large degree of uncertainty into this relation¬
ship both spatially and temporally. Changes in the
seasonality, source, or original isotopic composition of
precipitation may not remain constant through time,
particularly during glacial/interglacial transitions. And
in the low latitudes, temperature-rainwater 6 ls O
ENCYCLOPEDIA OF CAVES
PALEOCLIMATE RECORDS FROM SPELEOTHEMS
579
relationships are replaced by "amount effects" in which
the intensity of rainfall exerts the strongest control on
the oxygen isotopic composition of rainwater. Other
complicating influences include enrichment of ls O in
water by evaporation prior to infiltration or by kinetic
effects on fractionation between HCO 3 and CaCOs.
Because of the complexities involved with extracting
absolute temperature records, later studies using
stable isotopes evolved to the use of stable isotopes
6 ls O and 6 13 C as tracers rather than absolute indicators.
Lachniet (2009) provides a comprehensive review of
stable isotope geochemistry and speleothems.
Devils Hole: The First Long, Continous
High-Resolution Speleothem-Based
Paleoclimate Record
For decades, the benchmark paleoclimate records have
remained the marine foraminifera oxygen isotope curve
and the ice records of Greenland and Antarctica. While
the information these datasets provide is enormously
valuable, each is limited by varying degrees in the preci¬
sion and accuracy of its chronology. For example, marine
sediments cannot be dated by radiocarbon techniques past
50,000 years and paleomagnetism techniques do not offer
precise age constraints. And while annual banding in
Greenland ice allows precise age determinations on less
compacted, Flolocene-age sections, the older, more com¬
pacted portions of the cores have age errors of hundreds to
thousands of years.
The earliest high-resolution paleoclimate study using
speleothems with a chronology based on mass spec¬
trometry is the Devils Flole mammillary calcite record
(Winograd et ah, 1992). Cave mammillaries are subaque¬
ous speleothems that form near the water table. The
mammillaries in Devils Hole formed near a water
table that apparently fluctuated vertically by only a few
tens of meters over the past half a million years.
Continuous deposition of the Devils Hole mammillary
calcite resultantly preserved the carbon and oxygen
stable isotope values of the groundwater paleoclimatic
information.
Variations in both oxygen and carbon isotope values
roughly mimic those of SPECMAP (the marine record)
as shown in Figure 3. But the timing of these variations
offered the first significant challenge to the traditional
orbital (Milankovitch) theory of climate change. While
similar to SPECMAP, glacial-to-interglacial transitions
occurred approximately 15,000 years earlier than
predicted by orbital theory or as dated in marine sedi¬
ments. What the timing of glacial to interglacial transi¬
tions in the Devils Hole record represents remains the
subject of much debate.
A - Devils Hole, Nevada, USA F - Dongge Cave, China
B - Hulu Cave, China G - Cold Water Cave, Iowa, USA
C - Sofular Cave, Turkey H - Villars Cave, France
D - Fort Stanton Cave and Pink Panther Cave, New Mexico, USA 1 - Shanboa Cave, China
E - Botuvera Cave, Brazil J - Actun Tunichil Muknal Cave, Belize
FIGURE 2 Speleothem records globally are producing impressive accurately dated high-resolution paleoclimate records. Figure 2 shows
the location of numerous study sites used in this chapter.
ENCYCLOPEDIA OF CAVES
580
p
kyr BP
0 200 400 600
0 200 400 600
kyr BP
FIGURE 3 Devils Hole, Nevada, produced the first lengthy con¬
tinuous speleothem-based climate record, going back half a million
years and variations in oxygen and carbon isotope values in the spe-
leothem calcite correlated with the marine (SPECMAP) and Antarctic
ice (Vostok) core records. The Devils Hole record is shown above in
comparison to SPECMAP. These records show that glacial periods
(labeled 1—5) have been occurring on a roughly 100,000-yr periodic¬
ity over the past million years, which is similar to the cycle of eccen¬
tricity of Earth's orbit around the Sun.
The Devils Hole record, unlike almost all other
speleothem-derived records, was retrieved from a sub¬
aqueous speleothem. Other high-resolution cave-based
records are being retrieved almost exclusively from
stalagmites.
The Last Glacial: Contribution from Stalagmites
The Greenland ice cores record large shifts in air tem¬
perature during glacial periods, referred to as stadials
and interstadials. These include the Bolling-Allerod oscil¬
lations and Dansgaard—Oeschger events, the Younger
Dryas, and Heinrich events. A major question that spe¬
leothem studies are helping to answer is "are stadials
and interstadials in the ice core record coincident with
similarly profound climate changes in continental
regions around the globe?" Many of the stalagmite
records documented so far have not only demonstrated
the existence of significant climate change coincident
with stadials and interstadials, but they are providing
increasingly precise timing of these events as well.
Three records from stalagmites showing remarkable
agreement with climate changes throughout the last glacial
period are from Hulu Cave, eastern China, Sofular Cave,
northern Turkey, and Fort Stanton Cave, southwestern
United States. Each of these records is shown in Figure 4
Yr BP
10000 20000 30000 40000 50000 60000
FIGURE 4 Stalagmites are showing that land masses
throughout the Northern Hemisphere experienced changes in
climate synchronous with the changes recorded in the
Greenland ice cores. Records in this figure were retrieved
from stalagmites in North America (stalagmite FS2, Fort
Stanton Cave, southwestern United States), Europe (stalag¬
mites So-1 and So-2, Sofular Cave, northern Turkey), and Asia
(stalagmites H82, MSD, MSL, PD, and YT, Hulu Cave, eastern
China), and are compared to the Greenland ice core record
GCC05. These records show that changes recorded in the
Greenland ice cores drove changes in moisture regime/inten¬
sity or vegetation as well as the variations in climate that
accompanied these shifts. Increasingly detailed studies will
ultimately provide absolute accuracy of the timing of these
major shifts in clime.
ENCYCLOPEDIA OF CAVES
PALEOCLIMATE RECORDS FROM SPELEOTHEMS
581
in comparison to the Greenland ice core record (GCC05,
Svensson et al., 2005).
The record from Hulu Cave, China (Wang et al.,
2001), reveals precise changes of monsoonal activity
related to the last ice-age termination, and to all the
stadials and interstadials of the last glacial cycle. This
record, based on multiple stalagmites (stalagmites
H82, MSD, MSL, PD, and YT) became the stalagmite
record to which later speleothem-based studies of the
last glacial period compared their records.
The Sofular Cave data, produced from stalagmites
So-1 and So-2, yielded a largely continuous 50,000-year
record (Fleitmann et al., 2009) through the last glacial.
Both the carbon and oxygen values were characterized
by isotope variations that matched the Greenland sta¬
dials and interstadials. The S 13 C values in stalagmite
So-1 (Fig- 4) produced a record more similar to the
Greenland ice core 6 ls O records. Timing of
Dansgaard—Oeschger events in stalagmite So-1
matched those of Hulu Cave and suggested some dis¬
crepancies with the timing of events as dated in the
Greenland ice core.
The Fort Stanton Cave stalagmite yielded a continu¬
ous, 45,000-year-long record of the last glacial from
56,000 to 11,500 yr BP (Asmerom et al., 2010). This record
includes large shifts in oxygen isotope values similar to
and synchronous with those of the Greenland ice core
records and which are related to moisture sources affect¬
ing the American southwest. Heinrich events drive more
Pacific-dominated winter-like moisture into the south¬
western United States resulting in significantly lower
stalagmite 6 ls O values, while Dansgaard—Oeschger
events drive more Gulf of Mexico-dominated summer-
like moisture into the region, resulting in significantly
higher 8 ls O values. The outcome is a record that pre¬
serves large shifts in oxygen isotope values, similar to
those of the Greenland ice cores. As with the Sofular
Cave stalagmites, the Fort Stanton time series also con¬
tains some discrepancies with the timing of events in the
Greenland ice core.
Tracking Insolation and Paleomonsoons
Kutzbach (1981) described how orbital changes in
solar radiation drive monsoon variability over tens of
thousands of years. As summer insolation increases,
largely due to the precessional (wobble) Milankovitch
cycle, so do the land-sea temperature gradients that, in
turn, drive monsoon circulation in the subtropics. Thus,
according to this so-called orbital monsoon hypothesis,
wetter monsoons accompany higher summer insola¬
tion. Monsoon strength can be tracked by speleothems
because the oxygen in speleothem carbonate is closely
tied to the 6 ls O values of meteoric precipitation, which
in monsoon areas is dominated by the amount effect.
A remarkable relationship between insolation and
monsoon variability from a South American stalagmite
was reported by Cruz et al. (2005). Soon after, another
record from the same cave demonstrated the repeat¬
ability and reliability of these records from stalagmites.
Figure 5 shows these two stalagmite time series from
Botuvera Cave, Brazil, which span the past 116,000
years, and that are characterized by oxygen isotopic
ratios that track the sinusoid of austral summer insola¬
tion. Periods of elevated summer insolation corre¬
spond to the lowest oxygen isotopic ratios (and thus
the heaviest rainfall) (Cruz et al., 2005; Wang et al.,
2007).
But the longest and perhaps most impressive insola¬
tion-monsoon relationship comes from China. This
record of monsoonal activity involved the study of
numerous stalagmites (Cheng et al., 2009) and tracks
the insolation record (Fig. 6) back to 360,000 years.
Ties between Solar Variability and Climate
Change
Stalagmites have also been used to test the role
played by changes in the Sun's output on the strength
of monsoons. Because solar irradiance is tied to radio¬
carbon production in Earth's atmosphere, radiocarbon
measurements made on objects of known age (such as
trees and stalagmites) allow the back-calculation of
changes in solar irradiance, described as A 14 C (Stuiver
et al., 1998). Overlying A 14 C values against stalagmite
oxygen isotopic time series from monsoonal regions
allows an investigation of the role of the Sun on mon¬
soon rainfall. Wang et al. (2005) constructed a stalag¬
mite record from Dongge Cave, China, that spanned
the past 9000 years. Their record, which was sampled
so as to yield 5-year temporal resolution, found a
strong positive covariation between the strength of the
Asian Summer Monsoon and solar irradiance. On the
opposite side of the Pacific, Asmerom et al. (2007)
found that a 12,000-year-long stalagmite record from
Pink Panther Cave, New Mexico, in the American
southwest revealed a strong negative covariation
between solar output and rainfall. Asmerom et al.
(2007) reconciled these seemingly contradictory results
by invoking changes in the Walker Circulation of the
Pacific Basin, which essentially led to La Nina-like con¬
ditions during periods of increased solar output, with
increased rainfall in the western Pacific and decreased
rainfall in the eastern Pacific creating a see-saw rela¬
tionship. Figure 7 shows these two records and their
comparisons.
ENCYCLOPEDIA OF CAVES
582
p
FIGURE 5 Insolation is the measure of
solar radiation impacting the Earth's surface
(watts per square meter—W/nT 2 ) over time
and is generated using Milankovitch's theory.
Some stalagmite studies are showing remark¬
able comparison with insolation. This, figure
shows an impressive match between isotope
records retrieved from stalagmites BT2 and
BTV3a, Botuvera Cave, Brazil, and solar inso¬
lation over the past 120,000 years.
kyrBP
0 100 200 300 400
FIGURE 6 A remarkable match between a stalagmite-based
record from China and insolation over the past 360,000 years. It took
tens of samples to construct this record. The gray isotope values in
the left y-axis correspond to the gray curves in the first part of the
record. Stalagmites are providing crucial information.
Annual Laminations in Stalagmites and Late
Holocene Climate Change
Similar to tree rings, annual bands can form in sta¬
lagmites, recording the growth rate history of these sta¬
lagmites. Numerous studies have been published
showing banding history related to general moisture
availability or temperature change. Two such studies
are offered in Figure 8, one from northeastern China,
near Beijing (stalagmite TS90501; Tan et al., 2003), and
the other one from the southwestern United States (sta¬
lagmites BC2 and BC11; Polyak and Asmerom, 2001;
Rasmussen et al., 2006). The relationship of annual band
thickness to climate is complex due to the numerous
variables involved. However, like the temperature-6 ls O
problem, such banding records have the potential to
accurately reflect climate change variability and to pre¬
cisely record the timing of climatic events. An example
of this is illustrated in Figure 8 where two stalagmites
preserve the same, significant change in stalagmite
growth between 1400 and 1600 A.D. These records may
help resolve climate change interpretations when com¬
bined with other annually resolved records coming
from tree rings, annually laminated lake sediments, and
other high-resolution climate proxies.
ENCYCLOPEDIA OF CAVES
PALEOCLIMATE RECORDS FROM SPELEOTHEMS
583
(A) 0
-40 —
2000 4000
Yr BP
6000 8000
10000
12000
atmospheric A 14 C
0 2000 4000 6000 8000 10000 12000
Yr BP
(B) 0 2000 4000 6000 8000
0 2000 4000 6000 8000
Yr BP
03
CO
o
tyi
Co
o
FIGURE 7 Variations in solar radiation are recorded
in trees for the entire Holocene expressed through the
carbon isotopes, specifically the radioactive carbon iso¬
tope 14 C. Fleitmann et al. (2003) demonstrated that
changes in oxygen isotope values in a stalagmite from
Oman correlated well with changes in atmospheric 14 C
(A 14 C) during the Middle Holocene. This figure (A)
shows a positive correlation between the atmospheric
A 14 C and the oxygen isotope record from stalagmite PP1,
New Mexico, for the entire Holocene, supporting the
interpretation that changes in precipitation regimes in the
southwestern United States were driven by changes in
solar radiation. In (B), stalagmite PP1 is compared to sta¬
lagmite DA, China, and the two records express a nega¬
tive correlation which suggests that monsoonal intensities
in these regions are opposite of each other—a "seesaw"
effect.
Vegetation and Soil Dynamics
Carbon isotopes, unlike oxygen isotopes, cannot be
used to reconstruct past temperature or moisture
regimes but still offers important insight into past
environmental conditions. Carbon in speleothems is
derived from multiple sources. Atmospheric CO 2 and/
or soil CO 2 react with water to form carbonic acid which,
in turn, dissolves carbonate bedrock. Outgassing of CO 2
or evaporation of water drive precipitation of calcium
carbonate via:
Ca 2+ + 2 HCO 3 ^CaC0 3 + H 2 0 + C0 2
While the 8 13 C value of bedrock carbonate is mostly
fixed in time, carbon isotopic ratios of soil CO 2 change
according to the type of vegetation present. Isotopically
lighter C 3 vegetation includes trees, shrubs, and forbs,
while isotopically heavier C 4 plants are made up of
many prairie grasses, with each type favoring different
climatic conditions. Under ideal circumstances, about
half of the carbon in speleothem carbonate is derived
from soil CO 2 and half from bedrock carbonate (atmo¬
spheric CO 2 is typically a trivial source of carbon in
water infiltrating through soils that are thick and/or
contain dense plant cover). Thus, changes in plant type
yield distinct isotopic signals that are transferred into
drip water and then into underlying stalagmites.
This relationship was illustrated by Dorale et al. (1992)
who tied a Middle Holocene increase in 8 13 C values
in stalagmites from Cold Water Cave, north-central
ENCYCLOPEDIA OF CAVES
584
p
Year (AD)
2000 1600 1200 BOO 400 0
0 . 2 -'
-g- 0.16-
E
2000 1600 1200 800 400 0
Year (AD)
FIGURE 8 Annual banding records in speleothems are similar to
those retrieved from trees. This figure offers two annual band
records, one from southwestern North America (stalagmites BC2),
and one from northeastern China (stalagmite TS90501). There are
some similarities in these three records, particularly the period from
1400 to 1600 A.D. Stalagmite BC2 records offered information about
general precipitation over the Late Holocene, while stalagmite
TS9501 yielded a summer temperature record.
United States, to encroaching dry Pacific air masses that
led to replacement of deciduous forest (C3) by prairie
(C 4 ). However, some complexities of speleothem carbon
isotopic dynamics were explored by the Brazilian record
of Cruz et al. (2006) who documented large variations in
6 13 C values that oscillated in concert with the 40,000-year
obliquity (tilt) Milankovitch cycle while fc ,s O values
oscillated with the precession cycle (Fig. 5). These
authors argued that tilt-driven changes in meridional
temperature gradients affected local conditions that
changed plant activity and thus the amount of soil CO2
in drip water. Similarly, Genty et al. (2001) documented
large swings in 6 13 C values in a stalagmite from Villars
Cave, France, that occurred synchronously with D—O
events and that were tied to vegetation dynamics, with
lower stalagmite 8 13 C values corresponding to higher
contributions from soil CO2 during wetter intervals
when plant density and activity were elevated.
Timing of Climatic Changes
The ice core and marine records document the
abrupt termination of glacial periods, but these records
cannot examine the full geographic and temporal differ¬
ences in glacial terminations around the globe. Many
stalagmite records are accurately recording the end of
glacial periods, however. In the Northern Hemisphere
but also in some areas of the Southern Hemisphere, sta¬
lagmites are showing that the termination of the last
glacial (Termination I) occurred around 14,500 years
ago (Fig. 4). The stalagmite records from China (Cheng
et al., 2009) and the Dead Sea region (Frumkin et al.,
1999) are revealing that the termination of the previous
glacial (Termination II) occurred at around 130,000
years ago, and more precisely at 129,000 ±100 years
ago from the Sanboa Cave record (Cheng et al., 2009).
The climate record from Chinese caves also reveals the
timing of Terminations III and IV at 242,000 and 336,000
years ago, respectively. These high-precision dates
allow us to test the orbital (Milankovitch) theory of the
ice ages, and to serve as a point of comparison for less
precise dates of similar events.
Paleotempestology
Shorter-term climatic variables such as tropical
cyclones (hurricanes) may also be assessed using spe¬
leothems. In an attempt to extend the record of tropical
cyclone activity in the past that are available through
direct observation or reconstruction using historical
documents, a variety of paleotempestology archives
have been developed. The most common method takes
advantage of the fact that rain from tropical cyclones is
highly depleted in ls O (Lawrence et al., 2002). Values
of summer rainfall 6%c below normal summer rain
values serve as markers of severe storms, and thus
negative spikes in stalagmite carbonate have been
interpreted as marking tropical cyclone activity.
Several studies have compared historical tropical
cyclones with ls O-depletion in speleothem (e.g., stalag¬
mite and soda straw) carbonate. Of these, the most high-
resolution analysis was that of Frappier et al. (2007) who
tied 8 ls O values of a fast-growing (~1 mm yr -1 ) calcite
stalagmite from Belize to historical (1977—2000 A.D.)
tropical cyclones that tracked through the region. This
study demonstrated unambiguously that through high-
precision microsampling (20-pm intervals, averaging
~75 samples per year of growth, subweekly resolution)
tropical cyclones can be identified individually by com¬
parison with historical records. Stalagmites will provide
valuable information regarding history of hurricane/
cyclone activity for the past several thousand years.
Other Studies
While this article focuses primarily on the stable
isotope geochemistry of stalagmites, other methods exist.
ENCYCLOPEDIA OF CAVES
PALEOMAGNETIC RECORDS IN CAVE SEDIMENTS
585
These include elemental analyses, other isotope systems
such as 87 Sr/ 86 Sr, 234 U / 238 U, magnesium and lithium iso¬
topes, gray and color histograms, growth records defined
by growth hiatuses, laser excitation records, and organic
materials/substances.
See Also the Following Articles
Cosmogenic Isotope Dating of Sediments
Uranium Series Dating of Speleothems
Paleomagnetic Records in Cave Sediments
Bibliography
Asmerom, Y., Polyak, V. J., & Burns, S. J. (2010). Variable winter
moisture in the southwestern United States linked to rapid glacial
climate shifts. Nature Geoscience, 3(2), 114—117.
Asmerom, Y., Polyak, V. J., Burns, S. J., & Rasmussen, J. (2007). Solar
forcing of holocene climate: New insights from a speleothem
record, southwestern United States. Geology, 35(1), 1—4.
Cheng, H., Edwards, R. L., Broecker, W. S., Denton, G. H., Kong, X.,
Wang, Y., et al. (2009). Ice age terminations. Science, 326(5950),
248.
Cruz, F. W., Burns, S. J., Karmann, I., Sharp, W. D., Vuille, M.,
Cardoso, A. O., et al. (2005). Insolation-driven changes in atmo¬
spheric circulation over the past 116,000 years in subtropical
Brazil. Nature, 434(7 029), 63-66.
Cruz, F. W., Jr., Burns, S. J., Karmann, I, Sharp, W. D., Vuille, M., &
Ferrari, J. A. (2006). A stalagmite record of changes in atmo¬
spheric circulation and soil processes in the Brazilian subtropics
during the Late Pleistocene. Quaternary Science Reviews, 25,
2749-2761.
Denniston, R. F., Gonzalez, L. A., Asmerom, Y., Baker, R. G., Reagan,
M. K., & Bettis, E. (1999). Evidence for increased cool season
moisture during the Middle Holocene. Geology, 27(9), 815—818.
Dorale, J. A., Gonzalez, L. A., Reagan, M. K., Pickett, D. A., Murrell,
M. T., & Baker, R. G. (1992). A high-resolution record of
Holocene climate change in speleothem calcite from Cold Water
Cave, northeast Iowa. Science, 258(5088), 1626—1630.
Fleitmann, D., Burns, S. J., Mudelsee, M., Neff, U., Kramers, J.,
Mangini, A., et al. (2003). Holocene forcing of the Indian monsoon
recorded in a stalagmite from southern Oman. Science, 300(5626),
1737.
Fleitmann, D., Cheng, H., Badertscher, S., Edwards, R. L.,
Mudelsee, E. M., Gdktiirk, O. M., et al. (2009). Timing and cli¬
matic impact of Greenland interstadials recorded in stalagmites
from northern Turkey. Geophysical Research Letters, 36, L19707.
doi:10.1029 /2009GL040050.
Frappier, A. B., Sahagian, D., Carpenter, S. J., Gonzalez, L. A., &
Frappier, B. R. (2007). Stalagmite stable isotope record of recent
tropical cyclone events. Geology, 35, 111—114.
Frumkin, A., Ford, D. C., & Schwarcz, H. P. (1999). Continental
paleoclimatic record of the last 170,000 years in Jerusalem.
Quaternary Research, 51, 317—327.
Genty, D., Blamart, D., Quahdi, R., Gilmour, M., Baker, A., Jouzel, J.,
et al. (2001). Precise dating of Dansgaard—Oeschger climate oscil¬
lations in Western Europe from stalagmite data. Nature, 421,
833-837.
Kutzbach, J. E. (1981). Monsoon climate of the Early Holocene:
Climate experiment with the Earth's orbital parameters for 9000
years ago. Science, 214(4516), 59—61.
Lachniet, M. S. (2009). Climatic and environmental controls on spe¬
leothem oxygen-isotope values. Quaternary Science Reviews, 28,
412-432.
Lawrence, J. R., Gedzelman, S. D., Gamache, J., & Black, M. (2002).
Stable isotope ratios: Hurricane Olivia. Journal of Atmospheric
Chemistry, 41(1), 67—82.
Polyak, V. J., & Asmerom, Y. (2001). Late Holocene climate and cul¬
tural changes in the southwestern United States. Science, 294
(5540), 148-151.
Rasmussen, J., Polyak, V., & Asmerom, Y. (2006). Evidence for
Pacific-modulated precipitation variability during the Late
Holocene from the southwestern USA. Geophysical Research
Letters, 33, L08701. doi:10.1029/2006GL025714.
Stuiver, M., Reimer, P. J., Bard, E., Beck, J. W., Burr, G. S.,
Hughen, K. A., et al. (1998). INTCAL98 radiocarbon age calibra¬
tion, 24000—0 cal BP. Radiocarbon, 40(3), 1041—1083.
Svensson, A., Andersen, K. K., Bigler, M., Clausen, H. B., Dahl-
Jensen, D., Davies, S. M., et al. (2005). The Greenland Ice Core
chronology 2005, 15—42 ka. Part 2: Comparison to 15 other
records. Quaternary Science Reviezvs, 25(23—24), 3258—3267.
Tan, M., Liu, T., Hou, J., Qin, X., Zhang, H., & Li, T. (2003). Cyclic
rapid warming on centennial-scale revealed by a 2650-year stalag¬
mite record of warm season temperature. Geophysical Research
Letters, 30(12), 1617. doi:10.1029/2003GL017352.
Wang, X., Auler, A. S., Edwards, R. L., Cheng, H., Ito, E., Wang, Y.,
et al. (2007). Millennial-scale precipitation changes in southern
Brazil over the past 90,000 years. Geophysical Research Letters, 34,
L23701. doi:10.1029/2007GL031149.
Wang, Y., Cheng, H., Edwards, R. L., He, Y., Kong, X., An, Z., et al.
(2005). The holocene Asian monsoon: Links to solar changes and
north Atlantic climate. Science, 308, 854—857.
Winograd, I. J., Coplen, T. B., Landwehr, J. M., Riggs, A. C.,
Ludwig, K. R., Szabo, B. J., et al. (1992). Continuous 500,000-year
climate record from vein calcite in Devils Hole, Nevada. Science,
258(5080), 255-260.
PALEOMAGNETIC RECORDS
IN CAVE SEDIMENTS
Ira D. Sasowsky
University of Akron
INTRODUCTION
The paleomagnetic record in cave sediments allows
us to answer one of the fundamental questions about
caves: "How old are they?" This is a hard question
because caves are essentially empty space created in
rock and, therefore, we are trying to determine the age
of something that is no longer there. Hence, we must
constrain the age of the cave by other strategies. We
know of two conditions that must always apply.
First, the cave must be younger (more recent) than
the rock in which it has formed. The only exception is
the case of lava tubes, where the cave is the same age
as the rock in which it was formed. However, the
majority of caves are formed in limestone, a long time
after the limestone rock was originally deposited. For
Encyclopedia of Caves.
© 2012 Elsevier Inc. All rights reserved.
586
p
example. Mammoth Cave in Kentucky is about 4 mil¬
lion years old, but the rocks in which it is found are
more than 300 million years of age. Because of this
large difference, dating a cave by the age of the rocks
that host it is not very satisfactory.
A second, and more useful, method to determine
the age of a cave is to date the deposition of materials
that we find within it. This can include things that
have been carried in by natural processes such as sedi¬
ments and speleothems. This provides a minimum age,
because the cave must have existed before the material
could have been placed in it. The paleomagnetism of
cave sediments is one such way to determine age.
Other methods (not addressed here) are uranium-
thorium radiometric dating of speleothems and cosmo-
genic isotope dating of quartz particles washed into
a cave.
EARTH MAGNETISM AND
PALEOMAGNETISM
To understand the paleomagnetism of cave sedi¬
ments, an understanding of Earth's magnetism and
paleomagnetism is required. It has long been known
that the Earth has a magnetic field. The field has sev¬
eral sources, but it is mainly generated in the core of
the Earth by rotation of the planet. Early natural scien¬
tists and explorers used magnetic minerals (lodestones)
to measure the direction of the field or to guide them
in their travels. This led to the development of com¬
passes, which are precision devices with a narrow
magnetic needle that points toward magnetic North
(N m ). The Earth also has geographic, or true. North (N)
and South Poles that are defined by the axis of rotation
of the Earth. The magnetic North and South Poles are
in the general region of the geographic poles and are
the areas to which a compass needle will point. In
addition to a north—south orientation (which is called
declination), the magnetic field has an up—down com¬
ponent (called inclination). At the North magnetic pole,
the inclination is +90° (straight down). At the magnetic
equator, the inclination is 0° (flat), and at the South
Pole it is —90° (straight up).
Measurements of the position of the magnetic poles
have been made since the early 1600s, and it has been
found that the magnetic pole position varies (drifts)
with time—a process called geomagnetic secular varia¬
tion, or simply polar drift. This drift is on the order of
0.05° per year. Then, in the 1900s, it was discovered
that the magnetic field had reversed many times
throughout geologic history. In these instances, rather
than gently drifting though time, the poles had actu¬
ally (and quite rapidly) flipped: The magnetic North
Pole moved to the geographic South Pole and vice
versa. These changes are called field reversals. When the
magnetic North Pole is in the Southern (geographic)
Hemisphere, the condition is considered to be one of
reverse polarity. When the North Pole is in the
Northern Hemisphere, as it is today, it is considered a
time of normal polarity.
Many rocks and sediments hold a weak magnetic
signal that is "locked in." The signal that they hold
points in the direction of magnetic North from the
time that they were deposited. Rocks deposited during
a time of reversed polarity have a magnetic signal that
points to the Southern Hemisphere. Through the study
of rocks worldwide, a chronology of the magnetic field
has now been constructed. This magnetostratigraphic
time scale (Fig. 1) shows the orientation of the mag¬
netic field during different portions of Earth's history.
Those times when the fields were reversed are indi¬
cated in white.
As a result of the creation of this worldwide time
scale, it is now possible to determine the age of other
rocks and sediments by measuring their magnetic
properties and placing them in their proper position in
the magnetostratigraphic time scale. Butler (1992) and
Tuaxe (2010) provide comprehensive reviews of the
topic.
PALEOMAGNETISM OF CLASTIC CAVE
SEDIMENTS
All streams carry small, broken up bits of rock (clas¬
tic sediment) that come from the weathering of the sur¬
rounding landscape. This includes fine materials such
as clay and coarser materials such as sand and gravel.
When water flow decreases, or stops altogether, these
materials are dropped (deposited) from the stream. As
the particles slowly settle from the water, tiny mag¬
netic mineral grains (primarily magnetite and hema¬
tite) orient themselves with the magnetic field. The
process is similar to a compass needle becoming ori¬
ented. When the particle reaches the stream bottom, it
is buried by additional particles, and the magnetic ori¬
entation of the grains is locked in. This is called detrital
remanent magnetism (DRM). If a sample of this material
is analyzed, the direction of the magnetic field during
the time of deposition can be determined.
Caves typically have a period of their existence dur¬
ing which a stream is flowing through them (Fig. 2). If
this stream comes from the surface of the Earth, it will
carry clastic sediments that can then be left in the
cave. Such deposits (Fig. 3) can range in thickness
from millimeters to tens of meters and can consist of
all grain sizes. Virtually all caves have some clastic
sediment. To measure the paleomagnetism, researchers
ENCYCLOPEDIA OF CAVES
PALEOMAGNETIC RECORDS IN CAVE SEDIMENTS
587
1.0 -
<D
<
O
*15
a>
g
,5
'S
2.0
3.0 -
4.0 -
5.0
Chron
<
2
on
on
o
p2
w
PQ
a
Subchron
Jaramillo
Cobb Mountain
Olduvai
Reunion
Kaena
Mammoth
Cochiti
Nrnhvak
Sidutjall
Thvera
FIGURE 1 Magnetostratigraphic time scale. The polarity of the
Earth's magnetic field has switched through time, with the North
magnetic pole alternating position near the North geographic and
South geographic poles. Courtesy of US. Geological Survey.
FIGURE 2 Typical stages in the evolution of a cave: (A) Initial
rock mass with open bedding plane, (B) enlargement of bedding
plane by groundwater circulation, (C) deposition of clastic sedi¬
ments, (D) deposition of speleothems (chemical sediments).
FIGURE 3 Clastic sedimentary deposit in Windy Mouth Cave,
West Virginia, U.S.A. A stream that once flowed through the cave
deposited these sediments.
collect a small, oriented sample of the material.
Duplicate samples and samples from many locations
and levels are usually obtained. Clays are the most
desirable material to sample, because they represent a
still-water environment, which allows for good orien¬
tation of the magnetic particles. But good results have
also been obtained from coarse, cross-bedded sands.
The samples are then taken to a laboratory and mea¬
sured in a magnetometer. This instrument determines
both the direction and strength (magnetic intensity) of
the sample. Oftentimes several steps of magnetic clean¬
ing, or demagnetization, are needed to reveal the DRM.
On an automated system, such measurement takes
about 30 minutes. The resulting plots, which show
magnetic direction and intensity (Fig. 4), tell whether
the sample has a normal or reverse polarity.
Now comes the tricky part: interpretation. The mag¬
netostratigraphic time scale is a binary one—it only
shows normal or reverse. If we have an isolated sam¬
ple of cave sediment that is normal, how can we know
which normal interval (chron) it is from? The short
answer is that we cannot—at least not with absolute
certainty. But we can make educated guesses with an
isolated sample. For example, if we find a reversed
polarity sample, we know that it must be at least
ENCYCLOPEDIA OF CAVES
588
p
DAS-UO
Tick interval = lCT’kAm 1
Up
N
S
Dozvn
Up
DAS-106
Tick interval = KPkAm' 1
Horizontal Component: • Alternating Field Demag
Vertical Component: a Geographic Coordinates
FIGURE 4 Orthogonal vector plot (Zijderveld diagram) of two
clastic sediment samples from caves. Each sample has two plots. The
lower sample has normal polarity, showing a northward declination
and downward inclination. The upper sample is reversed, and shows
a characteristic hooking, with southward declination and upward
inclination.
780,000 years old (the end of the last reversed interval;
also called the Matuyama chron). Likewise, if we find
an isolated deposit of normal polarity sediment, we
infer that it must be either less than 780,000 years old
or more than 990,000 years old. Using these con¬
straints, in context with other information such as
U-Th dates of speleothems or position in the land¬
scape, we can narrow down the age range of a cave.
The work of Victor Schmidt (1982) at Mammoth Cave,
Kentucky, U.S.A., was some of the first to date a cave
using this technique.
In many caves, deposits of clastic sediments are dis¬
continuous or complex to decipher. In these cases,
interpretation may be limited to that mentioned above.
However, in other cases, one or more "stacks" of sedi¬
ment may be found, and a fairly complete stratigraphy
may be constructed. When this occurs, the resulting
local magnetostratigraphy may be directly matched
with the global column (Fig. 5), and the age range of
the method is greater.
PALEOMAGNETISM OF
SPELEOTHEMS
Speleothems are deposits of secondary minerals
(such as calcite) that form on the ceilings, walls, and
floors of caves. Stalactites and stalagmites are the most
frequently studied of these features. As with the clastic
sediments discussed earlier, speleothems can retain a
record of the magnetic field from the time of their for¬
mation. In this case, however, the magnetism is a chem¬
ical remanent magnetism (CRM) rather than a DRM, and
the sampling and processing are slightly different.
Because speleothems are hard, drilling is required
to collect the oriented sample. This is done using a gas¬
oline-powered drill with a hollow, diamond-tipped bit.
Water is pumped on the bit to cool it and remove the
rock cuttings. It is a rather onerous procedure to carry
out in the difficult setting of a cave, and this may be
one reason why there have been relatively few such
studies. However, small cores have been taken, and a
project conducted by B. Ellwood collected very long
cores from large speleothems in Carlsbad Caverns,
New Mexico, U.S.A.
IMPORTANCE OF THE PALEOMAGNETIC
RECORD IN CAVES
In the preceding sections, the importance of the
magnetic record in caves with regard to determining
cave age has been discussed. Paleomagnetism provides
a relatively quick and inexpensive method for estimat¬
ing such ages. However, the method relies on establish¬
ing a complete record of polarity changes. Therefore,
difficulties arise due to discontinuous deposition or
erosion of the sediments.
The main advantage of paleomagnetism is that it
has a greater range than U-Th disequilibrium dating.
The former has been applied back to 4.5 million years,
whereas the latter has a range of about 450,000 years.
Paleomagnetism also requires less laboratory work.
Therefore, both methods have utility for the study of
caves.
ENCYCLOPEDIA OF CAVES
PALEOMAGNETIC RECORDS IN CAVE SEDIMENTS
589
Declination
O o © O o
JZ
>* at
« c
0)
c
>,2
~ -C
8
2 2
2%
O 3
a
O J=
LU
clo
CL (/>
Floor
Angular-
up to 50 cm contains silt,
clay, small (bat) and
large bones near base.
Appears to be artificial fill.
4 - 5 aaaKaAa/vaj
&
Limestone
<3j> ~ O
Green Clay Chert
Clay Intraclast Red Sandstone
FIGURE 5 Correlation of the magnetostratigraphic column from Cave of the Winds, Colorado, with the global record (Luiszer, 1994). An
apparently continuous record of deposition from 4.5 to 1.5 million years ago is shown. Courtesy of Fred Luiszer and the Karst Waters Institute.
Through age determination of caves, it is also possi¬
ble to relate the caves and other features to landscape
evolution, such as the incision of rivers. Therefore, the
study of caves, which may seem rather esoteric, serves
a purpose that is much broader.
The relatively new study of "environmental magne¬
tism" may provide an additional utility for cave
sediments. Rather than simply looking at reversals,
environmental magnetism examines the magnetic prop¬
erties of sediments to understand the environmental
conditions under which the materials were produced.
For example, Ellwood et al. (1998) identified glacial
effects on cave sediments. These methods, along with
continued development of our understanding of cave
processes, suggest that the magnetism of cave sedi¬
ments will continue to be an important research tool.
See Also the Following Articles
Clastic Sediments in Caves
Paleoclimate Records from Speleothems
ENCYCLOPEDIA OF CAVES
590
p
Bibliography
Butler, R. F. (1992). Paleomagnetism: Magnetic domains to geologic ter-
ranes. Boston: Blackwell Scientific Publications, www.geo.arizona
.edu/Paleomag/book/.
Ellwood, B. B., Zilhao, J., Harrold, F. B., Balsam, W., Burkhart, B.,
Long, G. J., et al. (1998). Identification of the last glacial maximum
in the Upper Paleolithic of Portugal using magnetic susceptibility
measurements in Caldeirao Cave sediments. Geoarchaeology, 13,
55-71.
Luiszer, F. G. (1994). Speleogenesis of Cave of the Winds, Manitou
Springs, Colorado. In I. D. Sasowsky & M. V. Palmer (Eds.),
Breakthroughs in karst geomicrobiology and redox geochemistry
(pp. 91—109). Charles Town, WV: Special Publication 1, Karst
Waters Institute.
Schmidt, V. A. (1982). Magnetostratigraphy of sediments in
Mammoth Cave, Kentucky. Science, 217, 827—829.
Tauxe, L. (2010). Essentials of paleomagnetism. San Diego: University of
California Press.
PALEONTOLOGY OF CAVES
Blaine W. Schubert and Jim I. Mead
East Tennessee State University
INTRODUCTION
Paleontology is the study of prehistoric life through
the analysis of fossils. Fossils are physical entities that
represent evidence of past organisms, and can be pre¬
served plant or animal remains, dung (Fig. 1), molds,
casts, or traces such as tracks and claw marks.
Remains do not have to be altered or mineralized to be
considered a fossil. Generally, caves are subterranean
openings or voids in rocks near the surface of the
earth. Once caves exist, organisms find their way into
these systems in many different ways, and some of
them avoid destruction by being cached in caves and
become fossils. Caves provide shelter and protection
from the primary elements of weathering, and this
combined with relatively stable temperatures and
humidity make subterranean environments an ideal
location for long-term preservation.
Geologically speaking, cave passages are short-lived
because the host rock is subjected to weathering and
erosion. For this reason, most fossils from extant caves
are relatively young, less than ~2 myr old. Thus, the
fossil record provided by active caves is essentially
equivalent to the Quaternary, the past ~1.8 myr
(Schubert et al., 2003). Older cave deposits do exist, but
the physical cave is typically gone or filled and the
sediments and fossils are often highly mineralized and
cemented, forming what are known as bone breccias.
These more ancient cave or karst features are usually
found during construction or quarrying operations
of bedrock. Karst deposits with fossils date back to
the late Paleozoic, with the earliest example being
limestone fissure fills in Iowa where the oldest
Mississippian-age vertebrate fauna of North America
is preserved (Lundelius, 2006).
Here we focus primarily on the Quaternary, with
some mention of the earlier Pliocene as well. It is
during the Quaternary that most of our existing travers¬
able caves formed and became accessible to vertebrates,
so the fossil record is rich and diverse. The earliest part
of the Quaternary is the Pleistocene, which ranges from
~1.8 million years ago until ~12,000 years ago. This
time is known as the Ice Age, and it is characterized by
extreme climatic oscillations, reflected in a series of gla¬
cial advances and retreats. Colder (glacial) and warmer
(interglacial) periods alternated, with short-term warm¬
ing during glacials (stadials) and cooling during inter¬
glacials (interstadials). The second part of the
Quaternary is the Holocene (or Recent)—the past
~12,000 years. Many geologists consider the Holocene
to be just another interglacial within the Ice Age.
Researchers have studied fossils from Quaternary
cave sites all over the world, particularly in regions
where limestone bedrock is common. Most research
has focused on mammals, although studies have also
occurred on arthropods, mollusks, fish, amphibians,
reptiles, and birds. In general, there has been a bias
toward studying the mammals because many of them
are large and extinct, and the teeth of mammals pre¬
serve well and are easier to identify to specific levels
than most fossil remains from other taxonomic groups.
In addition, mammal teeth can be highly informative
about diet and evolution. For these reasons, there has
been more interest in the mammals overall. It is impor¬
tant to note that the lack of studies on other animal
groups from cave deposits is not a reflection of relative
abundance. In fact, most cave collections in major
FIGURE 1 Rampart Cave, Grand Canyon, AZ, U.S.A. Floor of
cave covered in sloth dung. Photograph courtesy of the GRCA National
Park Museum Collections.
Encyclopedia of Caves.
© 2012 Elsevier Inc. All rights reserved.
PALEONTOLOGY OF CAVES
591
natural history museums house a plethora of nonmam¬
malian fossils just waiting to be studied.
Fossil deposits in caves vary widely, ranging from
carnivore dens to owl roosts, from natural traps to
dung deposits, and packrat middens to early hominid
sites. Hominids are the bipedal apes, which includes us.
Homo sapiens, others members of our genus, and species
of Ardipithecus, Australopithecus, and Paranthropus. Here
we do not cover human habitation in caves, and leave
this to the archaeologists and paleoanthropologists. The
overall diversity in site types indicates the potential com¬
plexity of interpreting each unique locality, and discern¬
ing which taphonomic processes contributed to the fossil
deposit. Taphonomy is the study of how fossils come
into being. In other words, what happened to the organ¬
ism from the time of death until it was discovered? This
aspect of paleontology is a crucial component in under¬
standing how fossils get into caves.
TYPES OF CAVE DEPOSITS
AND EXAMPLES
Natural Traps
Caves that have a vertical entrance can serve as nat¬
ural traps for organisms that fall in and die. Size of the
animals represented depends on the size of the
entrance. Natural traps are more effective at accumu¬
lating remains if they are deep and lack associated hor¬
izontal passages and entrances. These deposits are
characterized by relatively complete skeletons with lit¬
tle or no postmortem taphonomic damage, though
small rodents often find their way into these areas and
gnaw on bones. The trap accumulations are often scat¬
tered along a talus cone or rock debris slope, positioned
under the vertical entrance. If sedimentation occurs reg¬
ularly, skeletons become buried more quickly and will
be better preserved, more articulated, and a chronology
of deposition can be more easily determined. If sedi¬
mentation rate is low, more scattering and time averag¬
ing will occur.
One of the best examples of an extensive natural
trap deposit is Natural Trap Cave in north-central
Wyoming, North America. The pit entrance of 4 by
5 m is not visible until the edge, the drop is ~20 m,
and the cave is bell-shaped with no other entrances.
This pit has accumulated animals for at least the past
110,000 years, including remarkable Pleistocene mega¬
fauna such as the American cheetah ( Miracinonyx tru-
mani), American lion ( Panthera atrox), giant short-faced
bear ( Arctodus simus), dire wolf ( Canis dirus), bison
(Bison antiquus), woodland muskox ( Bootherium bombi-
frons), American camel ( Camelops ), extinct horses
(Equus ), and mammoth ( Mammuthus ) (Martin and
Gilbert, 1978). Other extraordinary examples of natural
traps are found in the Thylacoleo caves on the
Nullarbor Plain, south-central Australia. These caves
are named after the articulated and well-preserved
marsupial lion ( Thylacoleo carnifex) remains recovered,
but many other vertebrates occur in the deposits,
including new species of kangaroos.
Alluvial and Colluvial Transport
Many caves have active river systems (alluvial), and
these streams can transport carcasses or parts of skele¬
tons into caves, or move them from one place to
another in the cave system. This alluvial transport
often introduces animals that otherwise would not
venture into caves. For example. Powder Mill Creek
Cave, Missouri Ozarks, USA, contains numerous
remains of late Pleistocene semiarticulated deer. Based
on the associated sediments and adjacent stream, these
specimens were clearly transported to their current
location by water. The level of articulation indicates
the carcasses washed to these locations close to the
time of death.
Colluvial transport is down-slope movement of
sediments without the assistance of water. These types
of sediments can introduce fossils into caves through
fissure fills, sinkholes, or sloped entrances. Further,
existing deposits in caves can change over time as col¬
luvial action occurs, particularly when sediments are
on slopes. These sedimentary processes point out some
of the potential complexities of cave deposits.
Raptors
Here we refer to raptors, or birds of prey, in the
informal sense, meaning those birds that base their
diet on scavenging or hunting other vertebrates.
Certain raptors, such as owls, hawks, and vultures,
often utilize cave entrances as roosts and accumulate
bones. Owls are nocturnal hunters, while other raptors
hunt and scavenge during the day (diurnal); thus, their
prey base can differ dramatically. Raptors do not digest
the entire carcass; instead they regurgitate bones and
hair (or scales) as pellets. Over time, significant accu¬
mulations can develop below roosts. Owl roosts con¬
tain the most prolific and undamaged bones because
they swallow most prey relatively whole and cast their
pellets from a more alkaline digestive system. Diurnal
birds of prey have varying degrees of higher acidity in
their precasting digestion, and therefore dissolution of
bone occurs in most species. Also, some of these rap¬
tors tear their prey or scavenged remains into pieces,
further disturbing the skeletal material. These differen¬
tial characteristics, as well as prey size, allow
ENCYCLOPEDIA OF CAVES
592
p
paleontologists to discern the type of raptor responsible
for certain fossil deposits (Andrews, 1990).
Skeletal remains collected by raptors and deposited
in caves play a significant role in our understanding of
the distribution and evolution of small mammals dur¬
ing the late Cenozoic. An example of an extensive rap¬
tor roost is Screaming Neotoma Cave in Arizona,
along the southern border of the Colorado Plateau,
western North America. This cave is a natural trap pit
where owls have roosted on and off since the late
Pleistocene; thus, at least two different processes are
depositing fossils. The deposit has only received mini¬
mal excavations, but species such as the sagebrush
vole ( Lemmiscus curtatus ) indicate a geographical change
in distribution; this species now occurs much farther
north. Climatic change and habitat loss are possibilities
for this biogeographic change. Other examples of
extraordinary raptor roosts in southwestern North
America are found in the Red Wall Limestone of the
Grand Canyon, Arizona. Here condors ( Gymnogyps ) once
roosted and accumulated the remains of scavenged ani¬
mals during the late Pleistocene. Feathers, skulls, kera¬
tin, other tissues, and bones are exquisitely preserved
because of persistent arid conditions inside and outside
the caves.
Bats
Bats are a widespread group of vertebrates that
have been around since at least the early Cenozoic.
They are the only mammals with true flight, and this
is accomplished by "hands" that are covered in skin
membranes that form the wings. With over 1000 living
species bats, are one of the most diverse mammalian
orders. The majority (>70%) are relatively small insec-
tivores that hunt at night and roost during the day. All
bats can see, but night hunters rely on echolocation for
navigation and finding their prey.
When caves are available bats use them, particularly
for daytime roosting, rearing young, and winter torpor
in colder regions. Thus, many species are closely inter¬
twined with cave environments and have evolved
adaptations and behaviors that reflect this relationship.
Because bat bones are small and fragile, their fossil
record is sparse for most of their existence. However,
Pleistocene and Holocene deposits from caves are pro¬
lific and can give insight into old roosting sites, past
distributions of extant species, and climatic and envi¬
ronmental change. Since many living bat species have
restricted temperature and humidity requirements for
roosts, these ecological parameters are sometimes
applied to Quaternary fossils to better understand cave
paleoecology.
Not surprisingly, the primary fossil localities for
bats are found under abandoned roost sites in caves.
Skeletal remains are intermixed with piles or layers of
bat guano if the roost was used during a time of feed¬
ing. If a roost represents a maternity colony, a high
proportion of the bat remains will be juveniles. On the
other hand, if a deposit is rich in skeletal remains but
lacks an abundance of guano, it may represent a win¬
ter die-off during torpor.
Bat deposits in caves are numerous but under-
studied. The extensive Mammoth Cave system in
Kentucky, USA, provides a rare example of a thorough
survey for bat fossils. The oldest bat known from the
cave belongs to a single vampire bat ( Desmodus ) speci¬
men, and was found in sediments in the upper reaches
of the cave, dated over 1 myr old. Desmodus today pre¬
fers warmer temperatures in the more tropical regions
of the Americas, and its occurrence from Pleistocene
sites in North America indicates warmer periods than
today. One of the most extensive deposits in
Mammoth Cave is of free-tailed bats ( Tadarida ), a
genus that now has a distribution to the southwest.
This deposit is thought to represent an interglacial
when it was warmer than today in Kentucky. Tadarida
are known to form huge colonies; for example.
Bracken Cave in Texas houses over 20 million indi¬
viduals during part of the year. The Historic Entrance
section of Mammoth Cave provides a more recent
example of environmental change and its impact on
bats. Small brown bats of the genus Myotis inhabited
this region of the cave in vast numbers for roosting
during historic times. The primary area of the colony
is documented in early journals, and as guano, bat
bones, and staining on the ceiling. A survey of this
entire region of the cave indicates that modifications to
the main entrance altered temperature and humidity
patterns, making the roosting site no longer adequate.
There are other ways that bat remains get incorpo¬
rated into fossil sites. Some species roost alone or in
small groups and die while roosting. Others become
incorporated into deposits by carnivores such as rap¬
tors or mammals. For example, raccoons are known to
venture back into caves and eat bats while they are
hibernating. Raccoon scat, primarily composed of bat
remains in Mammoth Cave, provides evidence for this
behavior in the fossil record. In addition, some bats are
carnivorous and contribute remains of their prey into
cave deposits. An example is the carnivorous ghost
bat. Macroderma gigas, of Australia.
Large Carnivorans
Many species of large carnivorans (Order Carnivora)
utilized caves in the past for denning or other purposes.
Carnivorans are not all carnivores; carnivore is an eco¬
logical term meaning the diet is composed mostly of
ENCYCLOPEDIA OF CAVES
PALEONTOLOGY OF CAVES
593
meat. Evidence of cave use ranges from claw marks to
bedding areas, and from foot-prints and coprolites, to
chewed-up prey and the remains of carnivorans them¬
selves. Extinct Pleistocene carnivorans that used caves
include giant bears, hyenas, wolves, and cats of differ¬
ent types. Indeed, it would not have been the best time
to be a caver. Examples of living large carnivorans that
use caves include the American black bear ( Ursus ameri-
canus), Asian black bear ( Ursus thibetanus), giant panda
(Ailuropoda melanoleuca) , lion ( Panthera atrox), leopard
(Pathera pardus), cougar ( Puma concolor), and spotted
hyena (Crucuta crocuta).
Bears
The most abundant fossils of a cave-dwelling carni-
voran are from the cave bear ( Ursus speleus), making
it one of the best-known Ice Age mammals. This spe¬
cies appeared around 300,000 years ago and went
extinct near the end of the Pleistocene. Its distribution
was limited to Europe, and ranged from sea level to
alpine areas. The cave bear was similar to the living
Holarctic brown bear ( Ursus arctos), but its limbs were
relatively shorter, its body was more barrel-like, and
its teeth were more broad and blunt. It was the largest
carnivoran from the Pleistocene of Europe and some
caves contain thousands of their bones. These were
not mass die-offs, but represent repeated use of caves
for hibernation and denning over extended periods of
time. The remains that are found in caves are usually
of juvenile, old, or diseased animals. Males were
much larger than females and sexual differentiation
in cave use may have occurred. For example, in
Drachenhohle (Dragons Cave) near Mixnitz, Austria,
~30,000 individuals were estimated to be in the
deposit, and the male specimens outnumber the
females by 3 to 1. In other sites, the females and asso¬
ciated juveniles are more abundant. Other evidence of
the cave bear's use of caves is found in polished
cave walls, claw marks, and tracks (Kurten, 1968;
Kowalski, 2005).
In North America, bear fossils are often found in
caves but are never represented by the same abun¬
dance as the cave bear in Europe. Bears, particularly
black bears ( Ursus americanns) , represent the most com¬
monly found articulated remains of large carnivorans
in North American cave deposits. Although the geo¬
logic age of most of these U. americanus has not been
determined, a great number of them likely represent
the Holocene epoch. In addition to their bones and
teeth, bear beds, claw marks, and footprints are rela¬
tively common. The brown bear, Ursus arctos, is not
typically associated with cave deposits, but numerous
records exist for Alaskan Caves.
The tremarctine bears (Subfamily Tremarctinae) are
known exclusively from the Americas and a number
of species ( Tremarctos floridanus, Arctodus pristinus, and
A. simus) are found in cave deposits in North America.
Only one living tremarctine exists, the spectacled bear
(Tremarctos ornatus) of South America. The extinct
Florida cave bear (T. floridanus) was a large robust
Plio-Pleistocene species that is known from the south¬
ern United States, Mexico, and Central America. This
bear derives its name from an abundance of cave sites
in Florida, where it is often found co-occurring with
U. americanus. It has been proposed that T. floridanus
was primarily a vegetarian like U. speleus (Kurten and
Anderson, 1980). Reports of T. floridanus existing until
~8000 years ago (Kurten and Anderson, 1980), based
on conventional radiocarbon dates from Devil's Den in
Florida, should be questioned. While this species cer¬
tainly went extinct around the end of the Pleistocene,
no reliable radiocarbon dates on the species have been
reported.
The lesser short-faced bear ( Arctodus pristinus) occurs
in some Pleistocene cave and karst deposits in eastern
North America, but it is not abundant. By the late
Pleistocene, A. pristinus was completely replaced by
A. sunus, the giant short-faced bear, which had a wide
distribution from Alaska to Mexico, and coast to coast.
The giant short-faced bear was larger, had a relatively
shorter face, more crowded teeth with some propor¬
tional differences, and longer limbs for its size (Kurten
and Anderson, 1980). These bears were highly sexually
dimorphic, with males being much more massive than
females. This is a trend that becomes more pronounced
among larger bears. Some estimates have placed the
largest A. simus males around 1000 kg. Well over 100
fossil localities are known for A. simus, and many of
these are cave sites. Interestingly, none of the large¬
sized specimens, interpreted to be males, are found in
cave deposits but 70% of the smaller, presumably
female, adult specimens are from cave sites (Fig. 2).
This has been interpreted to indicate denning among
the females (Schubert and Kaufmann, 2003). Dietary
studies on this bear indicate that it was highly carnivo¬
rous in some portions of its range, but additional stud¬
ies are needed to better understand variation across
space and time. Arctodus simus is one of the most
thoroughly dated late Pleistocene species in North
America, and these radiocarbon dates indicate that it
overlapped temporally and spatially with humans and
may have been one of the last megafaunal species on
the continent to go extinct (Schubert, 2010).
In South America, five different species of extinct
tremarctines are recognized, all within the genus
Arctotherium. Overall their fossil record is relatively
sparse, but some cave occurrences exist. The most
noteworthy is three specimens of the South American
ENCYCLOPEDIA OF CAVES
594
FIGURE 2 Giant short-faced bear forelimbs in Big Bear Cave,
MO, U.S.A. Photo courtesy of James E. Kaufmann. Used with permission.
giant short-faced bear Arctotherium angustidens, recov¬
ered during quarry operations in Buenos Aires Province,
Argentina. These specimens are interpreted to represent
a sow and cubs that perished while denning (Soibelzon
et ah, 2009). This species represents the largest type of
bear ever known, with males as large as 1600 kg. In con¬
trast to North America, where one of the youngest (or
latest) tremarctines was the largest ( Arctodus simus), in
South America it was the earliest known species
(Arctotherium angustidens) that was monstrous, and
apparently more carnivorous. This gigantic bear went
extinct in the middle Pleistocene and was replaced by
smaller and more omnivorous Arctotherium species,
some of which lasted until the end of the Pleistocene
(Soibelzon and Schubert, 2011).
The fossil record of giant pandas ( Ailuropoda ) comes
from cave deposits in China and Southeast Asia, where
three extinct species are recognized ( Ailuropoda microta ,
Ailuropoda wulingshanensis, and Ailuropoda baconi). The
oldest remains are from A. microta, and date to the late
Pliocene. The first discovered skull of this species,
from Jinyin Cave, south China, demonstrates that cra¬
nial and dental adaptations associated with bamboo
feeding were apparently in place by the late Pliocene
(Jin et ah, 2007).
Hyena
Fossil hyenas from caves sites in the Old World are
relatively common. Perhaps the most famous discov¬
ery of cave hyenas is from the Kirkdale Cave deposit
in North Yorkshire, England, described by Rev.
William Buckland in the early 1820s. The reverend ini¬
tially set out to show how the deposit was formed
by a Biblical deluge, but his thorough analysis in¬
stead demonstrated that the bone accumulation.
containing an abundance of hyena remains of all ages,
and the broken and gnawed bones of extinct hippo
(Hippopotamus ), woolly rhinoceros ( Coelodonta ), and
other megafauna, was the result of denning hyenas.
He also demonstrated that these hyenas were consider¬
ably larger than living species and that coprolites
found in the cave were from hyenas (Buckland, 1823;
Kurten, 1968). The cave hyena went extinct near the
end of the Pleistocene. While some morphological dif¬
ferences are known between this hyena and living spe¬
cies, there is disagreement over whether or not it is a
distinct subspecies or species or represents a larger
variant of the extant spotted hyena ( Crocuta crocuta ).
Another historically significant record of hyenas
as bone accumulators can be traced to South
African early hominid sites. Dr. Raymond Dart, the
anatomist who described the first Pliocene hominid,
Australopithecus africanus, from a South African bone
breccia cave deposit, was ostracized for his claims that
this was a human ancestor. Dart recognized and
described the specimen, known as the Taung Child, as
being bipedal with a brain the size of a chimpanzee's
brain. This characterization of human ancestors was
not to the liking of some British scientists, who con¬
ceived that the large brain would have predated biped-
alism, and that it should be found in Europe. In
the 1940s, Dart began to work on other South African
Australopithecus sites, particularly Makapansgat
Limeworks Cave where the broken up remains of
mammals were found in abundance. He imagined that
this hominid was a vicious killer and cannibal, and
that the remains in the caves were brought there by
members of this species. He termed this the osteodonto-
keratic (bone, tooth, and claw) culture and outlined his
ideas in a series of papers. Other researchers noted the
similarity in the fragmentary faunal remains to hyena
dens and Dart vehemently denied this possibility. The
fauna has now been carefully analyzed and is consid¬
ered to be a carnivore den, primarily formed by hye¬
nas. Thus, in this case the hominids appear to be the
hunted, not the hunters (Brain, 1981).
Wolves
Of the wolves, the extinct dire wolf ( Canis dims) is
the only species with an exceptional fossil record from
cave sites. This species was widespread, ranging across
North America and down into South America. Though
it was not necessarily taller than a large gray wolf on
average, it was more robust with a large broad head
and massive cheek teeth. Based on these morphologi¬
cal attributes, some researchers have suggested that it
was hyena-like in its behavior (Kurten and Anderson,
1980). Eastern and western populations in North
America differed significantly in proportions and
weight, with the eastern form being larger and heavier.
ENCYCLOPEDIA OF CAVES
PALEONTOLOGY OF CAVES
595
Most dire wolf sites are from caves and are thought to
represent dens. Examples of sites that have been inter¬
preted as dire wolf dens from caves or karst deposits
include Guy Wilson Cave in Tennessee and Cutler
Hammock in Florida. In addition to relatively intact
C. dints remains, these cave sites are littered with the
chewed-up skeletons of other vertebrates. In Guy
Wilson Cave, the most abundant prey items are flat¬
headed peccaries ( Platygonns compressus ) and deer
( Odocoileus ).
Cats
During the Pleistocene more types of large cats
existed on the landscape, and some extant species had
much larger geographic distributions. In Europe there
were lions larger than the African lion ( Panthera leo),
and these are sometimes referred to a distinct species,
the cave lion ( Panthera spelaea). Specimens of P. spelaea
from caves are usually not abundant, but there are
exceptions, such as Wierzchowska Cave, Poland,
where the remains of ~20 individuals were recovered
(Kurten, 1968). Similarly, in North America there was
a large cat known as the American lion ( Panthera
atrox). Its taxonomic position is also debated, and it
has been found in few cave localities. However, large
cat tracks from a Missouri cave (Fig. 3) have been
referred to this species (Graham et al., 1996). Leopards
(Panthera pardns) and jaguars (Panthera onca) ranged
much farther north during the Pleistocene; leopards
into Eurasia and jaguars over much of the United
States. While neither is particularly abundant from
cave sites, leopards are known for collecting remains
and taking them into caves. Thus, some faunal cave
deposits, for example, the South African hominid sites,
may have been formed in part by leopards. Jaguar
remains are found at Pleistocene sites across North
America, and these range from individual skeletons
or trackways that represent cave use, to specimens
that are part of faunal accumulations (Kurten and
Anderson, 1980). To date there are no fossil bone accu¬
mulations in caves attributed to jaguars.
The sabertooth cat, Smilodon, is perhaps the most
iconic of all American Pleistocene mammals; however,
its occurrence in caves is relatively low and there is no
evidence of denning and bone accumulation. However,
a record from Hurricane River Cave, Arkansas, pro¬
vides evidence of cave exploration by a sabertooth,
which unfortunately for the cat resulted in death at the
bottom of a pit. A related species, the scimitar cat
(Homotherium serum) seems to have had a closer associ¬
ation with caves. In contrast to Smilodon, Homotherium
serum had shorter sabers that are serrated on the front
and back, other teeth that were serrated, and excep¬
tionally long forelegs. Friesenhahn Cave in Texas
preserves multiple H. serum skeletons of varying
FIGURE 3 One of many large cat tracks from a cave in the south¬
ern Ozarks of Missouri, USA. The greatest width of this track is
~175 mm, and it is thought to be from the extinct American lion
(Panthera atrox) (Graham et at., 1996). Photo courtesy of James E.
Vandike. Used with permission.
individual ages, indicating that it was used as a den by
the cats. Associated with these remains are a multitude
of mammoth ( Mammuthus) and mastodon ( Mammut )
milk teeth and other bone fragments. Thus, it has been
hypothesized that these scimitar cats focused their
hunts on young proboscideans (elephants and rela¬
tives) and brought their prey back to the cave (Kurten
and Anderson, 1980). As with other Pleistocene mega¬
fauna, it is thought that these cats went extinct near
the end of the epoch.
Rodents
Some rodents are collectors and build dens and
debris piles (middens) within caves. These middens
are accumulations of the surrounding biota, including
plants, coprolites, bones, and teeth of whatever they
can carry. The collections are continually covered in
their own feces and urine, and in arid regions this cre¬
ates crystallized and cemented masses that can pre¬
serve for tens of thousands of years. In North America
the appropriately named packrats (or woodrat, genus
Neotoma) are well known for this peculiar collecting
habit, and studies on their middens are common. The
paleoecological record in these middens has provided
unique and significant information about climatic, eco¬
logical, and faunal changes that have occurred over
the past 40,000 years. In fact, without these packrat
midden deposits, there would not be a faunal
record for Quaternary reptiles and amphibians in the
American Southwest. Other rodents in South America,
Africa, and Australia make similar nests in caves that
can preserve fossils but none contains such diverse con¬
tents as those made by the packrat (Lundelius, 2006).
ENCYCLOPEDIA OF CAVES
596
p
Another type of rodent bone collector is the porcu¬
pine, both New World ( Erethizon ) and Old World
(Hystrix ) types. In Africa it seems fairly common for
porcupines to collect bones, store them in caves, and
gnaw on them periodically. Like other rodents, gnawing
on hard substances such as bones keeps the incisors
from getting too long and provides nutrients. There is
the potential that significant fossil accumulations in
caves exist from porcupine hoarding behavior (Brain,
1981). Porcupine dens are easy to recognize based on
the size and shape of gnawing marks on the bones.
Other Vertebrate Sites
Many other types of vertebrates contribute to cave
deposits. For example, small carnivores such as the
raccoon ( Procyon lotor), coati (Nasna nasua), ringtail
(Bassariscus astutus), skunk ( Mephitis mephitis), marten
(Martes martes), and red fox ( Vulpes vulpes), all use
caves to some extent today. Some use them as dens,
while others specifically seek out cave-based food,
such as raccoons and ringtails pursuing roosting bats.
Small carnivores often consume prey that is the same
size that raptors eat, but their accumulations are
marked by higher levels of bone fragmentation, some
evidence of digestive dissolution, and often intact
coprolites that contain an abundance of fragmented
bones (Andrews, 1990). There is no question that small
carnivores such as these have contributed substantially
to the fossil record in caves. In western Australia, the
Tasmanian devil ( Sacrophilus harrisi), a carnivorous
marsupial, used caves as lairs and accumulated mas¬
sive amounts of faunal material over time during the
Pleistocene (Lundelius, 2006).
In dry caves extraordinary preservation of soft tis¬
sues, hair, feathers, keratin, and dung occur. North
American caves of the Desert Southwest are particu¬
larly notable for these deposits. Many types of large
herbivores used cave entrances or large rock shelters
for latrines, and in some cases massive amounts of
dung accumulated over time. Dung attributed to
extinct mammals, such as the Shasta ground sloth
(Nothrotheriops shastensis), Harrington's mountain goat
(Oreamnos harringtoni), shrub ox ( Euceratherium ), mam¬
moth ( Mammuthus ), and horse {Equus), all give insight
into the diet of these animals and their paleoecology
(Mead and Agenbroad, 1992). The most extensive
deposit of fossil N. shastensis dung was found in
Rampart Cave, Grand Canyon, Arizona (Fig. 1).
Extensive deposits of mammoth dung are known from
Bechan Cave and Cowboy Cave in Utah, and indicate
a diet composed primarily of grass and sedges (Mead
et ai, 1986a). Dung from the extinct mountain goat,
O. harringtoni, is found throughout the entrances of
caves along the cliffs of the Grand Canyon, attesting to
the precipitous lifestyle of this species.
One of the earliest examples of soft tissue preserva¬
tion for an extinct Pleistocene mammal came from
Mylodon Cave in Chile, where skin, hair, and dung of a
mylodontid sloth was recovered. In arid western North
America, skin, hair, keratin claw sheaths, cartilage, and
other soft tissues are known from N. shastensis remains
from caves. The best known of these is from a lava
tube, Aden Crater, in New Mexico, where a relatively
complete mummified specimen was recovered. Horn
sheaths from bovids, particularly Oreamnos harringtoni
(Fig. 4), are also well known from cave sites on the
Colorado Plateau (Mead et ah, 1986b). In eastern North
America, caves are typically wet, which is not condu¬
cive to soft tissue and coprolite preservation. One exam¬
ple of an exception to this is Big Bone Cave, Tennessee,
where keratin and cartilage are preserved on Jefferson's
ground sloth ( Megalonyx jeffersonii) remains. This
includes keratinous claws (Fig. 5), intervertebral discs,
and other soft tissue components.
Two extinct ground sloths in North America,
N. shastensis and Megalonyx jeffersonii, are often found
in caves as relatively intact skeletons without any car¬
nivore damage. Thus, it has been proposed that these
sloths used caves for torpor, shelter, and / or maternity
denning (McDonald, 2003). One cave in northern
Alabama, ACb-3 Cave, contains multiple M. jeffersonii
skeletons of varying ages, including one female that
was apparently pregnant. This site has more known
M. jeffersonii than any other site and supports the
hypothesis that this species utilized caves for denning.
One of the most abundant Pleistocene mammals
found in caves is the flat-headed peccary, Platygonus
compressus. In some cases this peccary clearly found its
FIGURE 4 Extinct Pleistocene Harrington's mountain goat with
preserved horn sheaths from a Grand Canyon Cave, AZ, U.S.A.
Photo courtesy of Lorry L. Coats. Used with permission.
ENCYCLOPEDIA OF CAVES
PALEONTOLOGY OF CAVES
597
FIGURE 5 Jefferson's ground sloth (Megalonyx jeffersonii) claw
with preserved keratinous nail from Big Bone Cave, TN, U.S.A.
Photograph taken by Ted Daeschler, courtesy of the Academy of Natural
Sciences, Philadelphia.
way into faunal deposits as the prey of carnivores,
such as Canis dims. However, there are many sites
where skeletons are intact and abundant, and speci¬
mens of all ages are represented. The most extraordi¬
nary example is Megenity Peccary Cave, Indiana,
where hundreds of individuals have been collected.
Extensive peccary trackways are also known from cave
sites, for example, the clay floor in River Bluff Cave,
Missouri, is covered in hundreds, if not thousands, of
their tracks. The abundance of fossil remains and
tracks of P. compressus in caves indicates that they used
caves, probably for shelter, and they apparently lived
in large groups. Caves in southwestern North America
are actually used today by the living collared peccary,
Pecari tajacu, as retreats from harsh weather.
IMPORTANCE OF FOSSIL VERTEBRATES
FROM CAVES
Caves provide one of the primary sources of faunal
remains from the Quaternary fossil record and these
fossils are often well preserved (Schubert et al., 2003;
Lundelius, 2006; Jass and George, 2010). Thus, the
record from caves provides a crucial component for
understanding the chronology, distribution, ecology,
evolution, and extinction of vertebrates during this rel¬
atively recent, yet exciting, time in Earth's history. In
sum, this was a period of profound climatic and envi¬
ronmental changes, and the Pleistocene still holds
many mysteries, like how did biological communities
react to climatic changes, and when and why did the
megafauna go extinct?
Near the end of the Pleistocene epoch, most large
mammals in North America and South America
became extinct. Megafaunal extinctions also occurred
in Eurasia and Africa but were minimal in compari¬
son, and Australia's late Pleistocene extinction appears
to have occurred earlier than other continents. Many
of these species are found in cave deposits, and in
some cases caves represent the primary source for fos¬
sils. This is especially true for carnivorans that used
caves as dens, like the scimitar cat, dire wolf (outside
of Rancho la Brea), European cave bear. North
American giant short-faced bear, Florida cave bear,
and cave hyena. In addition, particular ground sloths,
the flat-headed peccary, and Harrington's mountain
goat are well known because of their fossil records in
caves. Besides providing relatively complete speci¬
mens that allow the study of different age individuals,
a growth series, the record in caves gives a unique
paleoecological view into cave use for torpor, shelter,
and denning. Further, the high level of preservation of
remains from caves has contributed to better radiocar¬
bon chronologies of species, as well as enhanced
knowledge of dietary preferences and evolutionary
relationships.
Most fossil sites that are excavated today include
careful screenwashing procedures to recover microverte¬
brates, which typically entered the site as prey. Small
mammal remains from screenwashing, particularly
rodent teeth, have received the most attention and
research for three primary reasons: (1) the teeth are typi¬
cally well preserved and are often distinct at the species
level, (2) they have high reproductive and evolutionary
turnover, and (3) they are sensitive to climatic and envi¬
ronmental changes. Because small rodents evolve
quickly, they are used as biostratigraphic markers, and
assist in age correlation from site to site. However, cau¬
tion must be applied in making such age assignments
because evolutionary turnover can be time transgressive
over space, and remnant populations of earlier species
sometimes survive in refugia. An example of this is
found in Pleistocene deposits from Cathedral Cave,
Great Basin, USA. Two extinct voles from the site,
Microtus meadensis and M. paroperarius, represent chro¬
nologic range extensions of ~100 ka years. Thus, there
is a possibility that this area was a refugium for those
species (Jass, 2007).
While many large mammals went extinct at the end
of the Pleistocene, small species survived but tended
to change their distributions based on ecological pref¬
erences. This knowledge is based almost exclusively
on cave faunas. For example, several voles, shrews,
and moles that now occur much farther to the north
are found in Pleistocene cave sites across the south-
central and southeastern United States. In these
"southern" cave sites "northern" species are found
associated with taxa that now occur farther west, and
species that still live in the area of the cave; thus, the
faunal accumulation is composed of mammals that no
longer co-occur and prefer very different environmen¬
tal settings. These accumulations are known as
ENCYCLOPEDIA OF CAVES
598
p
nonanalog faunas and are thought to reflect climatic and
ecological conditions that are quite different from
today. Recent advances in radiocarbon dating pro¬
vided a means to test this hypothesis, and dates on
nonanalog taxa show that these associations represent
true past communities.
See Also the Following Article
Vertebrate Visitors in Caves—Birds and Mammals
Bibliography
Andrews, P. (1990). Owls, caves and fossils. London: Natural History
Museum Publications.
Brain, C. K. (1981). The hunters or the hunted? An introduction to
African cave taphonomy. Chicago, IL: University of Chicago Press.
Buckland, W. (1823). Reliquiae diluviame; or, observations on the organic
remains contained in caves, fissures, and diluvial gravel, and on other
geological phenonmena, attesting the action of an universal deluge.
London: Thomas Davison, White Friars.
Graham, R. W., Farlow, J. O., & Vandike, J. E. (1996). Tracking ice
age felids: Identification of tracks of Panthera atrox from a cave in
southern Missouri, U.S.A. In K. M. Stewart & K. L. Seymour
(Eds.), Palaeoecology and palaeoenvironments of late cenozoic mam¬
mals: Tributes to the career of C.S. (Rufus) Churcher (pp. 331—345).
Toronto, Canada: University of Toronto Press.
Jass, C. N. (2007). New perspectives on Pleistocene biochronology
and biotic change in the east-central Great Basin. Unpublished
PhD dissertation. The University of Texas at Austin, Austin, TX.
Jass, C. N., & George, C. O. (2010). An assessment of the contribution
of fossil cave deposits to the Quaternary paleontological record.
Quaternary International, 217, 105—116.
Jin, C., Ciochon, R. L., Dong, W., Hunt, R. M., Jr., Liu, J., Jaeger, M.,
& Zhu, Q. (2007). The first skull of the earliest giant panda.
Proceedings of the National Academy of Sciences, 104, 10932—10937.
Kowalski, K. (2005). Paleontology of caves: Pleistocene mammals.
In D. C. Culver & W. B. White (Eds.), Encyclopedia of caves (1st
ed., pp. 431—435). Burlington, MA: Elsevier, Academic Press.
Kurten, B. (1968). Pleistocene mammals of Europe. London: Weidenfield
and Nicolson.
Kurten, B., & Anderson, E. (1980). Pleistocene mammals of North
America. New York: Columbia University Press.
Lundelius, E. L. (2006). Cave site contributions to vertebrate history.
Alcheringa Special Issue, 1, 195—210.
Martin, L. D., & Gilbert, B. M. (1978). Excavation at natural trap
cave. Transactions of the Nebraska Academy of Science, 6 , 107—116.
McDonald, H. G. (2003). Sloth remains from North American caves
and associated karst features. In B. W. Schubert, J. I. Mead, &
R. W. Graham (Eds.), Ice age cave faunas of North America
(pp. 1—16). Bloomington, IN: Indiana University Press.
Mead, J. I., & Agenbroad, L. D. (1992). Isotope dating of Pleistocene
dung deposits from the Colorado Plateau, Arizona, and Utah.
Radiocarbon, 34, 1—19.
Mead, J. I., Agenbroad, L. D., Davis, O. K., & Martin, P. S. (1986a).
Dung of Mammuthus in the arid Southwest, North America.
Quaternary/ Research, 25, 121—127.
Mead, J. I., Martin, P. S., Euler, R. C., Long, A., Lull, A. J. T.,
Toolin, L. J., Donahue, D. J., & Linick, T. W. (1986b). Extinction
of Harrington's mountain goat. Proceedings of the National
Academy of Sciences USA, 83, 836—839.
Schubert, B. W. (2010). Late Quaternary chronology and extinction of
North American giant short-faced bears (Arctodus simus).
Quaternary International, 217, 188—194.
Schubert, B. W., & Kaufmann, J. E. (2003). A partial short-faced bear
skeleton from an Ozark Cave with comments on the paleobiology
of the species. Journal of Cave and Karst Studies, 65, 101—110.
Schubert, B. W., Mead, J. I., & Graham, R. W. (Eds.), (2003). Ice age
cave faunas of North America. Bloomington, IN: Indiana University
Press.
Soibelzon, L. H., Pomi, L. M., Tonni, E. P., Rodriguez, S., & Dondas, A.
(2009). First report of a short-faced bears' den ( Arctotherium
angustidens ): Palaeobiological and palaeoecological implications.
Alcheringa, 33, 211-223.
Soibelzon, L. H., & Schubert, B. W. (2011). The largest known bear,
Arctotherium angustidens, from the early Pleistocene Pampean
region of Argentina: With a discussion of size and diet trends in
bears. Journal of Paleontology, 85, 69—75.
PASSAGE GROWTH AND
DEVELOPMENT
Arthur N. Palmer
State University of New York
Of all cave types, solution caves have the most com¬
plex developmental histories. They are formed by the
dissolving action of underground water as it flows
through fractures, partings, and pores in bedrock.
Such caves must grow rapidly enough to reach travers¬
able size before the rock material that contains them is
destroyed by surface erosion. Because of their sensitiv¬
ity to local landscapes and patterns of water flow,
solution caves contain clues to the entire geomorphic,
hydrologic, and climatic history of the region in which
they are located. At the land surface most of this evi¬
dence is rapidly lost to weathering and erosion, but in
caves these clues can remain intact for millions of
years.
STAGES OF CAVE DEVELOPMENT
Solution caves develop in several stages that grade
smoothly from one to the next: (1) For the rock to trans¬
mit enough water to form caves, it must first contain a
network of presolutional openings, such as fractures,
partings, and primary pores. (2) Because these openings
are narrow, the initial groundwater moves very slowly
and becomes nearly saturated after only a short dis¬
tance of travel. Solutional widening of the openings
is likewise very slow. Nevertheless, among the many
alternate routes that the water follows, there is a great
variation in the amount of flow because of differences
in width and length of openings. Those paths with the
greatest flow enlarge the fastest. (3) These favored
routes eventually become wide enough that
Encyclopedia of Caves.
© 2012 Elsevier Inc. All rights reserved.
PASSAGE GROWTH AND DEVELOPMENT
599
groundwater is able to pass all the way to the spring
while still retaining much of its solutional capacity. (4)
Flow along these preferred routes increases rapidly,
enlarging them into cave passages rather uniformly
over their entire length. (5) The cave acquires a distinct
passage pattern that depends on the nature of ground-
water recharge, geologic setting, and the erosional his¬
tory of the region. (6) Most caves evolve by diversion of
water to new and lower routes as surface rivers deepen
their valleys and lower the elevations of springs. (7)
Caves are eventually destroyed by roof collapse and by
intersection of passages by surface erosion. At any
given time, different parts of the same cave may be
undergoing different stages in this sequence.
THE EARLIEST STAGES
At great depth beneath the surface there is very lit¬
tle groundwater flow because openings in the rock are
narrow and few, and hydraulic gradients are feeble.
Wide tectonic fractures may penetrate to great depth
and can allow substantial water flow, but they are
sparse. As rocks are gradually exposed at the surface
by uplift and erosion, increasing amounts of ground-
water pass through them. At first, the rate at which
any opening enlarges by dissolution depends almost
entirely on the discharge through it. In turn, the most
important controls over the amount of flow are how
much water is available to supply the upstream end
of the growing caves, and the widths of the initial
openings. Also influential, but less so, are the over¬
all hydraulic gradient, length of flow, and water
chemistry.
When the openings along a particular flow path
enlarge enough that water can pass through all the
way to the spring while still retaining much of its solu-
tional capacity, this flow path begins to enlarge much
more rapidly than its neighbors' do. From then on,
water chemistry becomes the main factor controlling
growth rate, and the entire path enlarges rapidly at a
maximum wall retreat of about 0.001—0.01 cm yr 1 .
Openings that have not reached this stage continue to
enlarge only slowly, often at diminishing rates with
time. The typical result is a cave with only a few nego¬
tiable passages surrounded by openings that have
enlarged only slightly.
The time needed to reach this stage of rapid cave
enlargement can be considered the "gestation time"
through which an incipient cave must pass in order to
grow into a true cave. It is difficult to specify exactly
when this time begins. Some researchers argue that it
should include the entire age of the soluble rock,
including early depositional conditions, compaction,
and burial by other rocks, as well as later uplift and
perhaps folding and faulting. But before cave growth
can truly begin, there must be a substantial hydraulic
gradient through the rock. This usually requires that
the soluble rock be exposed at or near the surface,
where zones of groundwater recharge and discharge
are well defined. Most people date the age of a cave to
the onset of these conditions.
Field evidence and computer modeling show that
short fissures in limestone (up to a few hundred
meters long) with initial widths of 0.01—0.1 cm would
require no more than a few thousand or tens of thou¬
sands of years to reach their maximum enlargement
rates. Relatively long and/or narrow paths require
much longer times, typically on the order of hundreds
of thousands of years.
DEVELOPMENT OF CAVE PATTERNS
Competition between Initial Flow Routes
Most groundwater infiltrates in upland recharge
areas and emerges at lower elevations such as river
valleys. Patterns of solution caves typically reflect
these trends. Such caves can form only along paths
that gain discharge with time, which requires one of
the following conditions:
• Where water leaks from a stream and drains to a
lower outlet, the underground flow increases
dramatically as the initial openings enlarge by
dissolution. When the incipient cave enlarges
enough to carry all of the available flow, the surface
stream becomes a sinking stream. From then on the
mean-annual flow rate remains roughly the same,
but the water level in the cave drops as the passages
enlarge. The cave reaches its maximum enlargement
rate early in this enlargement history, so the leveling
out of the flow rate does not interfere with the
growth of the cave.
• Water that seeps through soil into underlying
soluble rock can gain discharge only by increasing
its catchment area. Because the upstream ends of
the major flow paths enlarge most rapidly, soil
subsides into these growing voids to form
sinkholes. As sinkholes grow they increase their
catchment areas, delivering progressively more
water to the passages they feed. Water input to the
caves increases in an irregular manner, and much
less rapidly than in routes fed by leaking
streambeds.
The routes fed by leaky surface streams increase
their flow much more rapidly, so they are usually the
first to form traversable cave passages. Passages fed by
sinkholes have a limited catchment area and require
ENCYCLOPEDIA OF CAVES
600
p
more time to form. They usually join the earlier pas¬
sages as tributaries of a branching cave system. The
first passages to form are usually short and direct.
With time, as their water levels drop, these early pas¬
sages serve as targets for later passages that drain
more remote areas. Although the growth of any single
passage propagates in the downstream direction, the
overall system grows in the upstream direction, away
from the springs, by the addition of new tributaries.
A typical sequence is shown in Figure 1.
Vertical Organization of Cave Passages
Most vadose passages, because their flow is purely
gravitational, tend to have a strong component down
the dip of the rock strata. This is especially true in
well-bedded rocks. Phreatic passages show no consis¬
tent relation to the dip, unless that is the only direction
to potential outlets, or where prominent fractures also
extend in that direction. In well-bedded rocks, the
intersections between dipping beds and low-gradient
water tables encourage many phreatic passages to
develop roughly along the strike of the beds. These
FIGURE I Typical stages in the early development of a solution
cave. Segment B-A forms first as a short, high-gradient path from a
sink point in the upland stream to the entrenched river. Segment C-B
forms next, aided by the decrease in head in the previous segment.
The passage fed by the sinkhole at D is slower to form because of its
smaller catchment and less rapid increase in flow with time.
relationships tend to be obscure where the geologic
structure is complex.
The vertical arrangement of underground water
flow is strongly controlled by the deepening of surface
valleys by river erosion, and cave passages reflect this
control. In general the largest passages form at the ele¬
vation of erosional base level when river entrenchment
is slow. At such times, rivers develop floodplains, and
springs remain at fairly constant elevations for lengthy
periods of time. Meanwhile the passages that drain to
those springs are able to reach large size. In contrast,
most passages that form during rapid river entrench¬
ment are small because the cave-forming water under¬
goes more frequent diversion to lower routes. The
major passages may be arranged in several different
levels that decrease in age downward. Partial filling of
river valleys with sediment tends to block nearby cave
passages, causing them to fill with sediment as well, or
at least to become ponded with water.
This conceptual model has been well validated in
many karst areas, but several complications can dis¬
rupt this simple interpretation. Vadose passages may
become perched on insoluble strata and grow to large
size well above the local levels of surface rivers. In
addition, most phreatic passages contain vertical loops
that descend below river levels. Even the ideal cave
levels controlled by pauses in river entrenchment
are not perfectly level. For this reason, many people
prefer to call them stories or tiers, either of which is
preferred in general applications. However, the term
cave level is still appropriate where there is a clear rela¬
tion to erosional base level. The critical measurement
is not the average elevation of a phreatic passage, but
instead the elevation where passages change from
vadose to phreatic types (for example, from a canyon
to a tube). This transition may not be clear where the
geologic structure is complex or in caves that experi¬
ence severe fluctuations in water level due to flooding.
Figure 2 is an idealized profile through a multisto¬
ried cave. Three main tiers of cave development are
shown, with decreasing loop amplitudes from the
highest tier to the lowest. This is a conceptual ideal
that is not characteristic of all multistoried caves.
Fissures may be sparse at first, so that passages are
constrained to only a few deeply descending loops. As
erosional unloading and cave development persist,
fissures become more numerous until eventually the
passages are able to form more or less along the water
table, with minimal phreatic looping. However, in
some caves the greater amplitude of loops in upper
passages is instead caused by floodwaters, which pro¬
duce ungraded, looping bypasses around low-flow
routes that have more uniform gradients. Furthermore,
the depth of phreatic loops also tends to be greatest in
ENCYCLOPEDIA OF CAVES
PASSAGE GROWTH AND DEVELOPMENT
601
FIGURE 2 Idealized profile through a complex, multitier cave
system showing decreased loop amplitude with time as fissure fre¬
quency increases from stage A to B to C. Loop amplitude is also pro¬
portional to the length of the cave system, to structural deformation,
and massiveness of the soluble rock.
highly deformed or massive rocks, and in broad
regional flow systems.
Origin of Branching Systems
Branching (branchwork) caves are roughly the under¬
ground equivalents of surface drainage. Many small
passages join together in the downstream direction to
form larger but fewer passages. Of all solution caves,
branchwork caves are by far the most common. There
are several reasons for the branching pattern:
• As phreatic passages enlarge, the local hydraulic
head within them decreases. Groundwater flows
from surrounding smaller openings, where the
water table is higher, toward the low heads in the
passages.
• Vadose passages have no inherent tendency to
converge, because their streams are independent of
one another. However, the structures that they
follow often intersect, forcing their streams to
converge as tributaries. Examples of convergent
structures include intersecting fractures, and
trough-like features in bedding-plane partings.
• Water from broad recharge areas converges toward
outlets of limited extent, generally discharging in
stream valleys. Thus there is a natural tendency for
conduits to join each other simply by competition
for space. There is little opportunity for water to
diverge farther downstream. The exception is in the
vicinity of spring outlets, where local distributary
systems may develop because of collapse, flooding,
and widening of the main stream routes by erosional
stress release. In those areas a tangle of diversion
passages may develop like the channels on a river
delta.
Development of Maze Caves
Most nonbranchwork caves are mazes in which all
passages form more or less simultaneously. A maze
cave can form only if the growth rate is similar along
many alternate flow paths. This requires that the ratio
of discharge to flow distance is large along many alter¬
nate flow routes. This condition can be met in any of
the following ways:
1. Floodwater recharge. A rapid influx of water into
caves during floods will fill the caves with high-
pressure water and force solutionally aggressive
water into all fissures in the surrounding rock.
Floodwaters combine steep gradient, high
discharge, and short flow paths. As a result, these
openings enlarge at approximately the maximum
possible rate. This process is most active in the
vicinity of constrictions in major stream passages
caused by collapse, sediment chokes, or poorly
soluble strata.
2. Short flow paths. Mazes can also form where many
alternate flow paths are short (z.e., from where
the water first enters the soluble rock). In this
case, all openings except for the narrowest enlarge
simultaneously at similar rates. The epikarst is an
example. This is a network of enlarged openings, all
of which have enlarged simultaneously at comparable
rates. The tendency for uniform growth diminishes
with flow distance, so only a few major drains lead
downward into the aquifer below.
3. Uniform recharge through a permeable insoluble rock.
Where water enters a soluble rock by first seeping
through an insoluble but permeable rock, all
fissures in the soluble rock receive roughly
comparable amounts of flow, regardless of how
wide they are. For example, mazes can be formed
by recharge through a permeable but insoluble
material such as quartz sandstone. Their uniform
enlargement is also aided by the short flow paths.
Rising water entering soluble rock through
permeable material can produce a similar effect,
and the resulting caves tend to be relatively
extensive.
4. Local boosts in aggressiveness. Where the groundwater
gains solutional capacity by mixing of waters of
contrasting CO 2 content or salinity, or by oxidation
of sulfide-rich water, many alternate routes are
enlarged at roughly comparable rates. Mazes are
formed along short paths of flow in and around the
mixing zone.
ENCYCLOPEDIA OF CAVES
602
p
The differences in maze patterns depend partly on
geologic structure. Netivork mazes consist of intersecting
fissures arranged in a pattern like that of city streets.
They require many intersecting fractures (joints or
faults), which are typical of massive or thick-bedded
rock. They can be formed by any of the four processes
above. Anastomotic mazes have a braided pattern of
intersecting tubes, usually arranged two-dimensionally
along a single parting or fault. They are nearly all
formed by process 1 above. Spongework mazes form
where primary (matrix) porosity is dominant. In pat¬
tern they resemble the intersecting holes in a sponge.
Most of them form by process 4, and less commonly
by process 1. A two-dimensional variety can form
along bedding-plane partings. Ramiform mazes consist
of rooms with offshoots extending outward from them
at various elevations. They usually include areas
of network or spongework maze development and
are formed mainly by process 4. Many network and
anastomotic mazes, and a few spongework mazes,
are merely superimposed on a basic branchwork patt¬
ern and constitute only part of the entire cave
development.
Figure 3 summarizes typical cave patterns and their
relation to sources of aggressive water and to domi¬
nant structural characteristics.
Adjustment of Caves to Changing Conditions
When an incipient cave finally reaches its maximum
growth rate, several other changes take place more or
less simultaneously. The flow regime of cave water
changes from laminar to turbulent, which increases the
solution rate slightly. The flow also becomes able to
transport sediment. For example, it can carry away the
soil that subsides into caves through karst depressions,
allowing the depressions to grow more rapidly. The
sediment load can also help to enlarge caves by
mechanical abrasion. But in places, sediment accumu¬
lates in thick beds that retard dissolution and erosion.
When a cave is able to transmit the entire flow from
its recharge area, the average flow can increase no fur¬
ther. Instead the head within the passage decreases as
the cross-section continues to enlarge. Much of the
upstream part of the cave becomes vadose, and
streams may entrench canyons in the passage floors.
Where water first enters caves, it is fairly rich in soil-
derived CO 2 and may acquire even more by oxidation
of organic materials within the caves. As caves acquire
entrances that allow air exchange with the surface,
many cave streams lose part of their solutional capacity.
Loss of CO 2 through entrances and other openings can
drive the stream water to supersaturation with
CAVE PATTERN
curvilinear
branchwork
rectilinear
branchwork
f
anastomotic
maze
V
network
maze
spongework
maze
4
ramiform
pattern
SOURCE OF
AGGRESSIVE WATER
sinkholes
•
•
•
•
sinking streams
•
•
•
•
•
uniform seepage
•
•
mixing of 2 sources
•
•
•
sulfuric acid
•
•
•
•
DOMINANT
STRUCTURES
bedding-plane
partings
•
•
•
•
fractures
•
•
•
intergranular pores
•
•
FIGURE 3 Relation of cave patterns to mode of groundwater recharge and geologic structure. Dot sizes show the relative abundance of
each pattern within the various recharge and structural categories. Single-passage caves are rudimentary or fragmentary versions of those
shown here.
ENCYCLOPEDIA OF CAVES
PASSAGES
603
dissolved calcite or dolomite. As a result, many cave
streams are aggressive only during high flow.
As the land surface becomes dissected by erosion,
patterns of groundwater recharge change. Initial water
sources are usually few and large, but with time they
may become divided into many smaller ones. Vadose
water must travel increasingly greater distances to
reach the water table, and extensive complexes of
vadose canyons and shafts can form. The resulting pat¬
tern of active cave streams tends to become denser
with time, and eventually much denser than that of
the original surface drainage. Growing dolines eventu¬
ally merge to form a continuous karst surface. In a
mature karst region the only surface streams that
retain their flow are the main entrenched rivers and
the ephemeral upstream ends of sinking streams.
CAVE DEGENERATION
As the land erodes, the surface intersects underlying
cave passages, segmenting them and eventually destroy¬
ing them entirely. Evidence for the cave may persist for
a while as a canyon-like feature or a rubbly zone of
collapsed blocks. This final phase in the life of a cave
passage usually occupies tens of thousands or even hun¬
dreds of thousands of years. However, newer passages
continue to develop if the soluble rock extends to lower
elevations. In dipping strata, new areas of rock are
uncovered by erosion at about the same rate as they
are eroded away in the up-dip areas. Thus some of the
caves we see today are the descendants of earlier caves
that once occupied parts of the soluble rock that are
now gone. This process ends when the entire soluble
rock in the cave region is eroded away.
CONCLUSION
Patterns of solution caves are highly sensitive to
their physical setting. With the aid of dating of cave
deposits, cave patterns can shed considerable light on
the history of landscape development, hydrology, and
climate changes in the surrounding region.
Ford, D. C., & Ewers, R. O. (1978). The development of limestone
cave systems in the dimensions of length and depth. Canadian
Journal of Earth Sciences, 15, 1783—1798.
Granger, D. E., Fabel, D., & Palmer, A. N. (2001). Pliocene-
Pleistocene incision of the Green River, Kentucky, determined
from radioactive decay of 26 A1 and 10 Be in Mammoth Cave sedi¬
ments. Geological Society of America Bulletin, 113(7), 825—836.
Klimchouk, A. B. (2011). Hypogene speleogenesis: Hydrogeological and
Morphogenetic Perspective (2nd ed.). Carlsbad, NM: National Cave
and Karst Research Institute, Special Paper Series 1.
Palmer, A. N. (1991). Origin and morphology of limestone caves.
Geological Society of America Bulletin, 103, 1—21.
Palmer, A. N. (2007). Cave geology. Dayton, OH: Cave Books.
White, W. B. (1969). Conceptual models of carbonate aquifers.
Ground Water, 7(3), 15-21.
White, W. B. (1977). Role of solution kinetics in the development of
karst aquifers. In J. S. Tolson & F. L. Doyle (Eds.), Karst hydrogeol¬
ogy (pp. 503—517). International Association of Hydrogeologists,
12th Memoirs. Huntsville, AL: UAH Press.
Worthington, S. R. H. (2004). Hydraulic and geological factors
influencing conduit flow depth. Cave and Karst Science, 31(3),
123-134.
Worthington, S. R. H. (2005). Evolution of caves in response to base-
level lowering. Cave and Karst Science, 32(1), 3—12.
PASSAGES
Qeorge Veni
National Cave and Karst Research Institute
Voids that transmit or have the capacity to transmit
turbulently flowing water in karst are called conduits, a
term that is used generically to describe all segments
of karst drainage networks or to refer specifically to
segments too small for human entry. Turbulent flow, a
key factor that distinguishes conduits from smaller
openings in the rock, occurs when an opening achieves
a width of about 1 cm. Conduits large enough for
human entry are called caves. Passages, rooms, pits,
and domes are types of conduits within caves that
have certain physical characteristics. Passages are hori¬
zontal to moderately sloping and narrow relative to
their lengths. A room is a type of passage that is wide
relative to its length; rooms often form where passages
intersect and are circular, oval, or irregular in shape.
Pits and domes are steeply sloping to vertical pas¬
sages. When encountered from the top, they are called
pits, and from below, they are called domes.
See Also the Following Article
Passages
Bibliography
Dreybrodt, W. (1990). The role of dissolution kinetics in the develop¬
ment of karst aquifers in limestone: A model simulation of karst
evolution. Journal of Geology, 98(5), 639—655.
DEFINITIONS AND CONCEPTS
Several factors affect the location, size, and shape of
cave passages. Because limestone, dolomite, and other
karstic rocks are poorly permeable, passages develop
along fractures and bedding planes where permeabil¬
ity is higher. Initially, groundwater will flow through
a variety of these structurally guided conduits until a
Encyclopedia of Caves.
© 2012 Elsevier Inc. All rights reserved.
604
p
preferential path is developed that enlarges to form
a passage. The pattern that the passages form, the
morphology or shape that they take, and the sediments
and features that they contain, reflect the lithologic,
structural, and hydrologic conditions within the karst
aquifer. Many of the factors and conditions discussed
in this article are shown and summarized in Figure 1.
This discussion describes typical conditions; local
hydrogeologic factors may produce variations and
exceptions. For in-depth discussions of cave and pas¬
sage development, see Bogli (1980), White (1988), Ford
and Williams (1989), Gillieson (1996), Klimchouk et al.
(2000), and Klimchouk (2007).
Passages can be classified into four general types
according to their position in the karst aquifer. Vadose
passages occur at the upgradient ends of the aquifer,
and they transmit water through the vadose zone from
the surface to the water table. They tend to have high
gradients and, depending on the gradient and volume
of flow, contain small or large accumulations of sedi¬
ment. Phreatic passages occur at and below the water
table (the phreatic zone) and transmit groundwater
from the vadose to the discharge ends of the aquifer.
They have relatively low gradients and usually contain
significant accumulations of sediment where intersected
by vadose passages. Tributanj passages feed groundwater
into the main vadose and phreatic passages. They
decrease in numbers down the hydraulic gradient while
generally increasing in size and volume of flow.
Discharge passages occur near springs where groundwa¬
ter flows to the surface. They often form a "distribu¬
tary" pattern that divides the water into multiple
outlets that form in response to changing conditions on
the surface, such as downcutting of valleys below the
springs. Some of these passage types may be hydrologi-
cally inactive or active only during flood events, having
formed when groundwater levels were higher.
PATTERNS OF CAVE PASSAGES
The distribution and geometry of passages within a
cave are dictated by a combination of local lithologic,
structural, and hydrologic factors. Variations in the sol¬
ubility of the strata may determine the particular bed
in which a passage forms. Interbedded rocks of very
low solubility, such as shales, clays, or chert, may
perch groundwater and produce passages at the base
of a group of soluble rocks. Fractures that breach those
poorly soluble units allow pits to form that may extend
down to the next poorly soluble unit, creating a stair¬
step pattern (Fig. 2).
Passages formed along fractures tend to be narrow
and linear, with sharp, often right-angle turns and
intersections with other passages (Fig. 1, angular pas¬
sages). Passages guided by bedding planes are usually
TYPE OF RECHARGE
VIA KARST DEPRESSIONS
DIFFUSE
HYPOGENIC
(LIMITED DISCHARGE
FLUCTUATION)
BRANCHWORKS
(USUALLY SEVERAL LEVELS)
& SINGLE PASSAGES
SINKING STREAMS
(GREAT DISCHARGE
FLUCTUATION)
THROUGH SANDSTONE
SINGLE PASSAGES AND
CRUDE BRANCHWORKS.
USUALLY WITH THE
FOLLOWING FEATURES
SUPERIMPOSED:
MOST CAVES ENLARGED
FURTHER BY RECHARGE
FROM OTHER SOURCES
INTO POROUS
SOLUBLE ROCK
MOST CAVES FORMED BY
MIXING AT DEPTH
DISSOLUTION BY ACIDS OF
DEEP-SEATED SOURCE OR
BY COOLING OF THERMAL
WATER
V
FISSURES,
IRREGULAR NETWORKS
FISSURES, NETWORKS
ISOLATED FISSURES
AND RUDIMENTARY
NETWORKS
CURVILINEAR
PASSAGES
/
ANASTOMOSES,
ANASTOMOTIC MAZES
SHAFT AND CANYON
COMPLEXES, INTERSTRATAL
SOLUTION
/
SPONGEWORK
RAMIFORM CAVES.
RARE SINGLE-PASSAGE AND
ANASTOMOTIC CAVES
k
RUDIMENTARY
BRANCHWORKS
X
SPONGEWORK
sandstone :
•V ■S Wf- »•■■ ■ - v
RUDIMENTARY
SPONGEWORK
SPONGEWORK
RAMIFORM & SPONGEWORK
CAVES
FIGURE 1 Summary of cave patterns and their relationship to types of recharge and porosity. From Palmer, 1991. Used with permission.
ENCYCLOPEDIA OF CAVES
PASSAGES
605
wide and meandering, with low-angle passage inter¬
sections (Fig. 1, curvilinear passages). When the struc¬
tural grain of the rock, such as the primary fracture
direction or dip of the beds, significantly differs from
the hydrologic gradient, cave passages may form
closely spaced meanders or zig-zags as competing fac¬
tors alternately control the development of different
short segments of passages.
Vadose tributary passages in caves formed in hori¬
zontal or nearly level beds join the main passages from
both sides in a roughly equal number; however, where
beds are dipping, the tributaries enter the main
passages from predominantly the up-dip direction,
and distributary passages branch off in the down-dip
direction. This is most noticeable in steeply dipping
strata but has also been observed in beds with <1° of
dip. Phreatically formed passages, usually the main
passages within a cave, tend to follow the strike of
dipping strata (Fig. 3).
The pattern of cave passages will vary depending
on the number of sources that contribute most of
the groundwater within the cave. Caves recharged by
multiple sites of roughly equal contribution will tend
Impermeable rock
FIGURE 2 Schematic model of stairstep cave development.
Adapted from Crawford, 1996.
to form the tributary systems described previously.
Numerous small passages will join and enlarge as they
flow downgradient, and their combined flow will form
the main drainage passage that leads to a spring (Fig. 1,
sinkhole type of cave). In contrast, caves formed by
capturing a single large source of water, such as a sur¬
face stream, quickly grow into single large passages
that maintain a relatively constant size between their
source and spring. Tributaries are few and provide rela¬
tively little water, and due to their lesser flows those
passages are smaller and more recently formed.
Occasionally, they will form floodwater mazes next to
the main passage (Fig. 1, sinking stream type of cave).
Passages that form down steep hydraulic gradients
are more linear than those formed on low gradients.
Nearly level gradients will produce anastomotic
patterns, where passages divide their flow as they
branch, rejoin, and vary in size. Ponded water may
dissolve low wide rooms along bedding planes or a
criss-crossing network of passages along fractures.
Ponding is often the result of a geologically short-lived
phenomenon, such as collapse along a cave stream or
inundation of the spring. Normal conditions are
restored when, for example, groundwater cuts a path
through or around the collapse. Flooding may also
inject water episodically into fractures and bedding
planes, temporarily creating ponded conditions until
the floodwaters subside. Flooding from chemically
aggressive surface streams is especially effective in
creating floodwater mazes. Some maze passages form
by vadose water moving down through poorly soluble
but permeable strata, such as sandstone, and enlarging
fractures in underlying soluble rocks so that they inter¬
connect linearly (Fig. 1, fissure networks).
Passage patterns also reflect their development in
shallow to deep portions of the aquifer. Roughly
FIGURE 3 Map of Honey Creek Cave, Texas, illustrat¬
ing passage development and groundwater flow relative
to dip of the beds. From Elliott, W. R., and Veni, G. (Eds.),
1994. Convention Guidebook, National Speleological Society,
Huntsville, AL. Used with permission.
ENCYCLOPEDIA OF CAVES
606
p
horizontal passages form in the shallow portion of the
phreatic zone. They may result from closely underly¬
ing poorly permeable strata but also develop without a
poorly permeable base in highly fractured rock. In
aquifers where the rock is less fractured, fewer perme¬
able zones are available to develop passages between
the vadose passages and the springs. Consequently,
groundwater pressure will move water through deeper
fractures, forming deeper passages. These passages
will ascend and descend en route along the more
permeable fractures. The greater the distance to
the springs, the greater is the potential depth of the
conduits (Fig. 4). Springs that discharge from such
systems in Mexico have measured depths of more
than 300 m.
While most passages form from the top down by
water entering the ground and discharging back to the
surface at a lower elevation, hypogenic passages form
from the bottom up. While initially described as the
result of rising gases, typically hydrogen sulfide, that
flow into groundwater to form acidic solutions that
more aggressively dissolve the rock, recent studies show
these hydrologic conditions create a distinct set of
morphologies independent of lithology and
FIGURE 4 Model of cave development below the water
table relative to fracture frequency. Adapted from Ford and Ewers,
1978 ; and Ford and Williams, 1989.
groundwater chemistry. These passages are character¬
ized by the absence of features that reflect vadose flow
from the surface; by decreased passage development
with increased elevation; by abrupt terminations of
rooms and passage (often with maze passages or honey¬
combed walls); by the occurrence of irregularly shaped
rooms and passages with little apparent relationship
to convergent, tributary flow patterns; and, where
exposed, by deep, narrow fissures in the floors where
groundwater rose (Fig. 1, hypogenic caves).
Mineralogical indicators may also be present, such as
gypsum deposits, where hydrogen sulfide produced sul¬
furic acid which dissolved the limestone. Carlsbad
Caverns (New Mexico) is the best known example of a
cave formed by this process.
PASSAGE MORPHOLOGY
AND EVOLUTION
Phreatically formed passages have circular to ellipti¬
cal cross sections. The ellipses will be vertical, horizon¬
tal, or diagonal, depending on whether the passage is
enlarged along a fracture or bedding plane and the ori¬
entation of those features (Fig. 5). Phreatic walls are
relatively smooth. The symmetry of phreatic passages
results from their development below the water
table where roughly equal water pressure is present on
all walls.
HORIZONTAL
VERTICAL
DIAGONAL
PHREATIC
-o-
0
a
VADOSE
A
INCISED
¥
8
Q
FIGURE 5 Cross-sectional patterns of phreatic, vadose, and
incised passages relative to bedding and fractures.
ENCYCLOPEDIA OF CAVES
PASSAGES
607
Vadose passages are relatively tall and narrow in
cross section, especially where the hydraulic gradient
is steep and/or the bedding is thick. They form by
vadose water downcutting the floor to reach the water
table in as short a distance as possible. Water pressure
and dissolution are concentrated at their floors and
lowermost walls. Pits are the ultimate form of vadose
passage. The walls of vadose passages are often rough
and irregular due to greater turbulence and differences
in water pressure and dissolution, as well as from
greater mechanical erosion by rocks and other sedi¬
ments carried by the streams. Passages formed along
vertical fractures tend to be narrow while those formed
along horizontal or dipping fractures and beds are
wider because they can more easily undercut, collapse,
dissolve, and wash away the strata (Fig. 5).
Water levels in an aquifer naturally decline over
time as the land surface erodes, allowing groundwater
to discharge from progressively lower elevations.
Consequently, phreatic passages will often be drained
and become modified by the new vadose hydrologic
conditions. The most common change is the incision of
passage floors to drain water to the descending water
table. Keyhole passage shapes are typical, where a
broad, smooth-walled phreatic tube sits over a narrow,
rough-walled vadose canyon that is cut into its floor
(Fig. 5). In caves where phreatic water moved up and
down pits, newly vadose passages at the tops of pits
will form waterfalls that lengthen the pit by cutting
headward, toward the water source, into the wall
below the passage.
During phreatic conditions, the buoyancy of the
water can support up to 42% of the weight of the pas¬
sage ceiling. In broad rooms and passages, especially
in thin- to medium-bedded limestone of low structural
strength, the change to vadose conditions can result in
ceiling instability and collapse. The influx of chemi¬
cally aggressive vadose groundwater may accelerate
the collapse. The result after the collapse is a more
structurally stable dome shape in the passage's ceiling
and a mound of breakdown on its floor. Because the
collapsed rock will fill more space than when it was
intact, it could extend to the ceiling and prevent
further collapse. In some caves, where jagged walls
demonstrate extensive collapse but there is no underly¬
ing mound of breakdown, groundwater on or under
the rubble is dissolving and removing the pile faster
than it is accumulating.
Some of the passage patterns, shapes, and modifica¬
tions described in this article can be hidden by sediments
and speleothems. Pits seemingly without passages are the
most common example. The passages exist but are buried
by sediments deposited on the pit floors where the water
that carries them changes from high-energy vertical flow
to low-energy horizontal flow conditions. Another some¬
what common example is where large passages end in
low crawlways that can barely be entered. These often
reflect where the once sediment-free phreatic conduit
dipped down and up in a V-shaped pattern (Fig. 4); dur¬
ing vadose conditions, the bottom of the "V" was filled by
sediments. In some cases, dense deposits of chert, clay, or
other poorly soluble materials may armor a passage floor
to prevent downward erosion. As more of this material is
deposited, groundwater will dissolve the walls and ceil¬
ing to enlarge the passage upward by a process called
pamgenesis.
Karst aquifers evolve at different rates according to
the geologic factors described in this article, but espe¬
cially in response to erosion of the overall landscape.
The downcutting of surface streams will drain old
aquifer levels, create new ones, and change groundwa¬
ter gradients toward new springs. These changes will
HIGH ELEVATION PASSAGE
FIGURE 6 Three-stage evolutionary model of tapoff passage
development. Adapted from Veni, 1997.
ENCYCLOPEDIA OF CAVES
608
p
be reflected in cave passages. In many cases, water
from one passage will be captured or pirated into
another passage with a steeper hydraulic gradient.
These passages are known as tapoff passages. Originally
meant to describe a newly formed passage that trans¬
mits water from a cave to a newly formed spring, its
definition has been expanded to describe a relatively
short, high-gradient cave stream passage that pirates
flow from a high elevation stream to a lower elevation
stream (Fig. 6). Tapoffs can form between any combi¬
nation of surface and subsurface streams, are usually
small relative to the surrounding passages, and have
few or no tributaries. With sufficient time, they may
enlarge to the point where they may be difficult or
impossible to recognize.
CONCLUSIONS
Cave passages form the framework of karst aqui¬
fers. Their proper interpretation provides insights into
the evolution of the aquifer and land surface, ground-
water flow behavior, potential fate and transport of
contaminants that might enter the aquifer, and the
definition and distribution of habitat for terrestrial and
aquatic species living in the cave and aquifer, as well
as providing key information on the portion of the
conduit system that remains humanly inaccessible.
See Also the Following Article
Passage Groivth and Development
Bibliography
Bogli, A. (1980). Karst hydrology and physical speleology. Berlin:
Springer-Verlag.
Crawford, N. C. (1996). The karst hydrogeology of the Cumberland
Plateau Escarpment of Tennessee. Nashville: Tennessee Department
of Environment and Conservation.
Ford, D. C., & Ewers, R. O. (1978). The development of limestone
cave systems in the dimensions of length and depth. Canadian
Journal of Earth Sciences, 15(11), 1783—1798.
Ford, D. C., & Williams, P. W. (1989). Karst geomorphology and hydrol¬
ogy. London: Unwin Hyman.
Gillieson, D. (1996). Caves: Processes, development, management. Cambridge:
Blackwell Publishers.
Klimchouk, A. B. (2007). Hypogene speleogenesis: Hydrogeological and
morphogenic perspective. Carlsbad: National Cave and Karst Research
Institute.
Klimchouk, A. B., Ford, D. C., Palmer, A. N., & Dreybrodt, W. (Eds.),
(2000). Speleogenesis: Evolution of karst aquifers Huntsville, AL:
National Speleological Society.
Palmer, A. N. (1991). Origin and morphology of limestone caves.
Geological Society of America Bulletin, 103( 1), 1—21.
Veni, G. (1997). Geomorphology, hydrogeology, geochemistry, and
evolution of the karstic Lower Glen Rose Aquifer, South-Central
Texas, Texas Speleological Survey Monographs (Vol. 1).
White, W. B. (1988). Geomorphology and hydrology of karst terrains.
New York: Oxford University Press.
POPULATION STRUCTURE
Valerio Sbordoni, Qiuliana A llegrucci, and Donatella
Cesaroni
Tor Vergata University, Roma, Italy
INTRODUCTION
Population structure is an important guideline to under¬
standing the evolution of cave-dwelling animals,
because it represents the outcome of their history and
adaptation as well as the groundwork for speciation in
the cave environment. Population structure can be
viewed from two different perspectives. Ecologists usu¬
ally view the composition of a population according to
age and sex of individuals, and population geneticists
keep in mind the organization of genetic variation
within and between populations, with special emphasis
on their spatial arrangement. In this article we address
the latter aspect. The advent of molecular techniques,
from allozymes to microsatellites and to DNA sequenc¬
ing analysis, led to significant improvements in under¬
standing the genetics of natural populations.
Particularly, the feasibility to estimate parameters, such
as the amount of genetic variation, gene flow, time since
isolation, provides important evidence to test alternative
hypotheses on patterns of colonization and evolution of
cave populations. Moreover, mitochondrial and nuclear
DNA genealogies have revealed genetic partitions that
can be interpreted in terms of habitat distribution and
geography. These scenarios appear to be influenced by
many factors intrinsic to the life histories of the organ¬
ism (i.e., the mating system, levels of gene flow) and by
historical biogeography (i.e., past distributional ranges)
and demographic factors. The most obvious cause of
structuring in cave populations is habitat fragmentation.
However, as we will discuss, habitat fragmentation may
not include the unique factor of population structuring.
Generally speaking, caves reflect the historical process
of habitat fragmentation quite well. Cave regions in
temperate countries have been subjected to profound
and repeated changes in climate and vegetation, which
are particularly well documented in the Pleistocene.
Cold and dry phases corresponding to glacial periods
have repeatedly led to changes from mesophyt forest to
mountain steppe. Consequently, previously widespread
Encyclopedia of Caves.
© 2012 Elsevier Inc. All rights reserved.
POPULATION STRUCTURE
609
ancestral populations of several cave-dwelling species
became confined to small refugial habitats, leaving relic -
tual populations at different steps of isolation, depend¬
ing on the biological properties and/or the history (i.e.,
time) of the organisms involved. Extrinsic causes for
fragmentation can also be detected in other circum¬
stances such as the structure and history of drainage
systems, tectonic changes, or lava flow events in volca¬
nic cave areas. In many cases, these phenomena can be
dated to a reasonable extent.
Other evolutionary processes can also account for
population structuring, especially where population
structure is perceived at much smaller scales, even
within a single cave. Double or multiple invasion of a
cave habitat at different times by different populations
of a given species represents a feasible mechanism.
Alternatively, extrinsic factors such as habitat and
resource patchiness may interact with the intrinsic bio¬
logical properties of organisms, like fecundity, dis¬
persal ability, and behavioral characters, leading to a
population structure maintained by different forms of
natural selection.
In this article, a series of case studies will be dis¬
cussed to underscore the role of different factors in
shaping population structure in a limited sample of
cave organisms that have been the object of appropri¬
ate population genetic studies.
ESTIMATING GENETIC STRUCTURE
The genetic parameters relevant to investigate popu¬
lation structure are effective population size ( N e ),
observed and expected heterozygosity ( H 0 and H e ,
respectively), genetic distance (D), amount of structur¬
ing between subpopulations (Fg T ), and gene flow ( N e m ,
where m is the migration rate).
Effective Population Size (N e )
The number of individuals that effectively partici¬
pates in producing the next generation is named effec¬
tive population size. Generally, the effective size of a
population is considerably less than the census size.
Evolutionary processes are greatly influenced by the
size of populations.
Heterozygosity
Mean heterozygosity, calculated across a number of
loci, is a valuable parameter used to estimate the
degree of genetic variation within a population.
Population structuring occurs when genotype frequen¬
cies deviate from Hardy—Weinberg expected
proportions, or panmixia is unfulfilled. If either
inbreeding or selection occurs, then populations can be
considered "structured" in some way.
Genetic Distance
When two populations are genetically isolated, both
mutation and genetic drift lead to differentiation in the
allele frequencies at selectively neutral loci. As the
amount of time that two populations are separated
increases, the difference in allele frequencies between
them should also increase, until each population is
completely fixed for separate alleles. Therefore, calcu¬
lation of genetic distance (D) between two populations
provides a relative estimate of the time elapsed since
these populations have existed as a single panmictic
unit. Small estimations of distance among completely
isolated populations indicate that they have only been
separated for a short period of time. Alternatively, in
the absence of isolation, small values of genetic dis¬
tance may indicate population structure (i.e., subpopu¬
lations in which there is random mating, but between
which there is a reduced amount of gene flow).
F-Statistics
F-statistics, developed by Wright (1965), represent
the basic method to measure the amount of subdivi¬
sion in populations. T-statistics can be viewed as a
measure of the correlation of alleles within individuals,
and they are related to inbreeding coefficients. An
inbreeding coefficient is really a measure of the non-
random association of alleles within an individual. As
such, F-statistics describe the amount of inbreeding¬
like effects within subpopulations, among subpopula¬
tions, and within the entire population. In particular,
the Fgj index (or Rg T , as estimated for microsatellite
data) is an estimator of the amount of structuring of a
population into subpopulations.
Migration
If there is no migration (gene flow) occurring
between two populations or demes, eventually alter¬
nate alleles will become fixed and will reach 1.
Alternatively, it has long been known that if migration,
measured in terms of N e m, is >1 (where N e is the
effective population size and m is the proportion of
migrants per generation, or migration rate), the allele
frequencies in the subpopulations remain homoge¬
nized (Wright, 1931). If, however, migration is present
but N e m < 1, an equilibrium based on the rate of muta¬
tion, migration, and genetic drift will be established.
ENCYCLOPEDIA OF CAVES
610
p
Phylogeography
Recently, a relatively new discipline named phylogeo¬
graphy has been applied to investigate the principles
and processes governing the geographic distributions of
genealogical lineages within and among closely related
extant species. Phylogeographic studies focus on under¬
standing the contribution of historical versus contempo¬
rary ecological processes in shaping present-day species
distributions. Phylogeographic inferences are based on
DNA sequences sampled from the same locus in many
individuals collected throughout the geographic range
of a species. Statistical analyses are based on coales¬
cence theory that employs a sample of individuals from
a population to trace all alleles of a gene shared by all
members of the population to a single-ancestral copy.
This uses sophisticated model-driven approaches that
answer specific questions for inferring population his¬
tory. Such studies can provide substantially new
insights into the processes responsible for shaping the
spatial patterns of genetic variation within and among
populations as well as their distributions.
POPULATION STRUCTURE AT
REGIONAL SCALE
It is well known that troglobitic species (/.£., obligate
cave-dwellers) occupy very reduced distribution ranges,
sometimes limited to a given karst region or even to a
single cave system. Troglophilic species (i.e., species
able to live and reproduce in the cave habitat as well
as in surface habitats) comparatively show much wider
distribution ranges.
Related troglobitic species within a genus are often
geographically vicariant in different karst areas or cave
systems. A huge number of taxa in terrestrial and
aquatic organisms, such as carabid and leiodid beetles,
spiders, pseudoscorpions, millipedes, isopods, amphi-
pods, exemplify this situation, particularly in limestone
areas of temperate regions. These vicariant species rep¬
resent the outcome of a geographical speciation pro¬
cess that was initiated with the population genetic
structuring of an ancestral troglophilic species at the
regional scale. As already outlined, the amount of
population structure is strictly dependent on intrinsic
factors, such as dispersal ability of the organism
implied, and extrinsic factors limiting gene flow
between populations, such as geographic distance and
the extent of ecological and geographical barriers to
migration. Therefore, we cannot automatically expect
that different organisms experiencing similar evolu¬
tionary pathways show the same geographical pattern
of population structure. Careful genetic analyses are
required to reveal the occurrence of actual or potential
gene exchange between populations. In turn, these
analyses represent a tool to disclose historical relation¬
ships between populations and to test hypotheses on
processes generating spatially structured population
systems.
Potential or actual gene exchange between cave
populations, and their resultant structuring at a
regional scale, can be understood by illustrating study
cases involving rhaphidophorid crickets and nesticid
spiders living in Italy and the eastern United States. By
means of allozyme polymorphisms, levels of gene flow
between populations were evaluated for Nesticus cave
spiders (Cesaroni et al., 1981) and cricket populations
belonging to Dolichopoda laetitiae from the Italian pen¬
insula (Sbordoni el al., 1987, 1991, 2000) and to
Euhadenoecus puteanus and E. fragilis from the eastern
United States (Caccone and Sbordoni, 1987). The
enhanced degree of population fragmentation in these
organisms is chiefly the result of a gradual reduction
of gene flow between populations caused by bioclimatic
changes that occurred in the Pleistocene. Hence, levels
of gene flow ( N e m ), as measured on the basis of the
present patterns of genetic differentiation (Wright's Fst),
are supposed to reflect historical gene flow occurring
between cave and surface populations living in a contin¬
uum of wet or mesophyllous woody environments. The
amount of genetic structuring was estimated for multi¬
ple combinations of populations at different geographic
scales. Fst and N e m values were calculated for different
groups of populations by considering increasingly
wider geographic windows. The size of the geographic
window for the population sampling was progressively
enlarged from a minimum average population distance
of 12—14 km for Italian caves and 30—40 km for
FIGURE 1 The amount of gene flow ( N e m ), as measured on the
basis of the present patterns of genetic differentiation (Wright's F ST ),
was estimated for multiple combinations of populations at different
geographic scales, by considering geographic windows of wider and
wider size. The four curves outline the trends of Dolichopoda laetitiae,
Euhadenoecus fragilis, Euhadenoecus puteanus, and Nesticus eremita.
ENCYCLOPEDIA OF CAVES
POPULATION STRUCTURE
611
American caves to a maximum average distance of
larger than 200 km (Fig. 1). Moreover, in D. laetitiae,
estimates of Fg T and N e m have also been performed at
smaller geographic windows by means of samples
caught in sites 50—500 m apart. At the same geo¬
graphic scale the extent of gene flow among E. putea¬
nus populations was higher than among Dolichopoda
populations, whereas the degree of population struc¬
turing in E. fmgilis is undoubtedly more than in both
D. laetitiae and E. puteanus. Population fragmentation
and geographical isolation in E. fragilis have already
proceeded to a very large extent, because almost irrele¬
vant gene flow is detectable at any scale. This outcome
is in agreement with the well-known features of these
three species. Both American species are found in caves,
but E. puteanus also commonly occurs outside caves in
forest habitats, whereas E. fragilis is a strictly cavernico-
lous species inhabiting a patchy limestone area.
Dolichopoda laetitiae shows halfway troglophilic habits
with respect to the two American species. Moreover,
troglomorphic features are more advanced in E. fragilis
than in E. puteanus, with D. laetitiae intermediate both in
the degree of leg elongation and in the rate of pigment
reduction. The general emerging picture corroborates
the idea that gene flow with epigean populations pre¬
vents or retards adaptation to the cave habitat, and that
isolation is an important prerequisite for the evolution
of troglomorphic adaptations.
In the European spider species, N. eremita, the rela¬
tion between N e m estimates and geographic distance
was similar to that observed for D. laetitiae. Nesticus
eremita is a widespread troglophilic species, colonizing
not only natural limestone caves, but also artificial
caves, cellars, and some epigean habitats, just like
D. laetitiae. Moreover, this spider can move outside
caves and disperse to some extent by means of bal¬
looning at juvenile stages. Therefore, both present and
past gene flow can explain the pattern of genetic dif¬
ferentiation between populations, which typically
reflects a pattern of isolation by distance.
The comparison to the examined cricket species indi¬
cates that the dispersal ability for N. eremita is lower than
that for E. puteanus, perhaps lesser than for D. laetitiae,
but undoubtedly higher than for E. fragilis. The study
cases examined here proved that different cave species,
even if closely related, could show different levels of iso¬
lation, fragmentation, and dispersal, depending on both
extrinsic obstacles to gene flow and intrinsic properties of
organisms. In some cases, different molecular markers,
used to investigate population genetic structures, or tech¬
nical procedure of data analysis may lead to slightly
different conclusions or may make the comparison of pat¬
terns of structuring difficult.
An example is provided by a species group of
Appalachian cave spiders, N. tennesseensis "complex,"
where population genetic structure was investigated
using mtDNA sequence data (ND1 gene). Hedin, in a
1997 paper, described high sequence homogeneity
within populations with large divergences between
populations for all species analyzed.
The revealed pattern was consistent with an ances¬
tral interbreeding population, which, at some time in
the past, was fragmented into several habitat-limited
subpopulations. Subsequent to fragmentation, the
exchange of migrants among subpopulations became
extremely restricted to nonexistent. The conclusion
was that Appalachian Nesticus species are character¬
ized by similar and essentially complete population
subdivision regardless of difference in species habitat
preference (surface, cave dwelling, troglobitic).
However, this outcome is in disagreement with many
other studies on cave animals, essentially based on
allozyme data, where troglobitic species showed lower
rates of gene flow than troglophilic or epigean species.
Hedin underlined that mtDNA, which is maternally
inherited, could reveal a reduced amount of gene flow
with respect to estimates based on nuclear loci. This
would explain the discrepancy between levels of gene
flow estimated in the American species of N. tennes¬
seensis complex and the European N. eremita. In addi¬
tion, Hedin suggested that studies on population
structure based on allozymes have to take into account
that populations currently exchanging no genes can
still share ancestral allozyme polymorphisms, that is,
the calculated extent of gene flow could be actually
interpreted as the trace of historical gene flow between
populations, as already outlined.
Histories of repeated isolation events leading to
populations structuring and relictual populations con¬
fined in subterranean habitats are frequently invocated
to explain the formation of species complexes in caves
according to an allopatric speciation model. Conversely,
a recent study by Niemiller et al. (2008) on Tennessee
cave salamander complex belonging to the genus
Gyrinophilus gave support to the hypothesis of a parapa-
tric speciation mechanism to explain the origin of the
cave populations from spring salamanders. The previ¬
ous classical hypothesis, based on the species distribu¬
tion, postulated that the subterranean species evolved
from an epigean, metamorphosing ancestor similar to
present-day G. porphyriticus during the Pleistocene as
climatic conditions forced surface populations at the
periphery of the species' range underground, thus iso¬
lating and facilitating speciation and evolution of troglo¬
morphic characters as predicted by the allopatric,
climate-relict model. Phylogenetic reconstructions based
on mitochondrial and nuclear DNA sequences indicated
that all three subterranean forms ( Gyrinophilus palleucus
palleucus, G. p. necturoides, and G. gulolineatus) are
recently derived from the surface-dwelling species
ENCYCLOPEDIA OF CAVES
612
p
G. porphyriticus and genealogically nested within, but an
allopatric speciation model, implying the isolation of
populations, should be rejected. In fact, the results of coa-
lescent-based analysis of the distribution of haplotypes
among groups supported a specification model implying
a continuous or recurrent genetic exchange during diver¬
gence (parapatric speciation).
SURFACE VERSUS UNDERGROUND
DISPERSAL ROUTES
How can extrinsic and intrinsic factors interact to
influence patterns of structuring populations? An
additional example is useful. The genetic structures
of populations were compared in two pairs of prey-
predator species in two regions of Kentucky: (i) the
cave cricket, Hadenoecus subterraneus, and the carabid
beetle Neaphaenops tellkampfi, which is a predator of
Hadenoecus eggs; and (ii) the cave cricket, H. cumberlan-
dicus, and its egg predator Darlingtonea kentuckensis.
The two regions where the species occur show differ¬
ent extents of fragmentation in the karst cover, with
the latter much more dissected into relatively smaller
limestone fragments. This difference is expected to
affect migration via underground routes more than
surface routes, because substantial forest cover was
almost continuous in both regions, at least in historical
times. This is why population fragmentation and
divergence could be strongly affected by extrinsic
barriers.
Results from allozymic analyses showed a similar
overall degree of genetic structuring in the two cave
cricket species (average Fs\S were 0.58 and 0.46 in
H. subterraneus and H. cumberlandicus, respectively). On
the other hand, although N. tellkampfi populations
showed a degree of genetic structuring (F g T = 0.56)
similar to cricket species, D. kentuckensis populations
displayed much higher values of genetic structuring
(Fst = 0.96, Nm = 0.01), suggesting that some of them
are already evolving into reproductively isolated spe¬
cies (Kane et al., 1992).
POPULATION STRUCTURE IN AQUATIC
TROGLOBITES: HYDROLOGICAL
ROUTES VERSUS ADAPTIVE DIVERSITY
In aquatic organisms, hydrologic relationships
appear to reflect population connectivity and gene
flow much better than geographic distance. In some
cases populations from different caves exposed to the
same groundwater aquifer may show no trace of
genetic differentiation, suggesting a single panmictic
gene pool, as reported in several crustaceans and fish.
In other cases, strong population subdivisions occur
even over very short geographic distances.
When groundwater habitats show high levels of
endemism, the most likely cause is strong hydrogra¬
phical isolation and low dispersal ability of inhabiting
species. Therefore, groundwater species are often con¬
stituted by highly subdivided populations and it is
also not rare to find cryptic species, especially
when habitat is fragmented. These general hypotheses
can be better understood by describing case studies
involving different taxa from Crustacea living in
the United States (Gammarus minus), in the Balkan
Peninsula (Troglocaris anophthalmus), and in France
(Niphargus virei).
The Crustacean, Gammarus minus, has been the
object of several detailed investigations by Culver et al.
(1995). This freshwater and cave-dwelling amphipod
species is found in caves throughout its range, but
morphologically highly modified cave populations are
found only in two relatively small regions of Virginia
and West Virginia in the United States. Populations of
G. minus reside in resurgences and related subsurface
basins, both in cave stream habitats and occasional
openings to the surface (karst windows). Genetic anal¬
ysis of allozyme polymorphisms pointed out that the
hydrological relationships among populations have
largely determined the levels of genetic differentiation.
Actually, cave and karst-window populations within a
given basin are similar to each other. These results
have been confirmed by molecular sequence data from
two loci, the nuclear gene ITS-I (internal transcribed
spacer) and the mitochondrial one, CO-I (cytochrome
oxidase I). Molecular data also indicated (a) much
lower genetic variation within cave populations than
within the surface ones and (b) a codon usage bias sig¬
nificantly lower in caves, suggesting that the cave colo¬
nization is rather recent. Overall data suggested limited
gene flow among populations and cave populations
more prone to bottleneck than spring populations,
likely resulting from differences such as temperature
variability and drought conditions between the two
types of habitats (Carlini et al., 2009).
On the other hand, the analysis of morphological
variation in G. minus generated a different pattern. In
this case, populations are similar by habitat rather than
by basin. Populations from cave and resurgence are
markedly different, and karst-window populations
exhibit a wide range of eye sizes, from very small (a
troglomorphic character) to the sizes seen in resur¬
gence populations. Both troglomorphic feature varia¬
tions and genetic structure of populations suggest
independent invasion of subsurface basins, and the
overall similarity of eyes in cave populations relative
to resurgence populations identifies a role for natural
selection.
ENCYCLOPEDIA OF CAVES
POPULATION STRUCTURE
613
A significant correlation between allozyme and
morphological distances was also found, suggesting
that morphological variation among populations is
largely influenced by evolutionary history. However,
there is no significant path linking habitat and genetic
distances, indicating that selection has little influence
on genetic structure. The significant path linking habi¬
tat distances and morphological distances does indi¬
cate that selection has a strong effect on morphological
variation among populations. A series of several inter¬
related analyses led to the conclusion that G. minus is
actually a species complex; populations in resurgence
habitats constitute one species, and populations in dif¬
ferent basins may be as many troglomorphic species as
independent isolations occurred in different basins.
The Dinaric karst area in the Balkan Peninsula hosts
the richest subterranean fauna in the world with several
endemisms and cryptic species as shown by different
studies based on molecular data. One of these species is
represented by Troglocaris anophthalmus, a cave shrimp
distributed in a large range of more than 500 km.
Morphologically, all populations are very uniform, but
on a genetic point of view, different studies carried out
on allozymes (Cobolli Sbordoni et al., 1990) and on
molecular analyses (Zaksek et ah, 2009) demonstrated
that the taxon is not homogeneous but composed by
different genetic units. In particular, considering the
study of Zaksek et al., based on molecular sequences
from mitochondrial (COI and 16S rRNA) and nuclear
(ITS2) genes analyzed in several samples collected
throughout the range, several (four—five) monophyletic,
geographically defined phylogroups can be observed.
Moreover, two of these groups showed biological repro¬
ductive isolation in sympatry. These results confirmed
the nondispersal hypothesis of subterranean fauna
although the southern Adriatic phylogroup showed a
pattern of recent dispersal across 300 km of hydrogra-
phically fragmented karst area. The authors explain this
result by combining the specific geomorphology of the
Dinaric Mountains with karst hydrology. Caves inhab¬
ited by Troglocaris are separated by flat depressions
(poljes), on insoluble ground, that are periodically
flooded. During periodical floods, aquatic cave animals
are frequently washed out of their subterranean habitat
and they could reach caves at other parts of a polje and
eventually reach a subterranean connection to an adja¬
cent polje. In this way, flooded poljes could act as
occasional stepping stones where no permanent hydro-
logical connection exists.
A phylogeographical analysis has been carried
out on the karstic groundwater dwelling amphipod
Niphargus virei, distributed in the Netherlands, Belgium,
and eastern France (Lefebure et ah, 2006). Two indepen¬
dent molecular loci, one nuclear (28S rRNA) and one
mitochondrial (COI), were used in specimens collected
throughout the species range. Results from both genes
support the same tripartite structure: a northern group
from Belgium and the Netherlands, a central eastern
group in France, and a more western group including
populations from North and South France. These
lineages appear to be rather ancient, being separated at
least 13 Ma. From morphological point of view all
populations belonging to Niphargus virei are extremely
homogeneous although allometric variation has been
observed between the two genetic units in France.
These results argue for the occurrence of deep cryptic
diversity in N. virei and demonstrate either a long mor¬
phological stasis in the evolution of the three lineages
or that extensive convergent evolution occurred in this
taxon. To solve this issue, a phylogenetic study includ¬
ing the best possible sampling of closely related taxa
consisting also of surface taxa, a reconstruction of the
ancestral morphology of N. virei, and subsequent
hypothesis about evolutionary trajectories would be
necessary.
Although hydrologic relationships are quite deter¬
minant for population connectivity and levels of gene
flow, in some cases, a similar or even identical habitat
fragmentation could produce a different pattern of
population structuring and promote differently genetic
divergence and speciation in relation to some features
of the organisms. Guzik et ah (2009) found a quite
interesting evidence of these differences by employing
a comparative phylogeographical and population
genetic approach to investigate the origins of a sym-
patric sister species triplet of diving beetles. The three
species are found within a single calcrete aquifer in
Western Australia and differ in size class: Paroster
macrosturtensis is mm, P. mesosturtensis ^2 mm, and
P. microsturtensis ~1.8 mm. Previous morphological
and molecular studies showed that these three taxa
form a reciprocally monophyletic clade among other
stygobiontic and epigean Paroster species; thus they are
sister taxa. Two plausible hypotheses about the popu¬
lation structures of these beetles' species can be made
depending on the role of calcrete environment in limit¬
ing gene flow between populations. First, if the
groundwater matrix is the same throughout the cal¬
crete and the beetles are highly dispersive, then their
populations should be panmictic. Alternatively, if the
calcrete environment is heterogeneous with the beetles
dispersing locally, then patterns of strong population
genetic structure caused by abiotic barriers, isolation
by distance, or genetic drift should be observed. The
outcome from population analyses, based on mito¬
chondrial DNA sequence data (COI), showed that each
species experimented multiple expansion events and
that spatial heterogeneity in the distribution of genetic
variation occurred both within and among the three
taxa. A significant fine-scale differentiation with
ENCYCLOPEDIA OF CAVES
614
p
isolation by distance was found for P. macrosturtensis
and P. mesosturtensis, but not for the smallest species
P. microsturtensis. The lack of population structure and
isolation by distance in P. microsturtensis most proba¬
bly reflects its smaller size and better dispersal capabil¬
ity than the other two species in the calcrete
environment. Such calcrete matrix, in fact, varies in
physical structure and connectivity (pore size of cre¬
vices or holes) and could play a role in influencing
levels of gene flow and population genetic structuring.
POPULATION STRUCTURE IN AQUATIC
TROGLOBITES: HISTORICAL
DETERMINANTS AND
SECOND ARY CONTACT
In several instances, an apparent population structur¬
ing can be determined to some extent by a genetic
admixture between already differentiated gene pools. In
tropical karst areas rapid evolution of drainage basins
combined with tectonic events may alter connections of
surface and cave streams, thus producing either isola¬
tion or secondary contact between biota adapted to dif¬
ferent conditions. The Cueva de Los Camarones in
Chiapas, Mexico, offers an excellent example of such
situations (Cesaroni et al., 1992; Allegrucci et ah, 1992).
Two undescribed species of Procambarus crayfish
belonging to the P. mimndai species group inhabit the
subterranean stream. They were roughly distinguish¬
able only by comparing extreme phenotypes, ranging
from dark, thick, eyed, surface-dwelling-like individuals
to light, elongate, micro-ophthalmic, cave-dwelling-like
individuals. Analyses of allozyme polymorphisms and
morphometry were performed to enlighten evolution¬
ary relationships among individual crayfish and to
explain patterns of microgeographic variation previ¬
ously revealed along the cave stream. Results from mul¬
tivariate morphometric analyses showed a real
discontinuity between the two species mainly deter¬
mined by the shape of the rostrum, chelae, and telson.
Moreover, these same characteristics exhibited clinical
variation within the less cavernicolous species. The
genetic structure of the two species was investigated at
23 enzyme loci, revealing unusually high levels of het¬
erozygosity in both species. Results of analyses on indi¬
vidual allozymic profiles corroborated morphometric
results, yielding a genetic distance of D = 0.26 between
the two gene pools. Due to the occurrence of alternative
alleles, we could quantify patterns of introgression that
reveal absence of FI individuals and asymmetric gene
flow between the two species. In the light of these data,
the observed microgeographic variation in morphology
within one of the two species, as well as the occurrence
of aberrant phenotypes, could be interpreted as the out¬
come of introgression.
Studies on cave organisms based on DNA microsa¬
tellites are scantier than those based on sequencing, in
part because of the technical difficulties named above.
These markers are, however, very informative when it
comes to understanding processes at the population
level, as attested by the study published by Strecker
et al. (2003) on the neotropical fish Astyanax fasciatus
(Characidae), where both surface and cave populations
are known. The authors analyzed simultaneously four
cave and four surface populations; cave populations
all showed extremely low genetic variability, which
most likely resulted from bottleneck events. No appre¬
ciable extant gene flow was detected between surface
and cave populations, the origin of which is likely
the result of independent invasions from surface
populations.
POPULATION STRUCTURE AND
HABITAT HETEROGENEITY
The hypothesis that habitat heterogeneity affects pop¬
ulation structure was investigated in the D. laetitiae
population inhabiting the Cerveteri's Etruscan necropo¬
lis near Rome, Italy. This is a well-known necropolis
extending throughout a roughly elliptical area of 70 ha.
The necropolis includes approximately 300 tombs of dif¬
ferent sizes and locations. Small colonies of Dolichopoda
inhabit most of these tombs, raising the following ques¬
tions: How many Dolichopoda populations inhabit the
necropolis? Is there a case for a metapopulation (i.e., a
population of populations) or, alternatively, for a
unique population? To what extent does this peculiar
habitat structure affect the population genetic structure?
A. Sansotta and the authors of this article faced
these questions by means of a long-term series of eco¬
logical and genetic investigations to measure the dis¬
persal ability of crickets across tombs, to estimate the
effective population size, to determine the amount of
heterozygosity of the population, and to assess
whether Dolichopoda genotypes are randomly distrib¬
uted throughout the necropolis.
Extensive fieldwork was carried out by means of
mark—recapture techniques based on individual tag¬
ging. Both allozyme polymorphisms and RAPD-DNA
markers were employed to carry out genetic analysis. As
a first result from this study, crickets revealed remark¬
able dispersal ability as they proved to move signifi¬
cantly across tombs to such an extent that the necropolis
can be considered as the home range for a panmictic
population as a whole. Second, a population sample col¬
lected across the necropolis exhibited relatively high
values of heterozygosity ( H 0 = 0.218; H e = 0.244)
ENCYCLOPEDIA OF CAVES
POPULATION STRUCTURE
615
compared to other Dolichopoda. Third, individuals carry¬
ing definite allele combinations (genotypes) were located
preferentially in some tombs within the necropolis to
such an extent that genotypes in different partitions of
the necropolis were not randomly distributed, but were
significantly associated with some cave ecological
descriptors, namely, temperature. In summary, these
crickets appeared to respond to the necropolis spatial
and environmental heterogeneity by means of habitat
choice expressed by different genotypes. Therefore, mul¬
tiple niche selection associated to habitat choice appears
to be the most probable selective process to explain the
observed high level of genetic polymorphism.
POPULATION STRUCTURE AND FOOD
RESOURCE HETEROGENEITY
The Frasassi Cave System is located in Central Italy,
on the Adriatic side of the Apennine Mountains. This
karst system is developed in at least four main horizon¬
tal levels. Each level shows a complex pattern, strongly
influenced by faults and by a hydrogeological setting.
The system is characterized by the occurrence of sulfide
streams in some sections and by deposits of guano in
others. The oniscidean isopod, Androniscus dentiger, is
the most abundant species in different habitats in the
Frasassi caves. Gentile and Sarbu (2004) studied the
possible occurrence of genetic structuring in this species
in population samples collected at different sites by
means of Fst analyses based on 18 allozyme loci.
Among other samples collected throughout different
caves in this system, three neighbor samples were col¬
lected within Grotta del Fiume (i.e., Guano room. Green
Lake room, and along a lateral sulfide section). Results
indicated departures from the Hardy—Weinberg equi¬
librium at several loci, generally due to heterozygote
deficiencies. The Fst value was 0.180, indicating a level
of genetic structuring comparable to values found
among geographically distant populations of the same
species. However, even by removing the Grotta del
Mezzogiorno sample from the analysis, the amount of
population subdivision remained unexpectedly high
(Fst = 0.180), especially considering the neighborhood
of sampling locations within the same cave. Such an
extent of population structure implies very low levels
of gene flow, as shown between the Green Lake room
and the lateral sulfide section, although each of these
demes shows a limited gene exchange with the subpop¬
ulation from the Guano room. Interestingly enough,
compared to other samples, individuals collected in the
sulfide section were proved to feed on different types
of resources such as chemoautotrophically synthesized
food, as established by stable isotope ratio analysis.
Again we are facing a situation where natural selection
could be involved in promoting and maintaining popu¬
lation structuring in A. dentiger in the Grotta del Fiume.
Caves are often expected to be very homogeneous
environments as a result of their physical parameters and
habitat stability. This may not necessarily be true, at least
in instances where localized trophic resources may con¬
vert a cave into a patchy habitat, constraining colonies to
set around resources, in spite of the potential dispersal
capability of individuals. If trophic resource heterogene¬
ity remains stable over time, we could expect that genetic
polymorphism is maintained by natural selection. This
interpretation explains why several cave-dwelling organ¬
isms show relatively high heterozygosity levels notwith¬
standing their reduced population sizes.
GENETIC VARIABILITY, POPULATION
SIZE, AND NATURALSELECTION
The amount of genetic variability expressed by
troglobitic populations is a controversial issue. In fish,
for example, a recent study based on RAPD markers
suggests that hypogean populations of balitorid fish
have lower genetic variability than related surface
populations. Similar results were obtained from
allozymic studies in the Mexican cavefish Astyanax fas¬
cia tus, amblyopsid fishes from North America, and a
trichomycterid species from Venezuela. The decreased
genetic variation observed was reputedly consistent
with the expectation that the troglobitic fish have smal¬
ler population sizes than the epigean species. Limited
to single caves and cave systems, they supposedly have
small population sizes. However, in a study of hypo¬
gean cyprinid fish from Somalia (Fig. 2), Cobolli et al.
FIGURE 2 Somali cave fish: Barbopsis devecchii and Phreatichthys
andruzzii. Photos from Ercolini et al. (1982).
ENCYCLOPEDIA OF CAVES
616 p
(1996) showed that populations of Phreatichthys andruz.- its closest relative, Barbopsis devecchii, a micro-ophthal-
zii, a troglomorphic stygobiont species, are more hetero- mic, less specialized fish also occurring in the ground-
zygous than most epigean cyprinid species studied waters of Somalia, statistically significantly higher
thus far. In addition, when comparing this species with heterozygosity values were revealed ( H e ranging from
TABLE 1 Estimates of Heterozygosity in Troglomorphic and Nontroglomorphic Cave Crickets
Orthoptera
No. Species
No. Pops.
Average H e
Range of H e
Noctivox
1
1
0.18
—
Longuripes
1
1
0.253
—
Hadenoecus
5
18
0.078
0.020-0.130
Euhadenoecus
4
24
0.064
0.030-0.110
Ceuthophilus
1
7
0.026
—
Dolichopoda
10
52
0.144
0.056-0.209
Troglophilus
9
19
0.06
0.000-0.178
TABLE 2 Estimates of Heterozygosity in Different Troglobitic Taxa
Crustacea
No.
No.
Average
Range of
Species
Pops.
He
He
Amphipoda
Crangonyx
1
6
0.118
—
Gammarus
1 +
8
0.108
0.075-0.130
Niphargus
4
9
0.274
0.104-0.347
Isopoda
Androniscus
5+
34
0.102
0.027-0.178
Trichoniscus
2+
12
0.089
0.034-0.218
Stenasellus
4
13
0.079
0.000-0.196
Proasellus
1 +
3
0.084
0.037-0.148
Typhlocyrcilana
2
3
0.04
0.029-0.061
Monolistra
3
3
0.28
0.261-0.316
Coleoptera
Carabidae
Duvalius
2
3
0.114
0.100-0.141
Neaphaenops
1
8
0.192
0.173-0.222
Dnrlingtonea
1
10
0.009
0.000-0.040
Pseudanophthalmus
3
13
0.094
0.053-0.170
Cholevidae
Bathysciola
1
1
0.121
—
Speonomus
5
44
0.113
0.060-0.192
Orcistygia
2
2
0.165
0.133-0.198
Leptodirus
1
1
0.168
—
Ptomaphagus
1
6
0.056
0.012-0.099
ENCYCLOPEDIA OF CAVES
POPULATION STRUCTURE
617
0.046 to 0.062 in Phreatichthys and from 0.014 to 0.020 in
Barbopsis). Due to the lack of appropriate population
estimates, we cannot test the hypothesis that these dif¬
ferences are related to population sizes. In Phreatichthys
it has been hypothesized that selective advantage for
individuals with high heterozygosity could be
involved. Negative correlations have been found in var¬
ious organisms between individual heterozygosity
levels and fitness components such as rate of oxygen
consumption, energy requirements for maintenance.
These and other metabolic features may represent
adaptations to low-energy input conditions in troglo-
bites and susceptibility to affect selective advantage to
a considerable extent. Comparing population struc¬
tures of Phreatichthys to Barbopsis, it was revealed that
the former is strongly structured into genetically iso¬
lated populations while consistent gene flow remains
genetically homogeneous in even geographically dis¬
tant populations of Barbopsis, indicating an earlier
cave isolation of Phreatichthys populations. On the
contrary, Barbopsis could have experienced a longer
direct contact with their epigean relatives. Because
isolation in the cave habitat by means of breakdown
of gene flow with epigean populations is a prerequisite
to enhance adaptation of cave populations, it could be
speculated that old established cave populations have a
higher probability of expressing fitness-related features in
their gene pools such as increased heterozygosity. The
tenet that troglobites or troglomorphic cave dwellers have
low genetic variability may not be generally applicable. A
comparison between troglomorphic and nontroglo-
morphic cave crickets like Dolichopoda versus Troglophilus,
Hadenoecus versus Ceuthophilus, and Longuripes versus
Noctivox leads to the invariable result that the former have
higher heterozygosity than the latter (Table 1). Moreover it
can be seen that, among other taxa, several stygobiontic
crustaceans such as amphipods and isopods as well as tro-
globitic beetles show evidence for very high levels of het¬
erozygosity (Table 2).
These findings can hardly be explained by large
population sizes, but they could reflect the existence
of genetic polymorphisms maintained by balancing
selection. Interestingly, high heterozygosity levels
have been revealed mainly in small arthropods,
which appear to perceive the environmental patchi¬
ness as coarse grained and therefore favorable in
which to adapt to habitat and/or resource heteroge¬
neity by means of a multiple niche polymorphism
strategy.
See Also the Following Article
Life History Evolution
Bibliography
Allegrucci, G., Baldari, F., Cesaroni, D., Thorpe, R. S., & Sbordoni, V.
(1992). Morphometric analysis of interspecific and microgeo¬
graphic variation of crayfish from a Mexican cave. Biological
Journal of the Linnean Society, 47, 455—468.
Caccone, A., & Sbordoni, V. (1987). Molecular evolutionary diver¬
gence among North American cave crickets. I. Allozyme varia¬
tion. Evolution, 41, 1198—1214.
Carlini, D. B., Manning, J., Sullivan, P. G., & Fong, D. W. (2009).
Molecular genetic variation and population structure in morpho¬
logically differentiated cave and surface populations of the fresh¬
water amphipod Gammarus minus. Molecular Ecology, 18,
1932-1945.
Cesaroni, D., Allegrucci, G., Caccone, A., Cobolli Sbordoni, M., De
Matthaeis, E., Di Rao, M., et al. (1981). Genetic variability and
divergence between populations and species of Nesticus cave spi¬
ders. Genetica, 56, 81 —92.
Cesaroni, D., Allegrucci, G., & Sbordoni, V. (1992). A narrow hybrid
zone between two crayfish species from a Mexican cave. Journal
of Evolutional!/ Biology, 5, 643—659.
Cobolli, M., De Matthaeis, E., Mattoccia, M., & Sbordoni, V. (1996).
Genetic variability and differentiation of hypogean Cyprinid
fishes from Somalia. Journal of Zoological, Systematic and
Evolutionary Research, 34, 75— 84.
Cobolli Sbordoni, M., Mattoccia, M., La Rosa, G., De Matthaeis, E., &
Sbordoni, V. (1990). Secondary sympatric occurrence of sibling
species of subterranean shrimp in the karst. International Journal of
Speleology, 19, 9—28.
Culver, D. C., Kane, T. C., & Fong, D. W. (1995). Adaptation and natu¬
ral selection in caves. Cambridge, MA: Harvard University Press.
Ercolini, A., Berti, R., Chelazzi, L., & Messana, G. (1982). Researches
on the phreatobitic fishes of Somalia: Achievements and pro¬
spects. Monitore Zoologico Italiano, 17(9 (N.S. Supplement)),
219-241.
Gentile, G., & Sarbu, S. M. (2004). Trophic sources partition and pop¬
ulation genetic structure of the isopod Androniscus dentiger from a
chemoautotrophy based underground ecosystem. Subterranean
Biology, 2, 7—14.
Guzik, M. T., Cooper, S. J. B., Humphreys, W. F., & Austin, A. D.
(2009). Fine-scale comparative phylogeography of a sympatric sis¬
ter species triplet of subterranean diving beetles from a single cal¬
crete aquifer in Western Australia. Molecular Ecology, 18,
3683-3698.
Hedin, M. C. (1997). Molecular phylogenetics at the population/
species interface in cave spiders of the Southern Appalachians
(Araneae: Nesticidae: Nesticus). Molecular Biology and Evolution,
14, 309-324.
Kane, T. C., Barr, T. C., Jr, & Badaracca, W. J. (1992). Cave beetle
genetics: Geology and gene flow. Heredity, 68, 277—286.
Lefebure, T., Douady, C. J., Gouy, M., Trontelj, P., Briolay, J., &
Gibert, J. (2006). Phylogeography of a subterranean amphipod
reveals cryptic diversity and dynamic evolution in extreme envir¬
onments. Molecular Ecology, 15, 1797—1806.
Niemiller, M. L., Fitzpatrick, B. M., & Miller, B. T. (2008). Recent
divergence with gene flow in Tennessee cave salamanders
(Plethodontidae: Gyrinophilus) inferred from gene genealogies.
Molecular Ecology, 17, 2258—2275.
Sbordoni, V., Allegrucci, G., Caccone, A., Carchini, G., & Cesaroni,
D. (1987). Microevolutionary studies in Dolichopodinae cave
crickets. In B. Baccetti (Ed.), Evolutionary biology of orthopteroid
insects (pp. 514—540). Chichester, UK: Horwood.
Sbordoni, V., Allegrucci, G., & Cesaroni, D. (1991). A multidimen¬
sional approach to the evolution and systematics of Dolichopoda
cave crickets. In G. M. Hewitt et al. (Eds.), Molecular techniques in
ENCYCLOPEDIA OF CAVES
618
p
taxonomy (NATO ASI Series, Vol. H 57, pp. 171—199). Berlin:
Springer-Verlag.
Sbordoni, V., Allegrucci, G., & Cesaroni, D. (2000). Population
genetic structure, speciation and evolutionary rates in cave dwell¬
ing organisms. In H. Wilkens, D. C. Culver & W. F. Humphreys
(Eds.), Subterranean ecosystems (pp. 453—477). Amsterdam:
Elsevier.
Strecker, U., Bernatchez, L., & Wilkens, H. (2003). Genetic divergence
between cave and surface populations of Astyanax in Mexico
(Characidae, Teleostei). Molecular Ecology, 12, 699—710.
Zaksek, V., Sket, B., Gottstein, S., Franjevic, D., & Trontelj, P. (2009).
The limits of cryptic diversity in groundwater: Phylogeography
of the cave shrimp Troglocaris anophthalmus (Crustacea: Decapoda:
Atyidae). Molecular Ecology, 18, 931—946.
POSTOJNA—PLANINA CAVE SYSTEM,
SLOVENIA
Stanka Sebela
ZRC SAZU Karst Research Institute
PHYSICAL SETTING
Hydrology and Topographic Setting
The Postojna Cave System is the longest (20,570 m
long; 115 m deep) known cave in the classic karst of
Slovenia. The term karst is derived from this region.
The Postojna Cave is connected to the Planina Cave
(6.656 km long; 65 m deep) by unknown passages.
There remains about 2000 m of unexplored separation,
although an underground water connection has been
known since the explorations of Gruber in 1781.
The Pivka River flows on impermeable Eocene
flysch and sinks into the underground system of the
Postojna Cave System at an elevation of 511 m on the
contact with Cretaceous limestones (Fig. 1). Several
kilometers downstream, the underground Pivka joins
the Rak River passage in the Planina Cave System and
FIGURE 1 Ponor entrance of Pivka River into Postojna Cave
System at 511 m above sea level. Photograph by the author.
emerges from the cave entrance to form the Unica
River at an elevation of 453 m. The average flow of the
Pivka at the swallow point is 5.26 m 3 s -1 . At Postojna
the average annual precipitation is 1565 mm (years
1982—2000). Dye-tracing experiments in 1988 provided
evidence for the bifurcation of the Pivka, for the surfi-
cial river also drains toward the sources of the Vipava
River and thus forms part of both the Adriatic and the
Black Sea drainage basins. The portion of the Pivka
that enters the Postojna—Planina Cave System belongs
to the Black Sea drainage basin.
The land surface above the Postojna—Planina Cave
System can be divided into two erosional levels. The
lower level is above the Postojna Cave System and
reaches 632 m. The land surface is higher than 630 m
above Planina Cave and can reach 753 m. It is a typical
karst surface with numerous dolines and collapse
dolines. There are also some horizontal caves, but ver¬
tical caves prevail and some end very close to under¬
ground Pivka River passages. The landscape was used
as pasture about 100 years ago. Today it is covered by
forest.
Geologic Setting
Slovenia is situated at the border between the Adria
microplate and the Eurasia plate and characterized by
complex and neotectonically active geological conditions.
Since the late Miocene to Pliocene paleomagnetic
data have indicated about 30° counterclockwise rota¬
tion of Adria microplate. During the Miocene to recent,
the thrust belts along the Adria margin include the
Dinaric thrust system, the South-Alpine thrust system,
and the Dinaric faults. The Dinaric faults cut and dis¬
place both Dinaric and South-Alpine fold-and-thrust
structures. Most of them are characterized by moderate
historic and recent seismicity.
The area is part of the Javorniki—Sneznik thrust unit
which has been overthrust over the Eocene flysch. The
Hrusica thrust unit, which is upper Triassic dolomite,
overthrusts the Javorniki—Sneznik thrust unit (Fig. 2).
Overthrusting took place after the deposition of the
Eocene flysch. During the Miocene and Pliocene, the
overthrusting was accompanied by folding. The princi¬
pal folding deformation in Postojna Cave is the
Postojna Anticline (Fig. 3). Between the Postojna and
Planina Cave Systems is the Studeno Syncline (Car
and Gospodaric, 1984).
It is important to distinguish older overthrusting and
folding deformations from younger faulting deforma¬
tions. The Postojna—Planina Cave System is situated
between two regionally important faults with the
NW-SE Dinaric orientation. These are the Idrija Fault on
the north and the Predjama Fault on the south (Fig. 2).
Encyclopedia of Caves.
© 2012 Elsevier Inc. All rights reserved.
POSTOJNA—PLANINA CAVE SYSTEM, SLOVENIA
619
FIGURE 2 Geological position of Postojna—Planina Cave System, Slovenia. (1) Alluvium (Holocene) covering the karst polje of Panina, (2)
Eocene flysch, (3) Cretaceous limestone (Javorniki—Sneznik thrust unit), (4) Upper Triassic dolomite (Hrusica thrust unit), (5) thrust, (6) fault,
(7) direction of underground water flow.
The tectonic structure of the area between those two
faults has all the characteristics of the intermediate zone
between two dextral strike-slip faults. The cave passages
of the Postojna Cave System follow the strike and dip of
the bedding planes, especially those with interbedded
slips. They are developed in both flanks of the Postojna
Anticline. They also follow Dinaric and cross-Dinaric
(NE-SW) oriented fault zones and mostly north—south
oriented fissured zones (Sebela, 1998).
During the development stages of Postojna Cave
System, some of the geological structures were reacti¬
vated and appear to be connected with the formation of
collapse chambers. The geological structural elements
were used as pathways for cave development. Along
the same fault zone in the cave up to four different reac¬
tivations can be detected. The fault zone that runs from
Pisani Rov and through Velika Gora shows sinistral hor¬
izontal movement in Pisani Rov, reverse fault movement
at Velika Gora, and dextral horizontal movement and
normal fault movement in Lepe Jame (Fig. 3). The same
Dinaric oriented fault zone is monitored at two places
with TM 71 extensometers. The same fault zone is cut
by the cross-Dinaric oriented fault zone in Lepe Jame.
This neotectonic fault zone was active post-cave devel¬
opment because it cuts older cave sediments which are
at least 0.78—0.99 Ma in age.
On the southwestern flank of Postojna Anticline
Dinaric oriented fault zones prevail, while on the
northeastern flank the cross-Dinaric oriented fault
zones prevail. The underground River Pivka passage
follows the strike direction of bedding planes in the
southern part and bedding strike dip direction in the
northwestern part.
The Postojna—Planina Cave System is developed in
Cretaceous carbonate rocks. The passages of Postojna
Cave are developed in upper Cretaceous (Cenomanian,
Turonian, and Senonian) mostly bedded and thick-bedded
limestones (Fig. 4). The Cenomanian and Turonian
ENCYCLOPEDIA OF CAVES
620
20/10
130/90
340/10
Matevze>
Rov /
Magdalena
■^Jama
$^ 30 / 6 (
220/50'
230/60
230/40
.Otoska Jama
Tartarus
170/80
\ t
240/40
>40/50
Kristalni
Rov
Ml 0/90
130/70 90
Pivka Jama
130/90
20/20
-A
B2-B1
Crna Jama
130/40
330/10
y
°N 0
30 *0/70 £0 ^
J \ Carobni /*■
Vrt \// 130/90
v»HkSO^?50/10 Pisani
Rov
^ 30/60
B2-B1
230/70
60 70/70
RjTl 4
' BiospeleolosKa
* Rostaja
130/80 \
240/30
Se
200 m
Sebela S
C8 66
B6 B5
(B)
FIGURE 4 Lithological columns. (A) Postojna Cave System litho¬
logical column. (Bl) Thin-bedded limestone with cherts (100 m), (B2)
thick-bedded limestone (35 m), (B3) very thick-bedded limestone
(100 m), (B4) thick-bedded limestone with rudists (225 m), (B5) very
thick-bedded limestone (225 m), (B6) thick-bedded limestone (140 m).
(B) Surface lithological column over the Postojna Cave System. (Cl)
Thin-bedded limestone (165 m), (C2) thick-bedded limestone (50 m),
(C3) very thick-bedded limestone with Chondrodonta horizon (up to
110 m), (C4) thick-bedded limestone with rudists (290 m), (C5) very
thick-bedded limestone with rudists (420 m), (C6) thick-bedded lime¬
stone (130 m), (C7) conglomerate (1-2 m), (C8) flysch.
FIGURE 3 Geology of Postojna Cave System. (1) Cave passage
with underground River Pivka, (2) strike and dip direction of bedding
plane, (3) anticline, (4) strike and dip direction of fault, (5) dextral hor¬
izontal movement, (6) vertical movement, (7) TM 71 exstensometer,
subsidence of SW block.
limestones are more thin bedded and can include chert
lenses. The Senonian limestones are thick bedded to
massive. The cave passages are developed in about an
800-m-thick lithological column. The Planina Cave
System is developed in lower and upper Cretaceous
limestone and calcarenite with inliers of limestone,
which in the northeast part traverses into lower
Cretaceous dolomite.
Recent microtectonic deformations have been moni¬
tored continuously in 3D in Postojna Cave with TM 71
extensometers since 2004. Two instruments, 260 m
apart, were installed on the Dinaric oriented (NW-SE)
fault zone that is situated about 1000 m north of the
inner zone of the regionally important Predjama Fault.
The locations for installation of the exstensometers in
Postojna Cave were selected to evaluate if the monitored
fault is tectonically active and to determine whether the
tectonic activity has any influence on the speleogensis.
The two devices were installed in the same fault zone
but on different fault planes. The device at Velika Gora
was installed at the contact point between a fault plane
(dip angle 70—90° to the NE) and a collapse block,
partly covered with flowstone. The second device (Lepe
Jame) was installed in a narrow natural passage, which
was enlarged artificially 30 years ago, between two fault
planes that are about half a meter apart. The northern
fault plane dips to the NE at 60° and has a horizontal
ENCYCLOPEDIA OF CAVES
POSTOJNA—PLANINA CAVE SYSTEM, SLOVENIA
621
striation—dextral movement. The southern one consti¬
tutes the normal fault, dipping toward the NE by
80—90° and showing vertical striation. The device at
Velika Gora is 68 m below the surface and the device in
Lepe Jame is 60 m below the surface.
Monitoring on both instruments has shown small
tectonic movements, that is, a general dextral horizon¬
tal movement of 0.05 mm in 4 years (Velika Gora) and
extension of 0.03 mm in 4 years (Lepe Jame). Between
the longer or shorter calm periods, eleven extremes
have been recorded regarding characteristic changes in
displacement. The largest short-term movement was a
compression of 0.04 mm in 7 days, detected in March
2005, which coincided with the 25-km distant Ilirska
Bistrica earthquake (Ml = 3.9). About two months
before the earthquake, an extension of 0.05 mm
occurred and one month before the earthquake the
strain changed into a compression of 0.05 mm. The
largest permanent peak was detected at the end of
2004. Along the y-axis (Velika Gora), there was a dex¬
tral horizontal movement of 0.075 mm in one month
(November 10 to December 15, 2004). After the sinis-
tral horizontal movement of 0.02 mm (December
15—27, 2004), the y-axis retained its permanent position
on 0.05 mm where it remained until today. According
to speleogenesis, the monitored fault zone represents a
stable cave environment (Sebela et ah, 2010).
Other Interesting Research in the Postojna Cave
System
In Kristalni Rov more than 20-year-long studies of
percolation water from the surface through the 100-m-
thick limestone cave roof are going on. Polluted waters
rich in nitrates and chlorides continued for 11 years
after the source removal, while sulfates and phosphates
continued even longer (Kogovsek and Sebela, 2004).
In the catchment area of the Unica River, more com¬
bined tracer tests with fluorescent dyes have been per¬
formed to characterize the properties of groundwater
flow and transport of contaminants through the karst
system under different hydrologic conditions. During
the tracer test when injection was followed by rain, the
flow velocities were 88—640 m h 1 (Gabrovsek et ah,
2010 ).
Since 1932 in Postojna Cave in Tartarus passage, an
underground station with horizontal pendulums was
in operation to measure Earth tides. The study that
stopped before 1945 was important for both karst
hydrology and seismology.
In the past few years, the idea to establish an under¬
ground geophysical station in Postojna Cave has been
prevalent. The cave passages provide good under¬
ground conditions for such studies. First steps started
in 2004 when two extensometers TM 71 were installed
to measure microtectonic displacements in three
dimension. Since December 2009, data on Velika Gora
have been automatically taken from the TM 71 exten-
someter. Parallel with microtectonic displacements, the
radon air measurements are organized. There are con¬
nections between extension periods and higher radon
concentrations (Sebela et ah, 2010).
Since July 2010, the tilt data on a vertical 20-m-long
pendulum in Magdalena Jama have been collected by
Czech explorers. The two-dimensional optical mea¬
surement of tilt of the rock mass and continous digital
evaluation of the results are important to understand
the behavior of the stress changes within the rock
mass.
In the same place as the 1932 station with horizontal
pendulums, the accelerometer instrument was installed
within the common project with Italian colleagues in
the beginning of 2010. Our aim is to establish the first
underground seismological station in Slovenia with
online registration.
The Postojna Cave preliminary underground geo¬
physical station includes two TM 71 extensometers
(with air temperature, air pressure, and radon moni¬
toring), one vertical pendulum in Magdalena Jama,
and an accelerometer instrument in Tartarus passage.
Paleomagnetic properties and magnetostratigraphy
of karst sediments in Slovenia (Zupan Hajna et ah,
2008) indicated that the age of the oldest samples
ranged from about 1.8 to more than 5.4 Ma. The
study indicates that part of Postojna Cave System
has sediments being much older than 0.78 Ma.
Sediments from Planina Cave System are younger
than 0.78 Ma.
All karst caves in Slovenia are natural worths and
are the property of the state. The Postojna Cave
System is three-fourths managed by a private company
and one-fourth by the Comune of Postojna. Both need
to accomplish expert control and recommendations for
management of cave system and climatic and biologic
monitoring of the cave system. Expert control and
recommendations for sustainable management of natu¬
ral worths, formulation of suitable directives for sus¬
tainable use of natural worths, and climatic and
biologic monitoring of cave system need to be fulfilled.
Besides this the cave guardian collects data on old and
actual scientific and popular researches of the cave
system.
The first studies of the Postojna Cave underground
climate were published by Crestani and Anelli in 1939.
Air temperature in different parts of the cave, under¬
ground River Pivka temperature, and also temperature
of the rock were obtained. The ventilation and humid¬
ity in the cave were studied. The connections between
blowholes on the surface and cave passages were
described.
ENCYCLOPEDIA OF CAVES
622
p
A systematical study of underground meteorologi¬
cal conditions in Postojna Cave has been started in
2009 to determine human impact on natural worth.
Besides two meteorological stations (air temperature,
water temperature, humidity, wind direction, wind
speed, C0 2 ), 15—20 other places have been selected in
the Postojna Cave as monitoring sites where air tem¬
perature, air pressure, humidity, and wind are contin¬
uously measured. The cave is well ventilated deep
inside. The average air temperature (2009—2010) is
11.10°C on the top of Velika Gora and 10.66°C for Lepe
Jame (Fig. 5). For the same period the average air tem¬
perature on the surface outside Otoska Jama was
9.20°C.
HISTORY OF EXPLORATION
The oldest inscription on a cave wall is from 1213.
In 1818, the native Luka Cec discovered new parts of
the cave and this is the official year of discovery of the
tourist section of the Postojna Cave System. The
Postojna Cave System has a long history of tourism
(Shaw, 2000; 2010).
Valvasor in 1689 described the cave as one of the
most remarkable caves in the world. Hacquet provided
a description of the cave in his work Oryctographia
Carniolica written from 1778 to 1779. Nagel produced
the first cave map in 1748. After the first guidebook
written by Agapito in 1823, Hohenwart in his 1830/32
guidebook also explained the geology of the cave.
Schaffenrath provided cave illustrations. In 1854
Schmidl described the importance of the discovery of
new cave passages by Luka Cec. Schmidl is also the
author of the first monograph on the Postojna Cave
FIGURE 5 Lepe Jame is rich in speleotheme decorations.
Photograph taken by the author.
System in 1854. He was one of the first explorers of
some parts of the underground Pivka River in the
Postojna Cave System. In the Planina Cave System, he
accomplished the exploration of the Pivka Passage to
the sump and also part of the Rak Passage in 1850. In
1889 the first Slovene caving club, named "Anthron,"
was established in Postojna.
The first underground post office in Postojna Cave
was situated close to Kongresna Dvorana in the years
1899—1927. In 1927 it was removed to Koncertna
Dvorana.
Since 1872 the railway was in operation for tourists
in Postojna Cave. A small man-powered railway was
introduced for all who could afford it. In 1924 man¬
power was replaced by locomotive power (Fig. 6).
Permanent electric lighting has been installed since
1884, which means that Postojna Cave was the third in
the world to be lighted electrically.
The French speleologist E. A. Martel in his book Les
Abimes (1894) included many interesting explorations
and studies of the Postojna Cave System. In the begin¬
ning of the twentieth century, Perko was one of the lead¬
ing names connected with the Postojna Cave System. He
was a manager of the cave from 1909 to 1941.
The Speleological Institute was established in
Postojna in 1929. Its first task was to find new natu¬
ral caves and passages between Postojna and Planina
Cave systems and to connect the caves. Gallino and
Petrini surveyed the Postojna Cave System by theod¬
olite in 1924—1934. In 1947 the Karst Research
Institute of the Slovene Academy of Sciences and
Arts was established. Systematic speleological, geo¬
logical, morphological, and hydrological studies of
the Postojna Cave System and other caves were car¬
ried out.
FIGURE 6 The tourist railway runs under fallen stalagmite,
Postojna Cave System. Photograph taken by the author.
ENCYCLOPEDIA OF CAVES
POSTOJNA—PLANINA CAVE SYSTEM, SLOVENIA
623
DESCRIPTION OF THE CAVE
The Postojna Cave System has several known
entrances:
• Ponor entrance of the Pivka River (511 m)
• Entrances from the artificial platform (529.5 m)
• Entrance to the Otoska Jama (531.7 m)
• Magdalena shaft entrance (562 m)
• Crna Jama entrance (531 m)
• Pivka Jama entrance (550 m)
Besides known entrances, there are also other, less
known, accesses from the surface. Some are just wid¬
ened fissures not big enough for a man to pass, repre¬
senting breathing holes; others are small entrances used
100 or more years ago that are no longer accessible {eg.,
Matevzev Rov). There are also some old entrances that
are still completely filled with sediments as was the case
with entrances east of Biospeleoloska Postaja. And there
are examples where the karstological processes on the
surface draw the dolines or collapse dolines near the
underlying cave passages as is the case with Pisani Rov.
On the area of Postojna karst there are some hori¬
zontal caves and shafts that do not have the connection
with the Postojna Cave System but at least some of
them probably belong to the older stages of the speleo-
genesis, meaning that in the past they were connected
with active passages of Postojna Cave.
The area between the Postojna and Planina Cave
systems was always of great interest due to the
unknown underground connection between both
caves. The diving explorations in 1998 and 1999 pro¬
vided sufficient information to understand the under¬
ground Pivka flow toward the Planina Cave System,
where about 2 km of unknown passages still exist.
Postojna Cave is developed in two principal levels.
The higher level begins at an altitude of 529.5 m and is
represented by dry passages (Figs. 5 and 6) which are
filled with cave sediments, flowstone, and collapse
blocks. The lower level (Pivka sink at 511 m) contains
the underground channels of the Pivka River which
are shifted more to the southwest and northwest with
respect to the higher levels. Both levels are connected
by side passages. The river channels of Postojna Cave
are 10—20 m lower than the older dry passages. In
some parts, the phreatic passages are still preserved.
There are also paragenetic passages and collapse
chambers. The largest collapse chambers are aligned to
principal fault zones and developed in thick-bedded
limestone. Velika Gora with the volume of 240,000 m 3
is the biggest collapse chamber in Postojna Cave.
Studies of the cave sediments show that there were
ten principal development stages of the cave system
between the first sedimentary deposits and the
present. The stages have been determined by absolute
dating of flowstone and the relative ages of the cave
sediments (Gospodaric, 1976). The Postojna Cave
System was active at least 0.99 Ma ago. Samples from
a small natural passage which is accessible from an
artificial passage showed reverse polarity and are at
least 0.78—0.99 myr old belonging to the Matuyama
Reversed Epoch.
Most of the passages in the Planina Cave System are
active water channels with some higher elevation dry
passages. Paleomagnetic analyses of sediment in smal¬
ler passages showed that the cave sediments were
deposited during the Brunhes Normal Epoch (younger
than 0.78 Ma) (Sebela and Sasowsky, 1999).
SPECIAL ATTRIBUTES
Biology
In 1831, one of the first guides in the Postojna Cave
System, Luka Cec, found the first beetle in the cave.
Hochenwart gave it to Schmidt who described it as a
new species Leptodirus hochenwarti. Because the first spe¬
cimens of cave animals were described from Postojna
Cave, the cave is called the birthplace of biospeleology.
The first specimen of the cave amphibian Proteus angui-
nus was found in Crna Jama Cave, which is part of the
Postojna Cave System, by Josip Jersinovic Lowengreif in
1797. The only known cave vertebrate in Europe Proteus
anguinus was otherwise first described in 1768 by
Laurenti from another cave in southeast Slovenia. The
Postojna Cave System is the type locality for 37 species.
Mineralogy
The Postojna—Planina Cave System is richly deco¬
rated with flowstone (Fig. 5), which is the most com¬
mon form of calcite in Slovene caves. Allochthonous
mechanical minerals, composed of cave sands and
loams, were transported to the cave from elsewhere.
These minerals reflect the composition of the original
noncarbonate rock, mostly Eocene flysch.
Paleontology/Archaeology
The first traces of human settlement (remains of the
cave bear Ursus spelaeus, cave lion, wolf, deer, etc.) in
the valley of the Pivka River derive from the Postojna
Cave System. Cultural layers from Postojna Cave
belong to the last half of the Middle Paleolithic, which
is the end of the Mousterian. In the beginning of the
Wiirm glacial period, the presence of Paleolithic man
in the cave can be proved reliably.
ENCYCLOPEDIA OF CAVES
624
See Also the Following Article
(Brodar, 1969).
Bibliography
Brodar, S. (1969). Latest Palaeolithic Discoveries in the Postojna
Caves. Acta Arcliaeologica, 20, 141—145.
Car, J., & Gospodaric, R. (1984). About geology of karst between
Postojna, Planina and Cerknica. Acta Carsologica, 12, 93—106.
Gabrovsek, F., Kogovsek, J., Kovacic, G., Petrie, M., Ravbar, N., &
Turk, J. (2010). Recent results of tracer tests in the catchment of
the Unica River (SW Slovenia). Acta Carsologica, 39(1), 27—37.
Gospodaric, R. (1976). The Quaternary caves development between
the Pivka basin and polje of Planina. Acta Carsologica, 7, 7—13.
Kogovsek, J., & Sebela, S. (2004). Water tracing through the vadose
zone above Postojnska Jama, Slovenia. Environmental Geology, 45,
992-1001.
Shaw, T. R. (2000). Foreign travellers in the Slovene karst 1537—1900.
Ljubljana: ZRC Publishing.
Shaw, T. R. (2010). Aspects of the history of Slovene karst 1545—2008.
Ljubljana: ZRC Publishing.
Sebela, S. (1998). Tectonic structure of Postojnska jama cave system (Vol.
18, 112 pp.). Ljubljana: ZRC Publishing.
Sebela, S., & Sasowsky, I. D. (1999). Age and magnetism of cave sedi¬
ments from Postojnska jama cave system and Planinska jama
cave, Slovenia. Acta Carsologica, 28, 293—305.
Sebela, S., Vaupotic, J., Kostak, B., & Stemberk, J. (2010). Recent mea¬
surement of present-day tectonic movement and associated radon
flux in Postojna cave, Slovenia. Journal of Cave and Karst Studies,
72, 21-34.
Zupan Hajna, N., Mihevc, A., Pruner, P., & Bosak, P. (2008).
Palaeomagnetism and magnetostratigraphy of karst sediments in
Slovenia. In Carsologica (Vol. 8, 266 pp.). Ljubljana: ZRC
Publishing.
PROTECTING CAVES AND CAVE LIFE
William R. Elliott
Missouri Department of Conservation
PROTECTING CAVE LIFE
Threats to Cave Life
Caves are often thought of as unchanging environ¬
ments, but with modern instruments in the farthest
reaches of large caves there are detectable, annual
changes in air and water. Some caves are naturally
perturbed by flooding or temperature shifts, and these
events influence the types of communities found there.
Caves that flood violently usually lack truly cave-
adapted species, called troglobites or troglobionts, but
may have troglophiles (cave-loving species) and troglox-
enes (animals, such as bats, that roost in caves but exit
to feed or migrate). Indiana bats ( Myotis sodalis ) and
gray bats, ( Myotis grisescens) hibernate in the near¬
freezing zone of certain cold-air-trap caves, which
have deep or large entrances. Such entrances allow
cold air to flow inward during strong cold fronts. In
contrast, in summer gray bats and Mexican free-tailed
bats ( Tadarida brasiliensis mexicana) prefer warm caves
with high ceiling domes, which, combined with bat
body heat, help to create an incubator effect for young
bats in the summer, >30°C in grays and >40°C for
free-tails.
Different animals are adapted to the natural
extremes of caves and other habitats. Human-caused
threats, however, can severely tax the ability of wild¬
life to adjust. Some of the most destructive changes to
caves were brought about by quarrying and water
projects that completely destroyed, or flooded caves.
Human disturbance of bat roosts caused severe
declines in numerous bat species (Elliott, 2000), but
since 2006 white-nose syndrome (WNS) has killed
more than 1 million bats in the eastern United States.
Bats are discussed in more detail below.
Amphibian chytrid fungus, Bntmchochytrium dendro-
batidis (Bd), was found in southern Missouri caves by
Rimer and Briggler (2010). This disease has killed
many amphibians worldwide, but this was the first
report from caves. Eight of 12 caves sampled were pos¬
itive in all five counties visited. Five of the 7 species of
salamanders and frogs sampled were positive, with an
individual infection rate of 31%. Chytrid was con¬
firmed in the grotto salamander ( Eurycea spelaea), the
first report in a troglobite. It is not known if humans
spread the chytrid spores into caves, but once in an
area it can spread via groundwater, humans, or pick¬
erel frogs, which use caves seasonally. Chytrid and
WNS in caves point to the need to practice "clean cav¬
ing" and decontamination as a precaution against the
spread of wildlife diseases.
Other pressures on cave life act over long time
spans and are more difficult to measure. They include
hydrological changes caused by land development,
which can alter the normal hydrological cycle and
increase sedimentation. For example, residential devel¬
opments may cause an increase in nitrogenous wastes
and sediments washed into caves. Sediments can be
harmful to aquatic creatures with gills and/or soft
body tissues, such as cavefishes, salamanders, and
"cave snails" (eyeless, subterranean snails).
The enrichment of caves with wastes, such as
ammonia and other nitrogen-based compounds, in
infiltrating waters, can bring an increase in bacterial
growth. Bacteria in groundwater are not killed by sun¬
light and they can be transmitted for many miles.
Residential developments may also bring exotic spe¬
cies, such as the aggressive red imported fire ant,
which has caused many problems in caves in Texas
(Elliott, 2000).
Encyclopedia of Caves.
© 2012 Elsevier Inc. All rights reserved.
PROTECTING CAVES AND CAVE LIFE
625
Regional overpumping of wells can lower karst
groundwater to the point where important springs and
wells run dry, endangering species that live there and
threatening water sources. In Texas, endangered spe¬
cies of salamander, amphipod, and wild rice have
been affected by such trends.
Dramatic chemical spills can harm caves if the con¬
taminants seep into streams or other routes into the
cave. Once chemicals are in the groundwater, they are
difficult to remove. Nutrient loss seems to happen less
frequently than nutrient enrichment, but can cause
severe problems (see example of Shelta Cave, below).
Extinct and Endangered Species
Although about six North American troglobitic spe¬
cies are thought to be extinct, it is likely that others
became extinct before they could be discovered or
described. Local populations of invertebrates, fishes,
salamanders, and bats have been extirpated. Since
some troglobitic species are endemic to a single cave
or a small cluster of caves, and many caves have been
disturbed, filled, quarried, mined, or polluted, it is
possible that some species have disappeared recently
without our knowledge.
Bats
Bats are important contributors to the world's eco¬
logical health. Caves harbor numerous bat species,
which consume night-flying insects, some of which are
agricultural, forest, or public health pests (Cleveland
et al., 2006). In Missouri, about 800,000 gray bats,
which roost in caves year-round, consume 490 metric
tons of insects per year (Missouri Department of
Conservation, 2010). In the tropics, bats that eat fruit
and pollinate plants often roost in caves. So, even
though some bats do not use caves, the bat-cave con¬
nection is important.
Declines in North American cave bats became
noticeable in the 1950s. Six of the 47 continental U.S.
bats are currently on the U.S. endangered species list
and additional species have been petitioned for listing.
The 6 are dependent on caves for part of their life
cycle, and human disturbance has been the major fac¬
tor in their decline. For example, Indiana bats (M. soda-
lis) lost significant numbers through disturbance of
their hibernacula and improper gating decades ago. If
such bats cannot hibernate deeply, they use up their
body fat too fast, which results in starvation or death
before spring.
Large water projects can drown caves under reser¬
voirs, or use them as recharge wells. A recharge proj¬
ect caused violent flooding of Valdina Farms Sinkhole,
a large cave near San Antonio, Texas. In 1987, a large
flood pulse cleaned out the cave. The cave lost a col¬
ony of four million Mexican free-tailed bats ( Tadarida
brasiliensis mexicana) and a rare colony of ghost-faced
bats (Moormoops megalophylla). A salamander ( Eurycea
troglodytes ) which lived only in that cave, probably is
extinct as a result.
The mining of caves for saltpeter, bat guano, or
other minerals can have a drastic effect on bat colonies
and other fauna. Mexican free-tailed bats have been
disturbed by some guano mining in Texas, while some
miners may have aided the colonies by mining out
rooms that would have filled with guano. The better
operations mine only in the winter when the bats are
gone. The opening of large, second entrances can
severely alter the meteorology of a cave, causing bats
to vacate. Marshall Bat Cave, Texas, lost its free-tail
colony after 1945, when a large, 40-m-deep shaft was
dug into the rear of the cave to hoist out guano, caus¬
ing too much ventilation and cooling of the cave.
Mammoth Cave once harbored Indiana bats before
the entrance was modified to block incursions of cold
winter air. The National Park Service is currently try¬
ing to reinstate the natural temperature profile of the
cave.
Nutrient loss from the loss of gray bats apparently
caused a domino effect in Shelta Cave, in Huntsville,
Alabama. Shelta had the most diverse cave community
known in the southeastern United States, but land
development encroached on the cave in the 1960s and
the townspeople were concerned about youths enter¬
ing the cave. The cave harbored a large colony of gray
bats. The National Speleological Society purchased the
cave in 1967, and they moved their headquarters to a
building nearby. They gated the cave in 1968 with a
strong, cross-barred gate that had been taken from an
old jail. This gate, in hindsight, was inappropriate for
bats, which abandoned the cave within two years.
Urbanization of the area probably affected the colony
too. In 1981 a horizontal-bar, bat-friendly door was put
on the gate, but no bats returned to the cave. The loss
of bat guano to the lake in the cave probably contrib¬
uted to the decline of cave crayfishes there (Elliott,
2000). The gate was replaced with a high fence around
the area in the early 2000s, and some bats returned.
WNS is the most severe threat to cave- and mine-
hibernating bats to date, killing more than 1 million
bats of nine species from 2006 to 2010 in the eastern
United States and Canada. As the fungal infection,
probably originating from Europe, spread from the
northeastern United States into Canada and south and
west as far as Oklahoma, many agencies and land-
owners prepared for the worst and wrote response
plans. Thousands of caves were closed or restricted to
human access to avoid possible spread of the
ENCYCLOPEDIA OF CAVES
626
p
infectious, microscopic spores of Geomyces destructans,
the apparent causative agent. Bats spread the spores,
but it is not clear how much was spread on unwashed
caving gear and clothing. Strict decontamination rules
are required for anyone to enter many caves now, and
an effective treatment is not yet in sight. It is clear that
more and better cave gates and alarm systems will
be needed for important bat caves, but also more
research, bat monitoring, and disease surveillance. It is
impossible and undesirable to gate all caves containing
bats, which might number 90,000 or perhaps 100,000
caves in the United States. Even noncolonial bats are
affected, such as Perimyotis subflavans, the tri-colored
or eastern pipistrelle bat, and sites remain infectious
for long periods after bats have died off. Captive hold¬
ing of bats for study, treatment, or even propagation,
may be the last resort for some species that are rapidly
declining, such as Myotis lucifugus, the little brown bat.
It is likely that WNS will spread to the western United
States and even into the highland tropics, wherever
bats hibernate in chilly caves and mines.
Many cave bats produce small amounts of guano dur¬
ing migration and hibernation. If WNS causes a decline
in gray bat maternity caves in summer, there could be a
decline in nutrient input to some cave systems in the
southeastern United States. Some conservationists have
proposed replacing lost gray bat guano in caves, but
there are problems in that strategy, as outlined below:
1. Many caves with high biodiversity lack gray bats
(Elliott, 2007), so loss of gray bat guano is not a
general threat to all cave biodiversity.
2. Large-scale guano transfer is an untested idea. A
search of 25,000 references in the American cave
biology literature found 32 reports on guano, but
none was an experimental transfer of bat guano
from one cave to another.
3 . Large-scale guano transfer from one cave to another
might introduce infectious Geomyces destructans
spores, whereas many are trying to practice
contamination control within caves, as well as
decontamination before and after entering caves.
4. Large-scale guano transfer from a gray bat cave
would disrupt the source cave. Other types, such as
free-tail guano, contain different fauna and flora,
and it might be biochemically harmful to the
recipient community of guanophiles.
5 . The total energy contribution of gray bat guano to a
cave ecosystem has not been measured well. Many
nutrients come from rotting detritus, organics in
infiltrating groundwater, bacteria, fungi, dead
pickerel frogs and salamanders, cave crickets, and
so on. Most of the carbon flux in two Missouri cave
streams that were studied is from allochthonous
sources (Lerch et al., 2000, 2002).
6. Wildlife agencies and researchers do not have the
manpower for large-scale guano transfer in a
scientifically controlled fashion.
Further discussion of bats may be found below, in
the subsections "Cave Gating Criteria" and "Cave
Restoration."
CAVE PRESERVE DESIGN
Cave conservation encompasses many aspects,
which can be applied to developed show caves as well
as to "wild" (unmodified) caves. Show caves, however,
have special problems, such as growth of cyanobacter¬
ia and plants near trail lights, accumulation of "cave
lint" and trash, and general disruption of the cave's
ecology. Some show caves avoid the use of trail lights
but still have good trails; the visitors are provided
with electric hand lanterns, and viewing native wild¬
life along the trail is a goal of the tour. Such show
caves usually provide a more educational experience
for the public.
Good cave management includes rules of access to
the cave. Many publicly owned caves can be left open
to the public as long as they do not vandalize or dis¬
turb cave life.
Besides WNS considerations, some caves require a
permit system for entry, based on flooding hazards or
other considerations of safety or sensitivity. Usually a
permit or restricted access cave would be gated to con¬
trol access, but appropriate signs are needed to inform
people of the availability of permits (Fig. 1). Certain
caves may be considered closed for recreation, but not
for monitoring and research. Examples include a few
caves that are especially pristine and rich in multiple
resources, or which have overlapping seasons for
endangered bats. Heretofore, a few eastern American
caves harboring gray bats in the summer and Indiana
bats in the winter were accessible between bat seasons
in May and September, but now few are managed that
way.
Cave preserves have been set aside for the protec¬
tion of endangered, nonbat species. Too often, how¬
ever, such preserves surround only the entrance area
and do not include the entire extent of the cave, much
less the recharge area to the cave (often called the
watershed in American usage). It is essential to have
good scientific information about the cave: an accurate
map, a description, inventories of the cave's resources,
and a hydrogeologic assessment. The assessment may
require a dye-tracing study in which tracer dyes are
put into streams, sinkholes, and other input points.
The dyes are recaptured with charcoal traps placed in
cave streams, springs, or wells. Maps can then be
ENCYCLOPEDIA OF CAVES
PROTECTING CAVES AND CAVE LIFE
627
NOTICE
CAVES ARE CLOSED
TO ENTRY UNLESS SIGNED OPEN
OR YOU HAVE A VALID PERMIT
You may not enter a Missouri Department
of Conservation cave that lacks a sign giving access.
Written permits may be available for selected caves.
For information see http://www.mdc.mo.gov or contact:
Proper disinfection is required before entering caves to reduce the spread of
White Nose Syndrome (WNS), a wildlife disease that has killed millions of bats
(search www for “FWS WNS"). Bats die more often when disturbed.
It is against the law to enter closed caves or caves with no signs, bypass or
tamper with a cave gate or fence, dig, remove natural, prehistoric, or historic
objects, build a fire, leave trash, take vehicles or animals into, deface or
vandalize caves. Persons violating these requirements will be prosecuted.
The Missouri Wildlife Code protects native wildlife. Disturbance of certain
wildlife at caves may be punishable by fines up to $50,000 for each violation
FIGURE 1 A 2010 version of a cave sign with information on
white-nose syndrome.
drawn that delineate the cave's water sources, which
makes it possible to more scientifically manage the
landscape.
Lack of detailed information should not stall conser¬
vation planning, however. For example, foresters
in Missouri try to maintain water quality to ensure a
pesticide-free food supply for gray bats. They maintain
a continuous forest canopy 60 m wide along streams,
in the 8 ha around and above gray bat cave entrances
and as travel corridors 60 m wide from gray bat caves
to riparian foraging areas. This canopy provides pro¬
tection from predators and a substrate for insect
production.
Buffer zones around small caves in Texas lacking
streams have been as small as one hectare just for the
protection of terrestrial invertebrates. It is important
to maintain native vegetation and drainage patterns
on the epigeum (surface). Even intermittent cave
streams may have sources beyond a few hectares, so
the preserve would necessarily be larger in such cases.
Krejca et al. (2002) recommended a minimum preserve
size of 28—40 ha (69 to 99 acres) around a small cave
or cave cluster, as well as maintenance and adaptive
management against other threats, such as red.
imported fire ants (Solenopsis invicta). If the cave pre¬
serve is adjacent to undeveloped lands, then occa¬
sional visits to the cave by raccoons and other
nonresident species may continue to provide neces¬
sary nutrient inputs in the form of droppings. If the
preserve is isolated by developed lands, then it proba¬
bly should be larger to maintain native flora and
fauna. Camel crickets and harvestmen may exit the
cave at dusk and forage for carrion and feces in the
surrounding area, but these arthropods may not travel
more than 100 m from the cave entrance. Pesticide use
is banned or limited in cave preserves to avoid poi¬
soning the cave fauna directly or indirectly. Even a
small cave may have to be protected with a strong
cave gate to prevent heavy visitation and vandalism,
which can alter the cave habitat. Such cave gates
should be designed to freely allow bats and small ani¬
mals to pass back and forth. Preserve designs are dis¬
cussed in various species recovery plans and in
guidelines issued by the U.S. Fish and Wildlife Service
and other agencies.
Buffer zones around large caves may be more diffi¬
cult to achieve. For example. Tumbling Creek Cave in
Missouri has high biodiversity and a recharge area of
2331 ha. The cave was under two private properties for
many years with extensive forest cover, losing streams,
light farming, and cattle production. Even though
the cave was managed by the Ozark Underground
Laboratory, gray bats gradually declined for several
reasons. In the 1990s sedimentation from poor land
usage by a neighbor contributed to the decline of a
cave snail unique to the cave. A drought of several
years probably contributed to the decline too. Today,
more of the cave's recharge area is under careful land
use, but it remains to be seen if the cave snail will
increase again (USFWS, 2001; Elliott and Aley, 2006;
Elliott et al., 2008).
In southeast Alaska's Tongass National Forest, stud¬
ies found that some old-growth forest areas had to be
protected on intensely karsted terrain. The limestone
was so pure that little mineral soil had developed, and
trees grew out of a thin moss blanket. When clear-cut,
the thin soils washed off into the numerous sinkholes,
which fed cave streams, which fed salmon streams. In
a sense, the karst served as three-dimensional stream
banks feeding the local streams, and could be consid¬
ered as stream buffers removed at a distance, and pro¬
tected under existing laws and forestry standards.
Some cave entrances and sinkholes received slash and
runoff from logging and roads, in violation of the
Federal Cave Resources Protection Act. Many of these
areas are now protected from road building and tim¬
ber harvest (Elliott, 1994).
Some states publish "Karst BMP" sheets, providing
the "best management practices" for construction and
ENCYCLOPEDIA OF CAVES
628
p
other ground-disturbing projects on karst. These BMPs
should be applied to prescribed burns and other con¬
servation projects. Different types of karsts may
require different BMPs. The essentials of a karst BMP
are summarized below:
1. A karst BMP should generally describe karst
geology, hydrology, and biology, and emphasize
that karst resources are easily damaged by
sediments, spills, dumping, and construction
activities.
2. Identification of sinkholes, recharge areas, sinking
streams, gaining streams, caves, and springs is
needed to avoid problems. Consult geological,
speleological, and state natural heritage surveys.
Cave maps are useful for determining the extent
of a cave. Dye-tracing studies can delineate the
recharge area for a cave, sinkhole, or spring. Some
karst features are hidden under certain soils.
3. Staging and storage areas with fuel and oil should
be located away from karst features, and must have
erosion controls. Concrete and wash water is
disposed off-site. Temporary roads should not be
steep, and should have runoff controls.
4. Buffer zones of at least 30 m (100 ft) should be
maintained around caves, sinkholes, and springs.
Sediment controls are installed upslope of buffer
zones. Pesticides and prescribed burning are
generally banned within buffer zones.
5. Disturbed areas should be revegetated or allowed to
regrow after a project. Annual nonnative grasses
such as rye or wheat may be planted in conjunction
with native species to provide short-term erosion
control. Areas subject to erosion may be planted
with nonnative mixtures for rapid establishment
and erosion control. Follow-up inspections are
required and continued controls may be necessary.
6. Avoid altering drainage patterns, and properly
dispose of debris and trash away from karst
features. Temporary erosion controls should be
removed after serving their purpose, but some
permanent controls may be needed.
Cave Gating Criteria
Cave gates are steel structures built to protect cave
resources by keeping out human intruders while
allowing air, water, and wildlife to pass freely in and
out. Cave gates have locking doors or removable bars
so that authorized persons can gain access during
appropriate seasons for necessary work. Poor cave
gates can harm wildlife and cave resources, but good
ones can improve conditions for bats and other wild¬
life. Cave gating is not an automatic solution to cave
conservation problems and there are many reasons for
not gating a cave. Technical knowledge and experience
are needed to gate a cave; for example, it cannot be
done properly by a general welding contractor without
providing specifications, a design, and onsite supervi¬
sion by an experienced cave gater. Knowledge of cave
ecology, especially bats, is necessary before a gate is
considered. Similar techniques are used for gating
abandoned mines, which often harbor bats of great
value. Some governmental agencies and cave conserva¬
tion groups assist cave owners in cave gating, but first
a decision guide must be followed (below).
A few rules of thumb can be followed for cave
gates. Natural entrances should not be sealed, but
opening a long-sealed cave can also cause problems
for the cave unless some means of protection is
devised. Gates should not be made of reinforcing bar,
or rebar ; it is much too weak. Chain link fences are
easily breached, but can be used around sinkholes if
necessary; stronger "vertical bar fences" are now pre¬
ferred. Do not construct any raised footings, stone
work, or concrete walls on the floor or around a gate
because they can hinder air exchange and change the
temperature at the bats' favorite roosts. Gates should
be tailored for the wildlife and other resources in the
cave. A cave gate is not a substitute for good land
management, but it is a last resort.
Limited space does not allow a full discussion here
of the many construction techniques that have been
developed for cave gating. Cave gate designs and
specifications by the American Cave Conservation
Association (ACCA), Bat Conservation International
(BCI), Missouri Department of Conservation, and
National Speleological Society (NSS) are available on
the World Wide Web and in publications.
Cave Gate Styles
Depending on the needs of the cave, the type of
entrance, bats and other wildlife, the design could spec¬
ify a full gate (Fig. 2), half gate or fly-over gate (Fig. 3),
window or chute gate (particularly for maternal gray
bats), cupola or cage gate (Fig. 4), bay window gate,
enclosure, fence, or no gate at all. For example, some
bat caves that may need a gate for protection are not
feasible to gate for certain physical reasons. Many caves
that are feasible to gate ought not to be gated because
other modes of protection may work better.
It is important to rely on an experienced cave-gating
expert. Leading organizations are ACCA, BCI, U.S.
Forest Service, U.S. Fish and Wildlife Service, the
Missouri Department of Conservation, NSS, and
others. ACCA's designs were adopted by BCI and
many government agencies, and have become the
industry standard.
ENCYCLOPEDIA OF CAVES
PROTECTING CAVES AND CAVE LIFE
629
FIGURE 2 The National Cave Gating Workshop, 2009, built a bat-friendly, full cave gate at Cliff Cave, Missouri, to protect multiple
resources, including endangered Indiana bats.
FIGURE 3 A large half gate or fly-over at Great Spirit Cave, Missouri, completed in 2002. The gate weighs 18 metric tons and measures
31 m wide by 5 m high. It protects gray bats (Myotis grisescens) in summer, Indiana bats (Myotis sodalis) in winter, and multiple resources in
this Natural Area cave.
ENCYCLOPEDIA OF CAVES
630
FIGURE 4 A cupola or cage gate over a sinkhole
entrance, which allows bats to gain altitude and exit
laterally, easier than flying up through a horizontal
grate.
The above organizations teach annual, regional
cave-gating workshops to demonstrate the proper
decision-making process, design, and construction
techniques for ecologically sound cave gates (Fig. 2).
These gates have resulted in significant protection and
increases of colonies of endangered bats, such as grays,
Indianas, and others. Protection of other irreplaceable
cave resources is another benefit of properly built
gates. Such workshops may also lead to the formation
of regional cave-gating groups from different organiza¬
tions. The workshop comprises lectures supplemented
by building a real cave gate under the direct supervi¬
sion of cave-gating experts. Each gate is somewhat dif¬
ferent, and various problems in funding, logistics,
design, safety, teamwork, and construction must be
solved.
Gates are now made stronger than in the past, but it
is important to check and repair them because any
cave gate can be breached by determined vandals.
Designs have evolved to foil attempts to tunnel under
a gate or to destroy the door or lock. Specifications
vary for different gate styles. Gates are usually made
of mild or modified steel "angle iron," although stain¬
less and manganal steels may be used in corrosive
environments. The latter are more expensive and are
unnecessary in most applications. The service life of
most steel gates may be 30 years.
Cave Gating Decision Guide
Are There Poor Reasons Not to Gate the Cave?
For example,
• Purely aesthetic objections to a gate while the cave's
resources are being degraded anyway.
• It may "start a trend" toward too much gating.
• Because a few people consider themselves above the
rules and may threaten the gate.
Score no points for any poor reasons not to gate.
Are There Poor Reasons for Gating the Cave?
For example,
• For fear of liability, which probably is nonexistent.
Cave owners are protected by law in some states.
• For administrative convenience (instead of having a
comprehensive conservation program).
• To keep wild animals or competing explorers out.
Score no points for any poor reasons to gate.
Are There Good Reasons Not to Gate the Cave?
For example,
• The gate, as designed, will not comply with current
ACCA and BCI standards.
• A vigilant owner or manager lives nearby.
• Other controls can be used—road gates, signs,
surveillance.
• Visitors probably will comply with a good permit
system.
• Cave management experts are opposed to the gate.
• The cave gaters are inexperienced and overconfident.
• No one will commit to checking and maintaining
the gate.
• Technical reasons: The entrance is too small for a
proper gate ( e.g ., half gate for gray bat maternity
colony), or the environment or budget will not
allow a good design.
Score one point each against gating if any good rea¬
sons against gating hold true.
Are There Good Reasons to Gate the Cave?
• The cave is hazardous to casual visitors and no
other controls (permits and signs) are adequate.
• Endangered species inhabit the cave and can be
bolstered by protection.
ENCYCLOPEDIA OF CAVES
PROTECTING CAVES AND CAVE LIFE
631
• The cave is a target for vandals, looters, and
trespassers. A "better clientele" is needed.
• The cave has high value, is threatened, and it can
best be studied and appreciated with a good permit
system combined with a gate.
Score one point each for gating if any good reasons
hold true.
Final Results
Add up the points for and against gating, and deter¬
mine which seems more important. Other criteria may
have to be considered.
Security Systems
Electronic technology now allows cave conserva¬
tionists to deploy alarms and surveillance equipment
in lieu of cave gates, or to supplement gates that
are frequently attacked. Light, vibration and magnetic
sensors are becoming available that can distinguish
human intruders from bats and other small animals.
An option for a cave manager would be an audible
alarm versus a silent alarm. The former may frighten
off intruders or anger them, possibly leading to van¬
dalism of the equipment. A silent alarm would alert an
authority, who could apprehend or warn the intruders.
The key question to consider in designing a silent cave
alarm system is, "Who will respond, and how
quickly?" If law officers or managers cannot respond
within one hour to apprehend the intruders, then the
alarm system probably will not be a deterrent.
There are many options for these security systems, a
few of which are outlined below.
Sensors for detecting humans could include attractive
objects, pressure mats, light detectors, infrared light
beams, motion detectors, vibration detectors, magnetic
detectors, and infrared trail cameras with invisible flash.
The cheapest intrusion detector is an attractive,
inexpensive object, such as a coin placed in a known
place in the cave. If it disappears since the previous
visit, then the manager knows that an unauthorized
person or a packrat has been there.
Pressure mats were designed for indoor use. Such
mats are switches using no electrical power until they
are triggered by someone stepping on them. They are
not reliable sensors unless they are placed on a flat,
hard surface, and then covered with soil. They can be
linked together to create a security zone. It may help
to place something just beyond the mats that will
attract an intruder and lead him to step on the mat
unknowingly. A simpler method might be a sign with
a momentary switch to the system labeled "Do not
touch this."
Light detectors are available in custom alarms or as
data loggers, which record the time and date that a
light illuminated the detector in the dark zone of the
cave. Such data loggers may be used to measure the
amount of traffic in well-traveled caves, or to detect if
anyone has trespassed into a closed cave, but the data
are downloaded later and cannot determine anyone's
identity.
Custom light sensor alarms can alert a manager or
law officer via radio, telephone, or satellite communi¬
cation to the Internet. However, it is usually necessary
to install such sensors in the dark zone, leaving the
entrance and twilight zone unprotected. They require
continuous electrical power, as do motion detectors,
which can be accomplished with solar panels and bat¬
teries or other power sources and cables, which must
be concealed.
Invisible infrared light beams can be positioned at a
certain height so that bats and wildlife are unlikely to
trip them, and they can be deployed in the entrance,
but false alarms may occur. Arrays of infrared beams
or "beam breakers" are now used to monitor elevated
bat activity during WNS-related activity.
Vibration detectors can be concealed and tuned to
respond only to the seismic shaking caused by a
human walking, or by earthquakes. They are more
effective when attached to a cave gate, but like all cave
electronics, they must be hardened against the harsh,
damp, mineral-rich cave atmosphere, and flooding.
All of the above sensors must be linked to a system
that conveys information out of the cave, either via a
cable that must be concealed, or by wireless relays.
The system can then trigger an automated telephone
call, radio call, or satellite signal to an authority.
Concealed video and still cameras are routinely
used by law officers to obtain evidence leading to the
arrest and conviction of law breakers. The image must
be clear enough to identify the suspect. They are some¬
what labor-intensive and vulnerable to attack if the
culprit notices the equipment, but they are easily
deployed to new sites as needed.
MANAGEMENT AND EDUCATION
Cave Management
A general goal of land management over a cave is
to not alter the landscape much or build infrastructure,
such as sewer lines, pipelines, roads, and the like,
especially if the cave contains streams and species of
concern. Paving over the top of a cave cuts off much of
its water supply. Installing septic systems may relegate
the cave to a sewer.
ENCYCLOPEDIA OF CAVES
632
p
In prescribed burns one should insure that smoke
does not enter caves, especially those occupied by
endangered bats. For example, if one burns a hillside
under certain meteorological conditions, a cold front
may carry the smoke down the slope and into a cave,
especially if it is a cold-air trap. This could be harmful
for hibernating bats. Conversely, smoke could rise up
a hill and into a cave that serves as a summer roost,
because it may be a warm-air trap.
Cave Restoration
Many cavers and cave managers favor cave cleanup
and restoration projects for multiple reasons. Graffiti
removal may seem to be only an aesthetic pursuit, and it
usually does not directly help restore wildlife in the cave
unless toxic materials are removed. However, one has to
consider how some caves become targets for vandalism.
If a government agency or cave owner allows people to
vandalize and litter a cave, then many visitors assume
that it is all right to do so, and they will continue.
Inaction is considered condoning the bad behavior.
Vandalism sometimes extends to harassment or killing
of bats and other wildlife. Trashy caves become training
grounds for ignorant behavior, places to act up, and this
behavior spreads to other caves. The same behavior is
seen in illegal dumpgrounds, where a few leave some
household trash on vacant land. If no one objects, others
then opportunistically dump there and the problem
accelerates. Signage does not seem to help at that point.
We can involve volunteers in photographing, docu¬
menting, then carefully removing graffiti. Such a conser¬
vation project helps to restore public respect for a cave,
especially if we publicize the work and explain the effort.
An important educational need is filled in this way. We
can also take the opportunity to educate volunteers
about wildlife during the project, which they appreciate.
Most graffiti that this author has seen is not historically
significant, but important signatures and markings are
usually left in place or else documented in a photographic
file or database before removal. Certainly aboriginal and
other valuable markings and art must be protected.
Various rules and laws define the types and ages of mark¬
ings that must be preserved in place, so it is best to check
with the appropriate state historical program office about
such requirements. More stringent requirements may be
practiced by conservationists on their own initiative.
An example of restoration is Little Scott Cave in
Missouri, managed by the Missouri Department of
Conservation (MDC). This cave is near a highway, and
it became a target for repeated vandalism, including
extensive and gross graffiti, breakage, spray painting
two bats to death, hundreds of beer cans from under¬
ground parties, and candles from occult ceremonies.
Cavers repeatedly tried to clean up the mess and noti¬
fied MDC. Finally, MDC gated the cave in April 2000 to
stop the cycle. MDC simultaneously instituted an easy
procedure for getting access to the cave by telephone
interview. MDC awarded grants to two grottoes (caving
clubs) to clean up the cave using low impact techniques
in which no chemicals were used. This work was noted
in local newspapers and in Elliott and Beard (2000).
Before restoring a cave to a more natural state, one
should consider how altered the cave is and define
realistic goals for the restoration project. Is it a show
cave with many years of accumulated change and little
hope of complete restoration, or is it a wild cave that is
not so ecologically disturbed? In a show cave nuisance
species may be present: cyanobacteria near electric
lights, exotic species in cave lint along trails, or epi-
gean (surface) wildlife that is attracted to artificial food
sources in the cave. The following questions should be
asked before planning projects:
1. Do we know what native species should be in the
cave?
2 . How can we restore the cave to a more normal
aesthetic and ecological state without harming the
native species?
3. To what historic or prehistoric period should we
restore the cave?
The historic period is an important question in eco¬
logical restoration for bats, many of which species have
declined drastically within the past few decades. Some
species of bat can be protected by bat-friendly gates.
Some roosts may return to the "maximum past popula¬
tion" as measured by ceiling stains. Most colonies do
not fully recover. Very large bat colonies (e.g., free-tails)
may not tolerate a full gate at all because of the bottle
neck, traffic jam, and acoustical confusion it may cause.
It is important not to remove decayed wood and
organic matter from a cave without first checking it for
cave life, which may have colonized it over many
years (Elliott, 2006).
No cave is ever completely restored to its former
aesthetic or ecological state. Proper biological inven¬
tory and project planning will improve the success of
restoration efforts.
Prioritizing Caves
Cave protection usually involves crisis management:
caves and cave life that are under the greatest threat
receive the most attention. Rarely do we look at caves
over a large region and consider which ones deserve
protection before they become degraded.
Caves can be very different, and comparing and rank¬
ing them seems difficult. In some states adequate records
ENCYCLOPEDIA OF CAVES
PROTECTING CAVES AND CAVE LIFE
633
exist to allow cavers, managers, and scientists to evaluate
and prioritize caves for protection. Ideally, a cave would
receive a numerical score for each of its resources. A com¬
posite score can then be derived for the whole cave and
used for ranking and prioritizing caves or karst regions. A
cave could be scored for its length, depth, hydrology, bio¬
diversity, value as a bat cave, geology, paleontology,
archaeology, history, speleothems, aesthetics, recreational
value, and the threats against it. There are various ways
to score each aspect; for instance, for biodiversity one can
consider the total number of species present, how rare or
endemic the cave's troglobitic species are, the cave's
importance to endangered species, and its overall biodi¬
versity. Small numerical differences between caves are not
important, and the rankings are used as a general guide.
Most states have a Natural Heritage Database, usu¬
ally within a state agency, and many states have a state
cave database, usually managed by a nonprofit cave
survey organization. Increasingly these databases are
being carefully shared among organizations for the
purpose of protecting significant caves from road and
land development and other threats. The locations of
caves, especially those on private lands, are protected
in each system so as not to draw the attention of
potential intruders. Coordinates can be truncated to
coarse values for mapping over large areas without
revealing cave names or precise locations.
Public Education
Cave conservation includes publications, videos,
and educational programs for the public. MDC and
other organizations offer workshops on cave ecology
for biologists and teachers.
It is not necessary for the public to become experi¬
enced cavers to achieve conservation goals, but it is
helpful to inform the public about all the resources con¬
nected to caves. Two concepts are important to convey
to the public: (1) caves operate on a much longer time
scale than surface landscapes, and are essentially non¬
renewable resources. (2) Our caves have already lost so
much—when we visit a cave we should give something
back to it. For example, we can pick up trash, teach
others about cave conservation, or advocate for the
cave and the bats to those who have authority over it.
Even for those who do not enjoy caves, it is important
to know that bats provide ecosystem services to us in
the form of insect control, and that valuable karst
groundwater resources and caves are highly related.
Bibliography
Cleveland, C. J., Betke, M., Federico, P., Frank, J. D., Hallam, T. G., &
Horn, J., et al. (2006). Economic value of the pest control service
provided by Brazilian free-tailed bats in south-central Texas.
Frontiers in Ecology and the Environment, 4, 238—243.
Elliott, W. R. (2000). Conservation of the North American cave and
karst biota. In H. Wilkens, D. C. Culver & W. F. Humphreys
(Eds.), Subterranean ecosystems (pp. 665—689). Amsterdam: Elsevier.
Elliott, W. R. (1994). Alaska's forested karstlands. American Caves 7,
8 - 12 .
Elliott, W. R. (2006). Biological dos and don'ts of cave restoration
and conservation. In J. Hildreth-Werker & J. Werker (Eds.), Cave
conservation and restoration (pp. 33—44). Huntsville, AL: National
Speleological Society.
Elliott, W. R. (2007). Zoogeography and biodiversity of Missouri
caves and karst. Journal of Cave and Karst Studies, 69,135—162.
Elliott, W. R., & Beard, J. B. (2000). Cave restoration. Missouri
Conservationist, 61, 14—16 (also available at http:Hmdc.mo.gov/con-
mag/2000/10/cave-restoration).
Elliott, W. R., & Aley, T. (2006). Karst conservation in the Ozarks:
forty years at Tumbling Creek Cave. In G.T. Rea (Ed.), Proceedings
of the 2005 Natiotial Cave and Karst Management Symposium, Albany,
NY, Oct. 30-Nov. 4, 2005. (pp. 204-214). Albany, NY:
Northeastern Cave Conservancy.
Elliott, W. R., Echols, K., Ashley, D. C., Aley, T., Leary, A., &
McKenzie, P. (2008). Waterborne contaminants in Tumbling
Creek Cave, Missouri. In W.R. Elliott (Ed.), Proceedings of the 18th
National Cave and Karst Management Symposium, October 8—12,
2007, St. Louis, Missouri (pp. 107—123). Jefferson City, MO:
Missouri Department of Conservation.
Fant, J., Kennedy, J., Powers, R., Jr., & Elliott, W. R. (2009). Agency
guide to cave and mine gates. Austin, TX: American Cave
Conservation Association, Bat Conservation International.
Krejca, J., O'Donnell, L., Kennedy, K., Knapp, S. M., & Shull, A.
(2002). How much surface habitat is enough? Preserve design
and application for cave-limited species. In G. T. Rea (Ed.),
Proceedings of the 2001 National Cave and Karst Management
Symposium, October 16—19, 2001 (p. 141). Tucson, AZ: USDA-
Forest Service, Coronado National Forest.
Lerch, R. N., Erickson, J. M., & Wicks, C. M. (2002). Intensive water
quality monitoring in two karst watersheds of Boone County,
Missouri. In G. T. Rea (Ed.), Proceedings of the 2001 National Cave
and Karst Management Symposium. Tucson, Arizona, October 16—19,
2001 (pp. 158—168). Tucson, AZ: USDA-Forest Service, Coronado
National Forest.
Lerch, R. N., Erickson, J. M., Wicks, C. M., Elliott, W. R., & Schulte, S. W.
(2000). Water quality in two karst basins of Boone County, Missouri
(Abstract). Journal of Cave and Karst Studies, 62, 187.
Missouri Department of Conservation (2010). White-nose syndrome action
plan. Jefferson City, MO: Missouri Department of Conservation.
Rimer, R. L., & Briggler, J. T. (2010). Occurrence of the amphibian
chytrid fungus ( Batrachochytrium dendrobatidis) in Ozark caves,
Missouri, U.S.A. Herpetological Review, 41, 175—177.
U.S. Fish and Wildlife Service. (2001). Endangered and threatened
wildlife and plants, list the Tumbling Creek cave snail as endan¬
gered [Emergency rule] Federal Register, 66, 66803—66811.
See Also the Following Articles
Exploration of Caves—General
Recreational Caving
ENCYCLOPEDIA OF CAVES
QUARTZITE CAVES OF
SOUTH AMERICA
Augusto S. A uler
Instituto do Carste
INTRODUCTION
A significant portion of South America comprises
ancient geological units, especially in the highlands
east of the Andes, where Proterozoic carbonates and
quartzite host several significant caves. Quartzite
rocks, being usually very resistant to weathering, tend
to form the highest and most inaccessible ridges and
plateaus. Poor sand-rich soil does not favor agriculture
and these areas are usually little populated. Due to the
difficulty of access, quartzite ridges have received lim¬
ited attention from speleologists. Although the poten¬
tial for major caves has been recognized for several
decades, only in recent years has systematic explora¬
tion started to reveal outstanding caves.
The total number of quartzite caves is still very
small when compared to the number of carbonate
caves. It is likely that fewer than 1000 quartzite caves
have so far been documented in South America. This
is due to the lack of exploration, access difficulties as
well as the smaller density of caves in quartzite rocks
when compared to carbonate. Only one cave, 16.1-km-
long Cueva Roraima Sur (Ojos de Cristal) in Venezuela,
is known to surpass 10 km in length, but at least
twenty caves are now over 1 km long. Large cave
volumes have been recognized from a few sites, both in
Venezuela and Brazil. Very few caves can be consid¬
ered as possessing significant depth, the area contain¬
ing the largest collection of deep caves being the
Inficionado Peak in southeastern Brazil, where two
caves are known to surpass —400 m in depth.
Many quartzite caves occur in scenic mountain
ridges and plateaus, where the rocky landscape rends
it as of limited commercial value and the lack of settle¬
ments made the establishment of protected areas easy.
As a result, some of the more remarkable quartzite
caves are within the boundaries of state and national
parks.
LITHOLOGY AND OCCURRENCE
Quartzite caves occur in a number of different geo¬
logical formations throughout South America. The large
majority represent ancient (Late to Mid Proterozoic)
fluvial and eolic sandstones metamorphosed at various
degrees of intensity. Terms such as quartzite, meta¬
sandstones, and even sandstones have been applied to
these rocks. In southern Venezuela, close to the border
with Brazil and Guyana, the Gran Sabana area
hosts several quartzite plateaus known by the Indian
denomination of tepuis. These flat-topped plateaus
have vertical walls that may exceed 1000 m in relation
to the surrounding plain. Quartzite belongs to the
>1.5-G yr-old Mataui Formation that lies subhorizon tally
and displays limited tectonic disturbance. Annual rain¬
fall exceeds 3500 mm. Only a few of the tepuis have
been subjected to systematic speleological exploration,
the best known being Roraima, Chimanta, Auyan,
Kukenan, Akopan, Autana, and Sarisarinama. Roraima
tepui hosts the best-known cave, Roraima Sur (also
known as Ojos de Cristal), currently the longest quartz¬
ite cave in the world at 16.1 km in length (Fig. 1).
Recent exploration at Chimanta tepui has revealed very
extensive caves such as Cueva Charles Brewer (7.3 km
long) and Muchimuk-Colibri system (8 km long), both
containing 50-m-wide passages and spacious chambers
representing some of the largest cave volumes in
quartzite recorded so far (Fig. 2). Auyan tepui contains
a vast depression that hosts the vertically significant
Aonda Cave at —383 m in depth and over 2 km in
length. Autana tepui was among the first to be explored
in the 1970s, being home to the Autana Cave, a large
Encyclopedia of Caves.
635
© 2012 Elsevier Inc. All rights reserved.
636
FIGURE 2 Spacious river passage at
Muchimuk Cave. Photo courtesy of Marek Andy.
Used with permission.
are over 300 m deep. Several tepuis await exploration,
there being considerable potential for further major dis¬
coveries. South of the tepui area, the little explored
quartzite plateau of Aracas Ridge in Brazil has recently
yielded at least one major vertical cave.
Proterozoic quartzite comprises much of the Espinhafo
ridge, running north—south along eastern Brazil. The
well-lithified rocks of the Espinha^o Supergroup com¬
prise numerous cave-bearing quartzite units at various
degrees of tectonism and metamorphism. The 1.1-Gyr
Chapada Diamantina Group in south-central Brazil is
characterized by both conglomerate and quartzite, giving
rise to tabletop mountain scenery similar to that in the
tepui area, although much smaller in scale. Torras Cave,
3.8 km long, is one among a series of major caves within
and around Chapada Diamantina National Park. The
very large Lapao Cave is entirely within a conglomerate
containing quartzite pebbles and a silica-rich matrix. The
Espinhaco ridge is over 1000 km long and many other
relevant caves occur throughout its length.
In the tectonically complex Iron Quadrangle area in
south-central Brazil, metamorphosed quartzite occurs
together with several other rock types. Quartzites of
both the Moeda Formation (2.4 Gyr) and Itacolomi
Group (~1.8 Gyr) contain major caves. The most signif¬
icant area is the 0.9-km 2 top of the Inficionado Peak,
where several fissures open into major vertical caves
(Fig. 3). Three caves are over 350 m in depth, Gruta do
Centenario being 3.7 km long and —484 m deep.
The Andrelandia Group (~1 Gyr) quartzites at the
Ibitipoca ridge contain a large density of caves, includ¬
ing the 2.7-km-long Gruta das Bromelias. In the sur¬
roundings dozens of significant horizontal or gently
sloping caves occur, sometimes associated with cuesta
relief.
tunnel that cuts through the upper third of the tepui
vertical wall. The Sarisarinama plateau contains some
vast circular depressions, some of which (Sima Mayor)
FIGURE 1 River passage at Cueva Roraima Sur (Ojos de Cristal).
Photo courtesy of Marek Audy. Used with permission.
ENCYCLOPEDIA OF CAVES
QUARTZITE CAVES OF SOUTH AMERICA
637
FIGURE 3 Vertical fissures at the top of the
Inficionado Peak. Some of these fissures give
access to very deep and long caves such as
Gruta do Centenario. Photograph taken by the
author.
Important quartzite and sandstone caves occur in all
major biomes, from equatorial Amazonia to temperate
southern Brazil and semiarid northeastern South
America, comprising a large variety of silicate rocks of
widely different age and tectonic history. Paleozoic and
Mesozoic sandstone sequences, although devoid of the
tectonic processes that affected older quartzites, display
similar lithological features and geomorphological pro¬
cesses. Sandstone caves can, thus, be discussed in the
same context as quartzite caves. Devonian sandstone of
the Maecuru Formation in Eastern Amazonia hosts at
least four major caves over 1 km in length. Devonian to
Silurian sandstones of the Alto Garzas and Vila Maria
Formations in western Brazil (Chapada dos Guimaraes
National Park and surroundings) contain two impor¬
tant and large caves. The Furnas Formation in southern
Brazil is home to several shafts and important caves,
although few tend to surpass 1 km in length.
CAVE MORPHOLOGY
Quartzite caves display a variety of patterns, some
of which find equivalents in traditional carbonate karst
settings (Fig. 4). Branchwork caves, with many pas¬
sages containing active streams, are common. Gruta de
Torras (Fig. 4A) is a good example, in which a small
stream collects water from a few tributaries and runs
for a few kilometers along the dip and the strike of a
quartzite bed. Passages following the dip tend to form
laterally wide passages with a uniform slope. Passages
along the strike display a series of waterfalls followed
by deep pools. Interestingly, although it is very com¬
mon for quartzite cave passages to contain streams.
cave entrances are rarely associated with swallets,
streams originating inside the cave as localized outlets
of the water table. There is no clear relationship
between underground stream discharge and surface
catchment areas, especially considering that many of
the upstream entrances lie close to the top of the cuesta
slope.
Anastomotic caves are well represented by Cueva
Roraima Sur (Ojos de Cristal) (Fig. 4B). At the top of
the Venezuelan tepuis, a much more friable and uncon¬
solidated unit of the Mataui Formation is sandwiched
between more hardened quartzites. This easily eroded
unit allows for the development of a series of intercon¬
nected passages sometimes divided by remarkable
quartzite pillars.
Another common type of quartzite cave is repre¬
sented by enlarged deep fissures. Major fissure caves,
such as Gruta do Centenario (Fig. 4C) and Gruta da
Bocaina (3.2 km long and —404 m deep), occur at the
top of Inficionado Peak (Fig. 3). They display a number
of curious features not found elsewhere, such as
absence of a proper ceiling (twilight being provided by
the curving vertical profile of the fissures) and cave
levels generated by breakdown edged in the fissures,
voids being filled with smaller pebble and sand sedi¬
ments, some even containing cave streams. The com¬
mon presence of cave drainage in such a small area is
due to the very wet climate characteristic of these high
peaks.
Many quartzite caves are shallow features, closely
paralleling the slope of the surface. Cueva Roraima
Sur is seldom more than 20 m deep in relation to
the surface. Gruta de Torras is still shallower.
These caves exhibit frequent upper entrances due to
ENCYCLOPEDIA OF CAVES
638
Q
CONTROLS ON QUARTZITE CAVE
GENESIS AND DEVELOPMENT
FIGURE 4 Cave patterns (at the same scale) of some representa¬
tive quartzite caves. (A) Gruta de Torras, south-central Brazil, a
branchwork cave with few tributaries. (B) Cueva Roraima Sur (Ojos
de Cristal) in southern Venezuela, an anastomotic cave. Map by
Sociedad Venezolana de Espeleologia. (C) Gruta do Centenario, a vertical
network cave in southeastern Brazil. Map courtesy of Grupo Bambui de
Pesquisas Espeleolo'gicas.
collapse and interception by surface fissures.
Although some caves contain active river passages that
can be followed downstream from the upper entrance
to resurgence, they appear to represent perched
streams. At the Venezuelan tepuis as well as at
Inficionado Peak, river passages may terminate at the
cliff face, hundreds of meters above the base of the
mountain. At the gently sloping caves of eastern
Brazil, some caves, such as Gruta de Torras, end at nar¬
row impassable fissures, but a characteristic type of
passage termination, especially common at Ibitipoca
Ridge, is the "sand siphon," in which a passage not
being able to transport sand produced by the disaggre¬
gation of walls and ceilings ends up being choked with
sand.
Small-scale morphological features in quartzite caves
may include pendants, upper levels, canyon-shaped
passages, and remnants of phreatic tubes. The friable
nature of the rock may prevent preservation of original
speleogenetic landforms. Biokarst has also been reported
for Charles Brewer Cave in the Chimanta tepui.
Quartzite tends to be a relatively homogeneous
rock, comprised of quartz grains with (or without) a
matrix composed of amorphous silica. A series of other
minerals such as micas (biotite) and kaolinite are com¬
mon constituents. Dissolution of these rocks proceeds
slowly but tends to increase at higher temperatures.
Amorphous silica, in particular, is quite soluble under
tropical climates (>100 mg L -1 ). The geochemistry of
quartzite cave waters in many areas in Brazil and
Venezuela has demonstrated that surface streams and
percolation waters are very acidic but also has shown
very low content of dissolved silica, suggesting that
dissolution must proceed very slowly. Timescales for
speleogenesis may be orders of magnitude higher than
in carbonate areas. The very old nature of the rock
and stable tectonic condition of most of the quartzite
cave areas allow for such expanded timescales for
speleogenesis.
Quartzite cave genesis appears to involve a sequen¬
tial process comprising both chemical and erosional
mechanisms. The most applied hypothesis involves
dissolution of the matrix, prompting release of quartz
grains. Since many quartzite rocks do not possess
matrix, dissolution at the boundaries of quartz grains
may also occur. A number of alternative mechanisms
have been proposed. Removal of kaolinite has been
shown to favor quartz removal at Furnas Formation in
southern Brazil. Chemical alteration of micas (mostly
biotite) may also provide enough space for the loosen¬
ing quartz grains. Field observations in many caves in
Brazil and Venezuela have shown that quartzite rock
inside caves tends to be very porous, water soaked,
and friable. The initial process of quartz loosening
appears to create favorable conditions for erosional
removal of the rock, a process known as arenization.
Some authors have even suggested that dissolution is
not required to explain the formation of quartzite
caves, although most sources tend to agree that disso¬
lution, although responsible for a small portion of rock
removal, plays an essential initial role in karstification
and conduit generation.
Some caves appear to have developed upstream
from a scarp through sapping processes. These caves
usually contain a perennial or temporary stream and
tend to progressively narrow as one goes upstream.
Many quartzite caves exhibit a clear lithological con¬
trol. Tepui caves develop mostly along the more
unconsolidated unit of the Mataui Formation, the
harder overlying quartzite providing ceiling support.
At Chapada Diamantina area, some caves, such as
Torras Cave, show passage control by a laterite-rich
alteration zone. Kaolinite- and mica-rich layers also
ENCYCLOPEDIA OF CAVES
QUARTZITE CAVES OF SOUTH AMERICA
639
appear to provide a more favorable zone for conduit ini¬
tiation. In general, the absence of deep phreatic passages
or springs and the close relationship between surface
topography and cave profiles may suggest that most
quartzite caves may originate close to the water table. At
present many caves, such as in the tepui area or
Inficionado Peak, although containing active streams, lie
well above the present water table.
Fissure caves appear to display other additional spe-
leogenetic mechanisms. These caves tend to be concen¬
trated close to very steep scarps, being less common at
the inner zone of tepuis. They appear to be related to
unloading joints due to stress release associated to the
proximity to the scarp line. These caves are, thus, at
least in part tectonic features enlarged by dissolutional
and erosional processes. Observations at the very deep
entrance shafts of Inficionado caves show that cave walls
(and thus fissure width) enlarge through flaking pro¬
cesses, water-soaked slabs of quartzite undergoing dis¬
aggregation, and gravitational spalling. Fissures tend to
become progressively narrower at depth.
Breakdown processes are very common in quartzite
caves. The quartzite conglomerate Lapao Cave in the
Chapada Diamantina area is a massive cave passage,
now entirely filled with breakdown. In other caves,
cave streams will rapidly disaggregate the breakdown
and transport sand downstream.
CAVE DEPOSITS AND AGE
Clastic sedimentation in quartzite caves tends to be
monotonous, comprising basically quartz sand and
breakdown. The higher relief of these areas prevents
the occurrence of overlying formations and soil. In
areas where kaolinite forms the rock matrix, red fine¬
grained sedimentation may be ubiquitous. The profu¬
sion and abundance of speleothems commonly found
in carbonate karst is largely absent in quartzite caves.
Speleothems are mostly represented by small opal cor-
alloids and crusts on cave walls. Flowstone of gelati¬
nous composition is sometimes observed. An exception
is provided by the remarkable speleothems found in
some of the caves in the tepui area, especially at the
Chimanta tepui. Large globular or filamentous
speleothems are unlike anything described elsewhere.
Genesis is interpreted as being associated with con¬
densed cave moisture mediated by biological processes.
Some speleothems have been described as stromatolites
formed by silicified filamentous microbes (either bacte¬
ria or cyanobacteria). These opalline-laminated spe¬
leothems have been successfully dated by the U-series
method. Their slow growth and chemical composition
may yield useful paleoclimate information.
Low rates of dissolution, low content of dissolved sil¬
ica in the water, and the old age of the landscape have
led to the widespread assumption that quartzite caves
may be extremely old. This may well be the case,
although direct dating evidence is lacking. On the other
hand, erosional processes in unconsolidated quartzite
may proceed extremely fast once an underground route
is wide enough to evacuate the frequent sand residue.
Initiation of quartzite caves may proceed in timescales
an order of magnitude slower than in carbonate areas,
although cave evolution may be quite fast.
Bibliography
Aubrecht, R., Brewer-Carias, C., Smfda, B., Audy, M., & Kovacik, L.
(2008). Anatomy of biologically mediated opal speleothems in the
world's largest sandstone cave: Cueva Charles Brewer, Chimanta
Plateau, Venezuela. Sedimentary Geology, 203, 181—195.
Aubrecht, R., Lanczos, T., Smida, B., Brewer-Carias, C., Mayoral, F.,
Schlog, J., et al. (2008). Venezuelan sandstone caves: A new view
on their genesis, hydrogeology and speleothems. Geologia
Croatica, 61, 345—362.
Galan, C., Herrera, F., & Astort, J. (2004). Genesis del Sistema
Roraima Sur, Venezuela, com notas sobre El desarrollo del karst
em cuarcitas. Boletih de la Sociedad Venezolana de Espeleologia, 38,
17-27.
Lundberg, J., Brewer-Carias, C., Ik McFarlane, D. A. (2010).
Preliminary results from U-Th dating of glacial—interglacial
deposition cycles in a silica speleothem from Venezuela.
Quaternary Research, 74, 113—120.
Martini, J. E. J. (2000). Dissolution of quartz and silicate minerals.
In A. Klimchouk, D. C. Ford, A. N. Palmer & W. Dreybrodt
(Eds.), Speleogenesis: Evolution of karst aquifers (p. 527).
Huntsville, AL: National Speleological Society.
Piccini, L., & Mecchia, M. (2009). Solution weathering rate and origin
of karst landforms and caves in the quartzite of Auyan-tepui
(Gran Sabana, Venezuela). Geomorphology, 106, 15—25.
Wray, R. A. L. (1997). A global review of solutional weathering forms
on quartz sandstones. Earth Science Reviews, 42, 137—160.
ENCYCLOPEDIA OF CAVES
TYPES OF CAVES USED
RECREATIONAL CAVING
John M. Wilson
Marks Products, Inc.
RECREATIONAL CAVING DEFINITION
Recreational caving may be defined as the activity of
entering a void, such as a cave, for the pure joy of the
activity. When entering a cave, cavers would expect it
to be totally dark, even during normal daylight, for the
experience to be truly considered as caving. Most
cavers generally agree that the definition of cave is "an
underground void large enough for a person to get
into." Although some measure of fun can be had
checking out small cave-like openings, true recrea¬
tional caving involves exploring longer cave passages,
which also might include manmade voids such as
mines and tunnels. A few deep pits and caves have
long, relatively straight entrances that allow daylight
to penetrate the entire cave, but cavers still consider
these to be valid caving caves. Deer Cave in Malaysia
and Neversink Pit in Alabama are examples of excep¬
tions to the total darkness rule. The information in
this article is based on caving in true recreational caves
or voids.
Recreational caving as defined here includes sport
caving as well as relatively easy cave recreation that
might involve only walking passages and visiting
caves in large groups. Recreational caving may overlap
with more avocational or professional caving purposes
when recreational cavers add speleological activities to
their cave visits or vice versa. These avocational and
professional activities may include photography, map¬
ping, and various cave-related scientific specialties.
Cave activities or caving have become a subordinate
occupation for avocational cavers, who give many vol¬
unteer hours for the joy of it, as material compensation
for caving is rare.
The following five types of natural caves used in rec¬
reational caving are ranked by popularity, based on
information obtained from cave publications and other
sources. There are significant differences in cave popu¬
larity in some geographic areas, depending on the types
of caves available for caving.
Solution Caves
Limestone and other solution caves are by far the
most used by cavers for recreation. They are formed
by several types of solution processes described else¬
where in this book. Availability is the most important
factor in determining the popularity of types of recrea¬
tional caves in a given geographic area. This is why
limestone caves top the list of popular caves. This
group also includes caves formed in marble, dolomite,
and gypsum. These three types make up a small part
of the recreational caving in solution caves.
Lava Caves
Lava caves, which are often referred to as lava tubes,
are the second most common type of cave used for rec¬
reational purposes. Cavers cave in lava tubes as well
as lava voids that are not in tube form. These caves
result from flowing lava and other processes described
elsewhere in this book. Many lava caves have rela¬
tively flat floors and multiple entrances, making them
suitable for those with limited caving skills. Lava caves
do have their own special risks. They often have dark,
very rough-textured surfaces with sharp edges that
absorb light, reducing the effectiveness of caving light
systems. Many novices have been amazed to discover
the extent of torn clothing after traversing small pas¬
sages in lava caves, in contrast to an equivalent-sized
passage in a limestone cave.
Encyclopedia of Caves.
641
© 2012 Elsevier Inc. All rights reserved.
642
R
Sea Caves
Sea caves are formed primarily by wave action at
the shores of oceans and large lakes. They are the most
common type of erosion cave. These caves are less
common than limestone or lava caves and are always
located at past or present sea levels. The relative
predictability and attractiveness of their locations
make sea caves the third most popular type of cave for
recreational caving. Some sea caves have considerable
incidental visitation by people primarily doing other
forms of recreation. Many sea caves have no totally
dark areas; however, this apparent deficiency is more
than made up for by other factors. The interaction of
tides and waves and the often abundant variety of life
forms in and near sea caves are appealing to cavers
who would not normally visit these comparatively
small caves if they were not next to the ocean.
Some partially flooded caves at sea level were formed
primarily by solution processes or flowing lava.
Technically, these are not sea caves but are solution
or lava caves. Recreational cavers may not be able
to make this distinction, and visitation information in
some reports indicates visits to sea caves when it was
in fact that caving occurred in solution or lava caves.
Talus and Tectonic Caves
Talus caves are formed when very large rocks fall
from mountains or cliffs. The spaces between the rocks
are sometimes large enough to allow a person to enter.
Tectonic caves are formed by tectonic processes along
faults near the Earth's surface. Few people use talus
and tectonic caves for recreation, as they are usually
small and have few caving features of interest to most
recreational cavers, making these the fourth most pop¬
ular type of cave. Little documented caving occurs in
these types of caves; therefore, ranking them sepa¬
rately is not statistically meaningful.
Glacier Caves
Caves are formed in glaciers and are the least popu¬
lar type of the five natural voids used for recreational
purposes. These caves are cold, wet, and unstable.
Specialized skill and equipment are necessary to
explore ice caves successfully. These caves are located
only in glaciers and icefields; thus, they are rare and
remote from population centers. Sometimes lava caves
with permanent ice deposits are referred to as ice
caves, but these ice deposit caves are still considered
lava caves, despite the common mistake of referring to
them as ice caves.
Mines and Tunnels
These and other voids in the Earth created by peo¬
ple are rarely used for recreational caving purposes
because of their instability. In terms of absolute visita¬
tion, mines are used by mine workers far more inten¬
sively than caves are used by recreational cavers.
Miners use some of the same equipment as cavers.
Once abandoned, mines tend to become unstable, and
owners are often obligated to construct gates or other
barriers and post entrance restrictions. Among the few
interesting exceptions to these generalizations about
mines is the Wieliczka Salt Mine near Cracow, Poland.
Part of the mine is developed as a show cave, but
cavers should not expect to go caving in the mine, as
visitors are limited to the show cave area.
OTHER VOIDS AND URBAN CAVING
Some manmade voids are occasionally used by peo¬
ple for caving. They might even use some of the same
type of equipment, but in most places it is not recom¬
mended by caving organizations due to safety and
legal considerations. As with all caving, the permission
of the property owner should be obtained before enter¬
ing. Most owners of abandoned mines and buildings
will not grant permission for recreational purposes.
However, some urban areas have mines, aqueducts,
tunnels, and quarries that have become popular with
people for recreation. The location of these voids
near major population centers probably contributes to
their use. One of the best examples is "Les Carrieres
de Paris" or Catacombs of Paris. These tunnels under a
portion of the city of Paris were originally Roman-era
quarries that were modified in the eighteenth century
as mass tombs to resolve severe overcrowding in
cemeteries. Beside human bones, visitors will find con¬
siderable upscale graffiti in an extensive underground
maze.
ACTIVITIES NOT CONSIDERED CAVING
Some activities that appear to be caving are usually
not considered recreational caving by cavers. Visiting a
show cave and show caving are activities done by
many more people than recreational caving. Some sur¬
face activities, such as looking for caves (which cavers
call ridge walking), are not considered caving in this
section. This section presents a more precise descrip¬
tion of traditional recreational caving by not including
data from show caving and cave-related surface
activities.
ENCYCLOPEDIA OF CAVES
RECREATIONAL CAVING
643
Shelter Caves
Shelter caves are small, naturally occurring voids
that may be formed by any of the previously described
cave formation processes. In addition, aeolian caves
formed by wind are usually considered shelter caves.
Many people visit shelter caves as an incidental part of
other surface recreational activities, so there is little
documentation of their visits. Because shelter caves are
penetrated by daylight and require no caving equip¬
ment, visiting them is usually not considered recrea¬
tional caving.
Show Caves
Show caves have manmade improvements that
allow for easier passage; thus, almost anyone can go
into one of these caves without any special skills or
equipment. Show caves are managed as a commercial
business that charges an admission fee or are main¬
tained by a governmental agency, often as part of a
park system, and may also require a user fee. Show
caving is not considered caving by most cavers
because of the cave improvements. The major disquali¬
fying improvement is the artificial light provided, usu¬
ally as a permanently installed electric light system or
as a communal light carried by a tour guide. Some
show caves have supplemental offerings and are
referred to as wild caving. These cave trips may or may
not be led by a guide. In either case, wild caving in
show caves is considered recreational caving.
Manmade Show Caves
Manmade show caves are usually built to provide a
hands-on educational experience for the public. The
historic cave paintings in Lascaux, France, are shown
to the public in a manmade cave that was created
because the originals were being threatened because of
changes in environmental conditions associated with
visitation. Some such manmade structures do include
the kinds of caving challenges that appeal to some rec¬
reational cavers and provide an opportunity to practice
caving skills. Visiting these manmade show caves
might be considered recreational caving, depending on
the specific caving conditions.
Abandoned Buildings
Little information is available on the recreational
exploration of these types of structures, perhaps
because it is often illegal. Manmade surface enclosures
usually do not appeal to the traditional caver, although
some of the same caving skills and equipment are
used by the few people who do enter these structures.
Most cave organizations do not recommend caving in
abandoned buildings because of structural instability
and legal considerations.
Factors Contributing to Recreational
Cave Visitation
Some caves become very popular, with visitation
averaging several hundred people a week; however,
most caves rarely see humans, and the few who
do enter have other purposes, such as mapping or
research. Cave popularity can be affected by word of
mouth, the media, availability of other recreational
options, and cultural changes. Recent increases in the
popularity of other forms of recreation and extreme
sports may have provided alternatives for potential
cavers; thus, the number of recreational cavers is most
likely lower than it might be otherwise. Cave visitation
can be influenced by the following six criteria, listed in
a plausible order of importance:
• Known existence. The cave must be known to exist,
and such knowledge is often spread by those taken
to caves by previous visitors. The location of a
cave can also be spread by word of mouth and by
various print and electronic media. Keeping the
location of a cave secret, at least in the short run,
has significantly delayed the advent of recreational
caving to these caves.
• Physical access. For visitation, a cave must have
physical access. Factors affecting physical access to
a cave might include gates, barriers, or proactive
access control, as well as caves being difficult to
find or requiring long hikes from the nearest road
to reach the entrance. Other physical factors include
vertical entrances, which reduce visitation of
nonvertically equipped cavers to near zero but
attract those with vertical caving capabilities. Water
barriers or in-cave sumps usually keep all but
properly equipped cave divers from entering.
• Distance to cave. With all other factors being equal,
the farther a cave is from a given starting point, the
less likely it is to be visited. The percentage of the
population who has caved is higher in communities
located in karst areas than in areas more distant from
suitable caves; however, even moderate distances
may change the nature of how people organize in
order to cave. It appears that cavers living in urban
areas tend to organize more effectively, perhaps to
accommodate their transportation needs; thus, these
organized cavers may continue to remain actively
involved longer than those people who live close to
caves but are not organized into a cave group. It is
true that many very experienced cavers travel all
over the world to go caving; however, most
ENCYCLOPEDIA OF CAVES
644
R
recreational cavers are not experienced and will go
on only a few cave trips in their lives. Recreational
cave trips tend to be to caves that are relatively
nearby. It is reasonable to assume that a higher
percentage of existing caves have been discovered in
areas that are easy to access.
• Caver appeal. Popular caves have one or more
interesting features that appeal to the caver, such
as significant size and complexity, aesthetics,
unique or unusual geologic features, interesting
biota, fun caving features, caving challenges, or
cultural mystique. The attractiveness of a cave to
cavers often depends on the degree to which cavers
can do the things they like to do in a cave, such as
climbing through complex passageways, crawling
into different parts of the cave, and exploring and
seeing interesting cave features.
• Suitability for group caving. Almost all recreational
caving is done in groups of 3 to 30 people. Caves
without places where people can stop and socialize
or at least communicate are noticeably visited less
often by recreational cavers.
• Legal access. Anarchy exists in a few places in the
world, and social deviance occurs to some degree in
all cultures. These factors partially account for the
fact that the single act of legal posting or closing of
a cave is not always effective. Cave owners usually
take additional actions if they intend to achieve total
compliance with their cave restrictions. Some posted
caves with little or no enforcement may continue
to have cave visitation; the extent of the trespassing
depends on the respect for property rights among
the area's cave visitors. All respected cave
organizations recommend that cavers should never
trespass in a cave that has been gated or posted.
Serious consequences of ignoring such postings
include lawsuits, serious injuries, and fatalities,
partially due to spelunkers disregarding cave
closings.
WHY PEOPLE BEGIN CAVING
There is no single reason why people enter caves.
Attempts to sum up caving using a short, catchy
phrase have resulted in oversimplified explanations.
People who cave do so for many different reasons.
Non-cavers may find it difficult to imagine why any¬
one would crawl in the mud, climb in the dark, and go
into tight places. Most people start caving to have an
enjoyable experience. The most common reasons why
people cave are listed below. They apply to novice as
well as avocational cavers.
• To enjoy the company of others in a fun group activity.
The social aspect is a significant factor for novice
recreational cavers. Schools, youth groups, and
community groups often sponsor cave trips.
• To explore or have an adventure. (Zuckerman, 1994)
Maze exploration is considered great fun by many.
The excitement of cave exploration can be compared
to exploring a complex maze. The curiosity to see
what is around the bend in the passage or beyond
the extent of one's light leads the caver on. The
increased adrenaline and other hormone production
in the stimulating cave environment provides a
pleasurable sensation. Some people enjoy this
experience and will return for more. Psychologists
(Archer and Birke, 1983) have provided strong
evidence from studies of animals and people that
exploratory behavior is an innate characteristic. If so,
this behavior may explain why people cave; however,
it does not explain why they cave as opposed to
choosing other forms of exploration. It is unlikely that
an innate exploratory need completely accounts for
caving behavior, as most people explore other things
and never experience the inside of a cave.
• To accomplish something unique, to see things that most
people have not seen or done, or to learn something
about caves and nature. For example, some cavers have
the goal of discovering a "virgin" cave, a cave passage
thought to have had no prior entry by humans.
• To engage in a physical or sporting activity for the
personal challenge or just for the exercise.
THE OCCASIONAL DOWNSIDE
Sometimes a person goes on a cave trip with mis¬
taken assumptions, believing that it will be an exciting
adventure full of glamorous achievements and reveal¬
ing a hidden world of wondrous vistas, not to mention
the possibility of finding something of value such as
a hidden treasure. It is a considerable disappointment
when the new caver discovers reality. It is often hard
work getting to and through caves. Much of the time is
spent traversing passages that are anything but glamor¬
ous, and the wondrous sites are few and far between.
No crowds are cheering the caver on. Actually, only a
few other people care about these exploits. On top of all
that, there is no buried pirate or other kind of treasure
in caves, and in most places laws prohibit the removal
of mineral formations and artifacts. In fact, most cavers
help enforce these laws, and the National Speleological
Society (NSS) even offers a reward of up to $1000 for
the arrest and conviction of anyone guilty of cave van¬
dalism (check with the NSS about the current status of
this award). It is not unusual for a new caver to emerge
ENCYCLOPEDIA OF CAVES
RECREATIONAL CAVING
645
from a cave wet, muddy, bruised, and tired with sore
muscles he did not know he had and nothing much to
show for his efforts, with memories being mostly of
mud and rock seen from a distance of 20 cm while
crawling. He has probably decided this will be his first
and last cave trip. In fact, he has a lot of company, as
between one fourth and one half of the people who
have gone caving once will never go on another cave
trip. This figure does vary, depending on the type of
cave trip in which the person participated and the
types of future cave trips available. Cave trips that are
competently led in caves appropriate for the first-time
caver will lead to higher repeat caving rates.
NOVICE CAVERS
Since the founding of the National Speleological
Society in 1941, there have been significant improve¬
ments in caving techniques, equipment, and safety
procedures. Many recreational cavers are among the
best-equipped and most experienced cavers. As a result,
recreational cavers have penetrated farther into caves
and coped with more challenging situations.
Inexperienced cavers sometimes try to emulate com¬
petent recreational cavers with bad results when they
exceed their skill level and have inadequate equipment
for the situation. Novice cavers make up a significant
portion of all recreational cavers and sometimes can
be identified by their inappropriate equipment,
misdirected motivation, or lack of cave knowledge.
Some novices who have not had a good orientation
and/or have had no contact with the mainstream of
the organized caving community may do strange and
sometimes dangerous things in caves. For example,
while there may be some instances when the following
items might be appropriate in a cave, the possession of
some of these items in a cave and the kinds of reasons
given for having them usually identify the person as
being either unprepared for cave trips or someone
who has plans to vandalize the cave. In general, one
should not take any of the following into a cave on a
recreational trip:
• Large knives for killing snakes, bats, or other
wildlife
• Guns for shooting bears or outlaws
• Handheld flashlight or candle as the primary or
only light source
• Inadequate or unreliable light sources or no light
source
• Clothesline or other light-duty rope for rappelling
down a drop or pit
• Hemp rope for climbing hand over hand out of a pit
• Balls of string to unwind and mark one's route in
order not to get lost
• Hand-carried coolers for beverages, especially
alcoholic beverages
• Alcoholic beverages or mind-altering drugs
• Paint, especially a can of spray paint
• Equipment for collecting cave mineral formations
Except when it is clear that such equipment has been
brought along for certain professional or ceremonial
activities, observing someone with any of this equip¬
ment should cause enough concern to take appropriate
action. Some cavers might try to advise such a person
of the inappropriateness of the items and recommend
corrective action before going caving. Cave conservation
and appropriate equipment and techniques are dis¬
cussed in other chapters of this book.
Unfortunately, a few recreational cavers have
engaged in destructive activities such as painting graf¬
fiti in a cave, breaking speleothems, stealing mineral
formations and artifacts, and harming cave life. Many
states and some countries have cave protection acts
that provide for punishment under the law for harm¬
ing caves or their natural inhabitants. Cavers will
sometimes refer to cave vandals as spelunkers.
The term spelunker is often applied to recreational
cavers. The word is derived from the Latin word spe-
lunca ("cave") or the even earlier Greek word, spelaion.
However, in the organized caving community today,
comprised of those who have learned proper techni¬
ques for cave exploration and developed an awareness
of the fragile nature of this underground resource, the
term spelunker has become a derogatory term. The
majority of English-speaking cavers would not refer to
another caver as a spelunker, although there are a few
exceptions to this rule. Two cave clubs in Missouri are
composed of competent cavers with an independent
spirit who refer to themselves as spelunkers.
WHY DOESN’T EVERYONE CAVE?
Some avocational cavers have wondered why every¬
one does not go caving, while other cavers are glad
everyone doesn't. Millions of people living in North
America have caved at least once. Professional and avo¬
cational cavers number in the tens of thousands. For
these thousands, caving is the epitome of experiences
life has to offer. The stereotypical cave trip involves
crawling or wiggling through tight places, sometimes
no more than 20 cm high, where mobility is severely
restricted. Sometimes cavers must crawl through a
stream or slide through a passage on their back, with
only a few inches of air space between the water and
ceiling. Some cavers thrive on these challenges; others
ENCYCLOPEDIA OF CAVES
646
R
do not. Anyone with even mild claustrophobia will be
unlikely to return to this type of activity. The same is
true of people with acrophobia, as many caves have
floors that are little more than a series of rocks,
boulders, crevices, and pits. Bats and a variety of inver¬
tebrates do not appeal to everyone. Erroneous beliefs
and imaginary threats stop others from even consider¬
ing going into a cave. The vast majority of the popula¬
tion does not enjoy these types of experiences. They
either avoid exposure to such situations or accept them
as a one-time experience in life. Some people enjoy
a more sedentary lifestyle that is incompatible with
sport caving. Most caving requires effort, sometimes
strenuous effort, to participate.
Other limiting factors include vertical caves with
entrance drops or pits. These entrance drops may be
anywhere from a few meters to more than 300 meters
deep. Special skills and equipment are needed for this
type of caving. Many difficult cave rescues have been
a result of poorly informed and inadequately equipped
recreational cavers attempting to descend into vertical
caves.
A CAVING LEVEL FOR ALMOST
EVERYONE
People interested in caving are pleased to discover
that they can choose caving activities appropriate to
their training and interests. Caving varies widely in its
intensity, required skills, and equipment. Exploring
some caves requires little more effort than taking a
short hike; such caving requires no special equipment
beyond a good light system. Diversity is the very
nature of caves, and cavers can pursue their interests
in many different cave environments. Caves present
a continuum of difficulty requiring more organization,
planning, and stamina. Cave divers require special
training and underwater breathing equipment.
Someone seeking an athletic challenge or someone
who is an extreme sports aficionado can find caves
suited to their goals. Cavers who search for new caves,
draft cave maps, or take cave photographs have a
seemingly endless opportunity to find and record
nature's handiwork. A cave is usually a low-energy
environment that is often isolated from outside envir¬
onments and hence spawns many different life forms.
Caves are a natural attraction for geologists and biolo¬
gists. This variety of activity attracts people with
widely varying interests, nowhere more pronounced
than at a National Speleological Society annual
convention, where the program includes numerous
special-interest sessions, contests, and workshops
stressing various cave-related skills, as well as hun¬
dreds of different presentations.
WHY DO PEOPL E CON TINUE TO CAVE?
The transition from recreational caving status to
being an avocational caver or speleologist often occurs
as these cavers find that caving is more interesting if it
has a purpose. Caving merely to do informal explora¬
tion of a cave that many others have seen and studied
before may no longer satisfy the needs of the more
adventurous. A few cavers have adopted caving as a
life-long avocation. The careers of most speleologists
were preceded by some recreational caving. Others
who were recreational cavers at one time may have
stopped caving for various reasons but maintained
their support of cave organizations and goals, and are
involved in other cave-related activities.
The hierarchy of human needs model proposed by
Maslow (1954) describes the sequence in which people
focus their efforts on meeting their various needs—
physiological, safety, belonging and love, esteem,
self-actualization, cognitive, and aesthetic. People meet
their more basic needs first and then proceed to
address their higher needs. Considering the character¬
istics of large, complex, and scientifically interesting
caves, it can be easily understood how caving meets
the self-actualization, cognitive, and aesthetic needs of
an avocational caver. Participation in cave-related
organizations may also meet some belonging and
esteem needs. A person mainly focused on basic needs
and trying to survive is unlikely to participate in
caving.
DEMOGRAPHICFACTORS
Contemporary studies conducted by members of
the National Speleological Society have discovered
several significant differences between cavers and the
general population. None of these studies has been
published outside local caving newsletters, but the
results are intriguing:
• Age. The mean age of recreational cavers, as
determined by information gathered from cave
registers between 1975 and 1985, was 21 years
(mode, 18 years). The frequency of caving decreases
significantly with increasing age. Anecdotal
information obtained recently indicates that the
mean age of cavers may be greater by several years.
• Sex. Males make up 75% of cavers and an even
higher percentage of cavers on cave trips, which
indicates that males have taken a greater average
number of cave trips. There does not appear to be
any component of recreational caving that would
favor one sex over the other. Most people say that
cultural factors account for this difference. While
ENCYCLOPEDIA OF CAVES
RECREATIONAL CAVING
647
this may be true, it is not an adequate explanation.
A more explicit hypothesis suggests that caving
activity is closer to traditional male activities such as
outdoor physical activity, exploratory behavior, and
risk taking. These activities, plus being covered with
mud and dirt when caving, are not those normally
associated with traditional feminine activities. A test
of this hypothesis would occur in a society where
the roles of women have veered away from the
traditional feminine roles; if caving is primarily a
culturally linked phenomenon, then one would
expect the percentages of female cavers in such a
society to increase.
• Race. In North America, Caucasians are the most
numerous ethnic group, representing at least 95%
and probably more of all cavers. There are far fewer
African-American cavers than would be expected,
given their percent of the population. There is
anecdotal information that caving activity among
Hispanic populations is lower than for non-Hispanic
Caucasians. Recreational caving exists in most parts
of the world, but data indicate that the percentage of
the population engaging in recreational caving is
highest in Europe and North America, although no
definitive studies on this issue have been conducted.
One plausible explanation that may account for some
of these differences could be that recreational caving
is similar to other forms of recreation that require
significant leisure time and equipment. While the
cost of basic caving equipment is not considered
expensive by Western standards, it might be beyond
the reach of many. Maslow's hierarchy of needs may
also help explain this difference, as people will focus
most of their attention on meeting basic needs before
acting on higher needs. The hypothesis is most
plausible if one assumes that in Europe and North
America a greater percentage of the population is
focused on meeting their higher needs, and
recreational caving fulfills these needs for people
who are active cavers. While Maslow's hierarchy of
needs may account for some caving frequency, other
factors are at work here that have not been studied
in relation to recreational caving.
• Length of cave trips. The length of cave trips can last
from a few minutes to several days. Most recreational
cave trips fall within 2 to 8 hours, and the average is
about 4 hours. The length of cave trips is mostly a
factor of the size of the cave, difficulty traversing a
particular cave, and the interest and ability of each
caver.
• Cycles in cave visitation. Cycles in the frequency of
caving do occur (e.g., seasonal variations), with the
warmer months being more popular than the colder
ones. Because most caves have a relatively constant
temperature year-round, it appears that factors
other than temperature are more important in
determining when to go caving. Also, weekends are
generally significantly the most popular time to visit
caves.
CONTEMPORARY TRENDS
The Outdoor Foundation (2010) published its most
extensive survey ever on outdoor recreation. The
results show a continual decline in participation in
many forms of outdoor recreation from 2006 through
2009. A few activities such as adventure racing, snow-
shoeing, and kayaking did have significant increases
in participation. Caving was not included as a category
and the report does not mention caving or caves at all.
However, an activity such as sport climbing could
serve as a proxy for recreational caving in lieu of any
other systematic study. The survey showed a modest
drop in participation in the four years of the study.
This study gave some of the reasons young people
gave for not participating in outdoor recreational activ¬
ities. They include lack of interest, no time, no access,
prefer TV and movies, computer or video games, and
hanging out with friends.
From 1950 to 1980, young people from the age of 15 to
30 dominated participation in caving. That demographic
is relatively much less common now. This change in
society away from outdoor activities is dramatically on
display at many gatherings of cavers in which the older
ones far outnumber the young. Historically, the vast
majority of older cavers and members of groups such as
the National Speleological Society started caving when
they were young.
Social norms evolve over time and can be driven by
technological change. Today, the ever-present Internet
and other electronic media offer an alternative method
for work, recreation, and finding information. The real¬
ity is that people now have more choices in how to use
their time. If one wanted to learn about caves or have
a visual cave experience, a case could be made that the
most efficient method would be to use the Internet.
For example, there are now many times more cave
photos on the Internet than there were available in all
printed material published more than 30 years ago.
SUMMARY
While recreational caving has parallels with other
forms of recreation, its diverse nature allows people to
enjoy an activity in the same places and sometimes in
the same organizations as professionals with similar
interests. People cave for many different reasons, at
many different skill levels, and in several different
ENCYCLOPEDIA OF CAVES
648
R
types of caves requiring a variety of equipment and
techniques. While young males are the most common
cavers, all types of people can be found caving. People
have clear preferences in the features and types of
caves they choose and in the distances they choose to
travel to cave. Recreational caving can be a risk to
both the cave environment and the caver. The effort
required to navigate through restricted passageways
filled with water and mud, drops, rough terrain, and
assorted other obstacles is enough to keep most people
from trying it or discourage them from returning.
These same features are part of the appeal of a cave
and contribute to the cognitive challenges, diverse
types of participation, and spectacular aesthetics asso¬
ciated with caves. These features may also be less
unique with the advent of newer forms of recreation
and electronic media which may offer alternatives
that are contributing to a declined in participation in
recreational caving.
See Also the Following Article
Exploration of Caves—General
Bibliography
Archer, & Birke, L. (1983). Exploration in animals and humans. New
York: Van Nostrand-Reinhold.
Maslow, A. H. (1954). Motivation and personality. New York: Harper.
The Outdoor Foundation (2010). Outdoor recreation participation
report. Boulder, Colorado, www.outdoorfoundation.org.
Zuckerman, M. (1994). Behavioral expressions and biosocial bases of
sensation seeking. New York: Cambridge University Press.
RESCUES
John C. Hempel
EE1 Geophysical
From time to time those who explore caves do not
return to the surface under their own power. Trained
teams of cavers take it on themselves to rescue cavers
who are in trouble. Rescues are required for many rea¬
sons: there may have been an accident and a caver is
injured, cavers may be trapped by rising water, and
cavers may simply have lost their lights and are unable
to move or become lost. Cave rescue is a highly orga¬
nized activity in the United States and Europe
(Hempel and Fregeau-Conover, 2001).
INTRODUCTION
Since humans began using caves for habitat, recrea¬
tion, or religious reasons, accidents have been a real
threat to those exploring the netherworld. Prior to elec¬
tronic communication, cave rescues went almost unno¬
ticed outside of the community in which they occurred.
Since the 1920s, media coverage of cave rescues has
turned unfortunate ordeals into public entertainment.
Cave rescues first attracted national attention when
Floyd Collins, a young caver and explorer, became
trapped in Sand Cave, Kentucky, in 1925. His attempted
rescue and death have been the subject of records,
movies, books, and documentaries. During the attempt
to free Floyd, hourly radio news reports held the nation
captive for days. Newspaper articles documenting the
efforts of his rescuers won a young journalist, Skeets
Miller, a Pulitzer Prize.
In 1925 few individuals practiced caving as a sport
and recreation in caves was a novel idea. As a fairly
new sport, experienced cavers were rare and few could
be found to help at Sand Cave. No one at the rescue
had been trained for cave rescue emergencies and they
lacked the specialized equipment to save Floyd's life.
The rescue team had no idea how to help Floyd
remove the rock that pinned him. Miners, farmers,
relatives, and hundreds of other people came to the
scene and tried to help. Eventually there were so many
bystanders that the rescue took on a circus-like appear¬
ance, complete with hot-dog stands and vendors sell¬
ing souvenirs. After days of entrapment and bitter
cold, Floyd died, imprisoned by a 26-pound rock.
HISTORY OF RESCUE
Since 1925, cave exploration has increased in popu¬
larity and frequency. Caving has become one of the
"new outdoor sports." With this increase, people of
many backgrounds have become attracted to caves.
They come for adventure, for recreation, and for sci¬
ence. Unfortunately, not all visitors come with the
proper equipment or training. It is often these poorly
prepared "spelunkers" that rescuers are called to help.
Currently almost 90% of cave rescue emergencies
result from poorly prepared or inexperienced people
entering caves. As the number of people entering caves
has increased, so has the need for rescue.
Like any sport, people can be injured in various
ways while exploring caves, and many rescues require
rescuers to possess special skills to reach and treat the
caver. However, most missions are initiated for simple
problems, such as a light failing or the group lost in an
unfamiliar area of the cave.
Cave rescues take many forms. The simplest rescue
mission might involve locating a lost person, and a
more complex rescue could involve negotiating multi¬
ple vertical drops or water problems while carrying an
Encyclopedia of Caves.
© 2012 Elsevier Inc. All rights reserved.
RESCUES
649
injured patient. These missions may involve as many
as 200 rescuers.
The National Speleological Society (NSS) first began
recording cave rescue information around 1950 and
each year since has published the rescue information
gathered. Currently, these accident reports are collected
in a single yearly publication entitled American Caving
Accidents.
Additional record keeping is done by the National
Cave Rescue Commission (NCRC) and by individual
regions of the NSS. One example is the eastern region
of the NCRC, which has maintained its regional mis¬
sion reports on file, on the web, and available for
research. These are available on the eastern region
website or by accessing its files. The NCRC records are
compiled each year as part of the publication American
Caving Accidents. These data banks have made it possi¬
ble to take a comprehensive look at the growth of cave
rescue in America. By studying mission reports com¬
piled over the past 30 years, rescuers can better pre¬
pare for future rescue situations and determine
training needs for an area.
DEVELOPMENT OF RESCUE TEAMS
In order to better understand the growth of cave
rescue in the United States, the historical records were
categorized and divided into several time periods. The
mission types, frequency, and occurrence showed that
there was an almost proportional increase in the num¬
ber of rescue callouts and the growth of the caving
population as determined by increases in NSS mem¬
bership. From this, one can infer that the growth of the
NSS reflects an overall increase in total hours people
spend sport caving each year. As more people take up
caving, more accidents occur.
The increase in total accidents graphs as a parallel
linear trend when compared to NSS growth but when
accidents are graphed by type, the growth trend shows
several sudden peaks followed by a gradual return to
a more linear trend. An example of this was the sharp
increase in the vertical and water rescue accident rates
when new equipment and skills were introduced in
the 1970s. It is theorized that this surge in vertical and
water missions reflects a lack of training or experience
in using the new equipment, resulting in higher acci¬
dent rates. As training and experience were gained,
the rate declined and eventually returned to a more
normal growth line.
As accidents increased over the 50 years of study,
more rescuers were needed to help on rescue missions.
This resulted in the formation of local cave rescue
teams. During the 1960s and 1970s, teams such as the
National Capital Cave Rescue Team and the Hondo
Underground Rescue Team (HURT) responded to res¬
cues all over the country. These teams were soon fea¬
tured in men's magazines such as Saga, which ran an
article entitled "Get the Heroes from HURT." The
backlash from all of this publicity set cave rescue team
training back a decade.
IN THE BEGINNING
Cave rescue teams were a relatively new concept in
the United States during the 1950s. Organized training
in cave rescue was done informally by local clubs but
not on a national scale. Before 1976, no structured res¬
cue classes were available to cavers. Because of this,
cavers learned rescue skills by doing rescues.
From training done by early clubs, the first true
American cave rescue teams emerged. The earliest
teams were formed in response to accidents occurring
on club trips or when club members were nearby.
During the early years an injured caver often waited
for hours as one of the established cave rescue teams
was summoned to the site. This often took many
hours, resulting in a number of body recoveries
instead of rescues.
Cavers recognized the need for more formal train¬
ing. Grottos (local NSS chapters) began training and
new teams were formed around experienced leaders.
As these teams gained in mission experience, they
began to amass information on which techniques
worked and which did not.
In the 1960s. the NSS formed the first Cave Rescue
Commission and began planning for a national
policy on cave rescue. The first commissioners met at
NSS conventions and began the first formal exchange
of information between teams. This effort continued
with several false starts and slowdowns until 1976
when the first formal certification in cave rescue was
offered by the West Virginia Fire Extension Service.
This class, taught by NSS instructors, was the begin¬
ning of the society's first organized effort to train cave
rescuers.
Since 1977, the NCRC of the NSS has taken the lead
in cave rescue training throughout the United States
and the Caribbean. Standards for cave rescue devel¬
oped by the NCRC are derived from mission experi¬
ences, mission data analysis, and current rescue
training standards.
LEARNING FROM THE PAST
Mission records reveal some interesting facts about
the types of missions each decade spawned, including
how training and experience influence the outcome of
ENCYCLOPEDIA OF CAVES
650
R
many missions. As the years progressed, better trained
teams succeeded in saving lives more often than
untrained teams and experienced fewer injuries to
team members.
Accident Frequency Rate during the 1950s
National Speleological Society cavers in the United
States spent a maximum of 1,440,000 hours caving per
year. There were 13 rescue missions or incidents
reported during those 1,440,000 hours of caving in the
1950s. For NSS caving this translates to one accident
per roughly 111,000 hours underground during the
decade.
Accident Frequency Rate during the 1960s
Caving in the 1960s started off with a bang. New
rappelling systems, such as brake bars and carabiners,
replaced the body rappel for descending. Prusik knots
were replaced by new and faster jumars. Ascending
devices and improved cable ladders opened previ¬
ously unexplored passages to a new generation of
cavers. All over the country cavers were off to push
greater depths and longer drops. With this quest to
push the limits, rescues became harder and more
frequent.
By the beginning of the 1960s, the NSS was record¬
ing several rescues per year. As a result the NSS
became more interested in defining its rescue capabili¬
ties. In 1960 the NSS polled its members to determine
who had rescue experience. The survey found that
only 81 members nationwide had cave rescue experi¬
ence, about 8% of the active members of the NSS at
that time.
Accident Frequency Rate during the 1970s
From 1970 to 1980 there were 198 incidents
reported. Caving was getting safer but people were
pushing harder and accidents were increasing. The
NSS had about 3200 members and cavers had one acci¬
dent per 23,272 hours underground. For the first time
chronic medical problems became a factor and rockfall
turned up as a cause of 8% of the rescue missions.
The 1970s saw additional improvements in equip¬
ment and technique. Descending devices now included
racks and figure eights in addition to brake bars and
carabiners. The introduction of better equipment was
not complete until braided nylon ropes appeared
and Charlie Gibbs introduced his Gibbs ascending sys¬
tem at the 1969 NSS Convention. During this period
southeastern cavers were experimenting with racks for
descending long drops and with new abrasion-resis¬
tant rope designs. Vertical caving took a giant leap for¬
ward and vertical incidents took a leap upward. The
percentage of vertical missions dropped back from
42% in the 1960s, but the number of vertical problems
increased.
Accident Frequency Rate during the 1980s
The 1980s found cavers well equipped to push the
hardest caves and "world class caving" was intro¬
duced. Cavers took the new equipment, mastered it,
redesigned it, and even manufactured it in order to
explore the world's largest caves. With equipment
mastery and vertical skills training a part of every
caver gathering, the vertical accident rate started to
trend back toward the linear caver-population growth
curve.
From 1980 to 1990, accidents continued to increase
as did membership in the NSS. However, vertical acci¬
dent rates began to drop. This decade saw 300 inci¬
dents with an NSS membership of 5200. The calculated
rate of accidents per hour underground remained near
that of the previous decade with one accident occur¬
ring per 24,960 hours of caving. Analysis of incidents
indicated that for the first time vertical accidents were
not the leading type. Water-related accidents sur¬
passed vertical accidents this decade because sport
diving became more popular and the great cave
springs claimed many lives. Chronic medical problems
doubled in response to the aging of the caving popula¬
tions and the availability of treatments that allowed
persons with chronic medical problems to become
more active.
THE PRESENT
The years from the 1990s and into the twenty-first
century were ones of consolidation. The NCRC spon¬
sored cave rescue training courses at beginner and
advanced levels in most parts of the country at regular
intervals. Rescue certification was well established.
Many involved with rescue had advanced from basic
first aid to emergency medical technician (EMT) and
paramedic status. Most caving groups have a local
cave rescue committee and relations with other search-
and-rescue organizations such as police and fire
departments have greatly improved. For the unfortu¬
nate victim of a caving accident this means a shorter
wait for rescue personnel to arrive, better medical
training when they do arrive, and an effective organi¬
zational structure that deals with everything from
proper packaging and transport of the victim from the
cave to news conferences with the media.
Most countries with active caving groups have found
it necessary to organize formal rescue teams. These
groups have interacted and exchanged information
ENCYCLOPEDIA OF CAVES
RESPONSES TO LOW OXYGEN
651
through the International Congresses of Speleology
which are held every four years.
See Also the Following Article
Exploration of Caves—General
Bibliography
Hempel, J. C., & Fregeau-Conover, A. (2001). On call. Huntsville, AL:
National Speleological Society.
RESPONSES TO LOW OXYGEN
Frederic Hervant and Florian M alard
UMR CNRS 5023, Laboratoire d’Ecologie des
Hydrosystemes Naturels et Anthropises, Universite Lyon 1
INTRODUCTION
Before the late 1970s, ecological studies were carried
out mainly in the unsaturated zone of karst aquifers,
particularly in caves. Because cave water bodies are
exposed to the atmosphere, they are usually saturated
with oxygen, dissolved oxygen (DO) was not measured
routinely, and oxygen availability in subterranean bio¬
topes was rarely considered as a key ecological factor
that governed the occurrence and spatiotemporal dis¬
tribution of hypogean animals. Biological activity, ani¬
mal density, and organic matter content in the
groundwater were assumed to be too low to induce
oxygen deficiency; therefore, the possibility that sub¬
terranean organisms may have to face hypoxic stress
was not considered. Ecophysiological studies essen¬
tially concerned the adaptive responses of hypogean
animals to low food supply including their higher
food-finding ability, strong resistance to starvation,
and reduced metabolism (Hervant and Renault, 2002;
Issartel et al., 2010; Salin et al., 2010).
Since the 1970s, this view has been reevaluated by
many researchers who have frequently reported low
dissolved oxygen concentrations from shallow ground-
water in unconsolidated sediments, particularly in the
hyporheic zone of streams; the groundwater environ¬
ment has been described as being hypoxic or weakly
oxygenated. Also, several authors have conducted lab¬
oratory studies to test the resistance and adaptive
strategies of animals in response to low oxygen
(Malard and Hervant, 1999; Issartel et al., 2009).
The ensuing material begins by examining the oxy¬
gen status of different groundwater systems including
deep and shallow-water-table aquifers and the hypor¬
heic zone of streams. Then, based on laboratory
studies conducted by one of the authors (Hervant
et al., 1996, 1997, 1998; Malard and Hervant, 1999;
Issartel et al., 2009), we examine the behavioral, respi¬
ratory, and metabolic responses of groundwater organ¬
isms, especially crustaceans and amphibians, to low
oxygen concentrations. Finally, we suggest that the
selection of organic-matter-rich habitats in groundwa¬
ter increases the risk of facing hypoxic stress.
SUPPLY AND CONSUMPTION OF
DISSOLVED OXYGEN IN GROUNDWATER
Because of permanent darkness, there is no photo¬
synthesis in groundwater, thus no production of oxy¬
gen; therefore, the oxygen status of groundwater is
determined by the rate of oxygen transport from
the surface environment and by the rate of oxygen
consumption in the subsurface. Replenishment of dis¬
solved oxygen occurs by air diffusion from the unsatu¬
rated zone or by recharge with rainwater or river water.
Fluctuations of the groundwater table enhance air
entrapment, thereby increasing dissolved oxygen trans¬
fer from entrapped air. Oxygen transport within
groundwater may occur as a result of oxygen diffusion,
convection currents caused by heat transfer, and advec-
tion of water in response to hydraulic gradients. The
diffusion of oxygen in water is negligible, and convec¬
tive currents are limited in groundwater by sediment
and weak thermal gradients (the geothermal gradient is
usually about 0.01°C m -1 ) (Malard and Hervant, 1999);
therefore, oxygen transport in groundwater is primarily
due to advective movement of water in response to
hydraulic gradients. Because groundwater velocity is
usually low ( i.e., 10 6 —10 _4 m s -1 ), the available flux
of DO in groundwater is much lower than in surface
waters. On the other hand, oxygen consumption
by microorganisms is limited in many aquifers by
the availability of biodegradable organic carbon.
Consequently, DO may persist at considerable distances
from the recharge zone in deep-water-table aquifers
where soil-generated dissolved organic carbon is
completely degraded during the transit of infiltrating
water in the unsaturated zone. In confined aquifers of
the Ash Meadows basin (south central Nevada),
Winograd and Robertson (1982) sampled ground-
water with 2 mg L -1 O 2 at a distance of 80 km from
the recharge zone. In contrast, dissolved oxygen may
be totally consumed over very short distances {i.e., a
few meters or even centimeters) in shallow-water-
table aquifers or in the hyporheic zone of rivers because
of the input of soil- or river-labile dissolved organic car¬
bon. Malard and Hervant (1999) reported strong vari¬
ability among groundwater systems in the length of
underground pathways for dissolved oxygen. Based on
Encyclopedia of Caves.
© 2012 Elsevier Inc. All rights reserved.
652
R
cross-system comparison from literature data, these
authors suggested that differences among hyporheic
zones reflect variation in the contact time of water with
sediment, whereas differences among confined aquifers
are primarily a result of differences in the rate of DO
consumption (see also Foulquier et al., 2010).
SMALL-SCALE HETEROGENEITY IN
DISSOLVED OXYGEN IN GROUNDWATER
Small-scale investigations of oxygen distributions in
subsurface waters revealed strong variations over
distances of a few centimeters or meters. This hetero¬
geneity, an essential feature of the groundwater envi¬
ronment, was observed in a number of subsurface
water habitats, including the saturated zone of karst
aquifers, the water-table region of deep- and shallow-
water-table porous aquifers, the halocline of anchiha-
line caves, the hyporheic zone of rivers, and the
interstitial environment of marine and freshwater bea¬
ches (Malard and Hervant, 1999) (Fig. 1). Small-scale
spatial heterogeneity in DO reflects changes in sedi¬
ment composition and structure, subsurface water
flow velocity, strength of hydrological exchanges with
the surface environment, dissolved and particulate
organic matter content, and activity of microorgan¬
isms. Strong temporal changes in DO may also occur
in the hyporheic zone of streams as well as in the
recharge zone of aquifers, but these fluctuations are
strongly damped with increasing distance from the
stream and the recharge zone. Whereas diminished
30-m-deep
water table
(B) 0 2 (mg/L- 1 )
0 12 3 4
FIGURE 1 Meso- (m) and micro- (cm) scale het¬
erogeneity in dissolved oxygen concentrations in:
(A) a deep-water-table aquifer; (B) a shallow-
water-table aquifer; (C) an anchihaline cave; (D) the
interstitial environment of a sandy marine beach;
and (E) the hyporheic zone of a backwater (Malard
and Hervant, 1999, and references therein.).
0 0.2 0.4 H 2 S |mM]
Stream length (10 m)
ENCYCLOPEDIA OF CAVES
RESPONSES TO LOW OXYGEN
653
oxygen concentration is typically not a rule for the
groundwater environment, the high spatial heterogene¬
ity of DO at meso- (meter) and micro- (centimeter)
scales is considered a peculiarity of groundwater habi¬
tats (Malard and Hervant, 1999). This implies that ani¬
mals living in groundwater have to experience highly
variable oxygen concentrations as they move through a
mosaic of patches with contrasted DO concentrations.
FIELD EVIDENCES OF SURVIVAL IN LOW
OXYGENATED GROUNDWATER
Results of field studies reveal that most animals can
be found living in a wide range of DO, even anoxia in
some cases (Malard and Hervant, 1999). The blind sal¬
amander Proteus anguinus, the only European verte¬
brate that lives exclusively in caves, has to frequently
cope with low DO (Issartel et al., 2010, and references
therein). Based on 700 faunal samples collected in the
hyporheic zone of several desert streams in Arizona,
Boulton et al. (1992) examined the field tolerance to
DO of 23 common invertebrate taxa. These authors
showed that most taxa could be encountered in sub¬
surface waters with less than 1 mg L _1 0 2 . Strayer et al.
(1997) obtained similar results based on 167 samples of
hyporheic invertebrates collected at 14 sites in the east¬
ern United States; however, in both studies, several
animal groups, particularly crustaceans and insects,
occurred more frequently in well-oxygenated sedi¬
ments than where oxygen was scarce. Several crusta¬
ceans (including the amphipod Niphargus hebereri
Schellenberg, the thermosbaenacean Monodella argen-
tarii Stella, and some species of remipedes) are known
to develop dense populations in the sulfide zone of
anchihaline caves (Malard and Hervant, 1999). It is not
yet clear, however, whether these crustaceans perma¬
nently live in deoxygenated waters or if they tempo¬
rarily seek shelter in a more aerated environment.
A diversified aquatic fauna was also found to live in
the uppermost hypoxic layer (DO <0.3 mg L _1 0 2 )
of sulfidic groundwater in Movile Cave, Romania
(Malard and Hervant, 1999). On the other hand, the
paucity of fauna in extensive areas of hypoxic ground-
water suggests that most hypogean taxa are probably
not able to survive a lack of oxygen for very long
(more than 3—4 days). Reports of groundwater animals
among poorly oxygenated groundwaters are equivocal
because their presence may be due to either tolerance
to low oxygen concentrations by hypogean inverte¬
brates or the existence of microzones of high DO.
Recent detailed laboratory studies on the ecophysio-
logical (i.e., behavioral, respiratory, and metabolic)
responses of several subterranean crustaceans and
amphibians to anoxia have enabled us to define more
precisely the degree of tolerance of groundwater ani¬
mals to low oxygen concentrations and to elucidate
some of the mechanisms responsible for extended sur¬
vival under oxygen stress (Hervant et al., 1996, 1997,
1998; Issartel et al., 2009).
SURVIVAL TIMES UNDER ANOXIA
Most data available on adaptations to low oxygen
in groundwater organisms arise from comparisons of
the adaptive responses of three hypogean crustaceans
(the amphipods Niphargus virei Chevreux and
Niphargus rhenorhodanensis Schellenberg and the iso¬
pod Stenasellus virei Dolfus) and two epigean crusta¬
ceans (the amphipod Gammarus fossarum Koch and
the isopod Asellus aquaticus L.) to anoxia (Hervant
et al., 1996, 1997). The responses of these five crusta¬
ceans were examined in darkness (at 11°C) under
three different experimental conditions: anoxia,
normoxia following an anoxic stress (postanoxic
recovery), and declining DO concentration (from 10.3
to 0.7 mg L -1 0 2 ). Information gained from these
experiments concerned (1) the lethal time for 50% of
the population (LT 50%) in anoxia; (2) the locomotor
activity (number of periods of locomotion per minute)
and the ventilation activity (number of pleopod beats
per minute) of animals; (3) the oxygen consumption
rates in normoxia and under declining PC) 2 (in order
to determine the critical partial pressure of oxygen.
Pc); and (4) changes in the concentrations of key
metabolites such as high energy compounds (e.g.,
adenosine triphosphate [ATP], phosphagen), anaero¬
bic substrates (glucose, glycogen, amino acids), and
anaerobic end products (lactate, alanine, succinate,
malate, propionate).
Groundwater crustaceans show survival times (LT
50%) of about 2 to 3 days under anoxia (Table 1) and
several months in moderate hypoxia (Malard and
Hervant, 1999). They are much more resistant to oxy¬
gen deprivation than morphologically closed epigean
species whose LT 50% values range from a few hours
to one day. Hypogean fishes, crayfishes, and urodel
amphibians also display much higher survival times
than their related epigean species. For example, as far
as we know, the cave-dwelling urodel P. anguinus is
the most anoxia-tolerant amphibian (exhibiting LT 50%
in anoxia of about 12 h at 12°C) and one of the most
anoxia-tolerant vertebrates. At high temperatures, only
the goldfish Carassius auratus presents a higher half-
lethal time (22 h at 20°C; Issartel et al., 2009). Proteus
anguinus could thus represent an excellent model
organism for the study of key mechanisms in anoxia
tolerance.
ENCYCLOPEDIA OF CAVES
654
R
TABLE 1 Comparison of Locomotory and Ventilatory Activities, Oxygen
Consumption (SMR), and Intermediary and Energetic
Metabolism during Normoxia, Anoxia and Postanoxic Recovery
in Two Surface-Dwelling and Three Groundwater Crustaceans
Epigean
Hypogean
G. fossarum A. aquaticus
N. rhenorho.
N. virei
S. virei
Survival during anoxic stress
• Lethal time for 50% of the population
6.3 h
19.7 h
46.7 h
52.1 h
61.7 h
Activity
• Number of periods of locomotion per minute (in normoxia)
19.3
6.7
12.5
1.4
4.5
• Changes during anoxia
X
X
X
X
X
Ventilation
• Number of pleopod beats per minute (in normoxia)
60
43
44
48
38
• Changes during anoxia
X
X
X
X
X
Respiratory metabolism
• Normoxic O 2 consumption (SMR; /.tL 0 2 /g dw/h)
940
1325
605
305
425
• 0 2 critical pressure (mg L _1 )
3.6
3.8
2.0
1.8
2.0
Body energy stores (in normoxia)
• Stored glycogen (pmol/g dw)
110
165
356
243
307
• Stored arginine phosphate (pmol/g dw)
7.0
10.5
26.5
26.5
31.5
Post-anoxic recovery
• Glycogen re-synthesis rate (p mol/g dw/min)
0.003
0.01
0.07
0.03
0.04
• % of glycogen re-synthesis
8
19
47
49
53
After Hervant et al., 1996, 1997,1998; Malard and Hervant, 1999.
BEHAVIORAL RESPONSES
Every spontaneous or stress-induced activity increases
energetic expenditures and, therefore, oxygen consump¬
tion. Some hypogean species show a very low activity
rate when compared to morphologically closed epigean
animals (Table 1). This reduced activity has been inter¬
preted to be a result of (1) food scarcity, stable envi¬
ronmental conditions, and general lack of predators
and territorial behavior; (2) cramped interstitial habitats;
and/or (3) low oxygen availability (Hervant et al., 1996;
Hervant and Renault, 2002; Salin et al., 2010). A low activ¬
ity rate in groundwater animals which is often associated
with an increase in food-finding ability is an efficient
adaptation to low food supply because it results in a
reduced energy demand.
Epigean Gamviarus fossarum and Asellus aquations
respond to experimental anoxia by a marked hyperac¬
tivity (Table 1) that, in natural conditions, probably cor¬
responds to an attempt to move to more oxygenated
habitats {i.e., to a behavioral compensation). In con¬
trast, all three hypogean crustaceans show a drastic
reduction in locomotory activity under lack of oxygen
(Table 1). This adaptive behavior reduces energy
expenditure during oxygen deprivation (by decreas¬
ing the oxygen consumption) and therefore increases
survival time in deoxygenated groundwater. This
"sit-and-wait" strategy, very advantageous under
hypoxic (and/or food limited) environments, was also
observed in starved stygobite crustaceans (Hervant
and Renault, 2002) and cave salamanders (Hervant
et al., 2001). The hypogean isopod Proasellus slavus
actively explored its interstitial environment but
slowed down exploratory movement under severe
hypoxic conditions {i.e., 0.1 mg L 1 O 2 ). The lack of an
escape behavior is probably not universal for
hypogean aquatic organisms, however, because active
animal migration in response to fluctuating DO con¬
centrations has already been documented in natural
biotopes and in a laboratory flume.
ENCYCLOPEDIA OF CAVES
RESPONSES TO LOW OXYGEN
655
RESPIRATORY RESPONSES
Subterranean animals generally have reduced stan¬
dard metabolic rates (SMRs) and reduced routine metabolic
rates (RMRs) when compared to their surface-dwelling
counterparts (Hervant et ah, 1998; Issartel et al., 2010).
For example, the cave salamander Proteus anguinus
Laurenti (Hervant et ah, 2001), the hypogean fish
Astyanax fasciatus mexicanus (Salin et ah, 2010), and the
groundwater crustaceans Niphargus virei, Niphargus rhe-
norhodanensis, and Stenasellus virei show SMRs in nor-
moxia from 1.6 to 4.5 times lower than their epigean
relatives (Table 1) (Hervant et ah, 1998). Low SMRs and
RMRs among hypogean animals have long been inter¬
preted only as an adaptation to low food availability;
however, reduced SMRs in some populations of subter¬
ranean animals may also reflect an adaptation to low
oxygen concentrations in the groundwater (Hervant and
Renault, 2002). A reduced metabolic rate (i.e., low ener¬
getic requirements) results in a lower oxygen removal
rate and an increased survival time in anoxia.
The hypogean amphipods N. virei and N. rhenorho-
danensis and the hypogean isopod S. virei reduce their
ventilatory activity during anoxic stress (Table 1),
thereby limiting their energy expenditure. Several ani¬
mals contract an oxygen debt during oxygen depriva¬
tion that is repaid on return to normoxia (Hervant
et ah, 1998). The repayment of this debt during recov¬
ery from anaerobic stress involves a significant
increase in SMR. The oxygen debt of hypogean N. virei
and N. rhenorhodanensis is 2.2 to 5.3 times lower than
that of epigean G. fossarum and A. aquaticus (Hervant
et ah, 1998). This lower oxygen debt indicates a
reduced energetic expenditure (i.e., an energy sparing)
which is probably linked to lower locomotory and ven¬
tilatory activities during anoxia.
The resistance to anoxia (and/or long-term fasting)
displayed by numerous hypogean species may be
explained by their ability to remain in a prolonged state
of torpor. This state enables hypogean organisms to tol¬
erate a prolonged reduction in oxygen (and/or food)
availability by maximizing the time during which metab¬
olism can be fuelled by a given energy reserve (or a
given food ration). This supports the classic suggestion
that difficulties in obtaining food in stressful environ¬
ments may select for conservative energy use (Hervant
and Renault, 2002). These adaptive responses may be
considered for numerous subterranean organisms as an
efficient energy-saving strategy in a harsh and unpre¬
dictable environment where hypoxic (and/or starvation)
periods of variable duration alternate with normoxic per¬
iods (and/or sporadic feeding events). Hypoxia-tolerant
(and/or food-limited) groundwater species appear to be
good examples of animals representing a low-energy
system (Hervant et ah, 2001; Hervant and Renault, 2002),
although starvation tolerance is not universal for cave
organisms (Salin et ah, 2010).
Theoretically, aerobic organisms can be described as
metabolic conformers or regulators if their oxygen con¬
sumption varies directly with or is independent of PO 2 ,
respectively; however, these two types are merely the
two ends of a large spectrum of respiratory responses
of species. Animals never fully belong to one or the
other type, and below a critical PO 2 (Pc) a regulator
becomes a conformer (Hervant et ah, 1998); therefore.
Pc is a good indicator of the tolerance or adaptation of
an organism to low P0 2 . Gammarus fossarum, A. aquati¬
cus, N. virei, S. virei, and N. rhenorhodanensis are able to
maintain a relatively constant rate of oxygen consump¬
tion that is independent of PO 2 and between normoxia
and Pc (Hervant et ah, 1998). The blind salamander
P. anguinus presented the same respiratory pattern
(Hervant, unpublished data). However, the PO 2 at
which respiratory independence is lost is significantly
lower for hypogean species than for surface-dwelling
ones (Table 1). This implies that groundwater species
are able to maintain an aerobic metabolism for a longer
time in declining PO 2 (progressive hypoxia) instead of
partly switching to a low-energy anaerobic metabolism.
The maintenance of an aerobic metabolism (i.e., sur¬
vival at a lower energetic cost) under hypoxic condi¬
tions is partly due to the lower SMR of hypogean
animals. The occurrence of respiratory regulation over
such a wide range of DO concentrations had, until
then, not been observed among epigean crustaceans
and had rarely been observed in other invertebrates
(Malard and Hervant, 1999, and references therein).
METABOLIC RESPONSES^
During Anoxia
A number of biochemical adaptations that permit
extended survival under prolonged hypoxia or anoxia
have been identified in various well-adapted epigean
groups, especially marine annelids and intertidal mol-
lusks. These include the maintenance of high reserves of
fermentable fuels (such as glycogen and amino acids) in
some tissues under normoxic conditions, the use of
anaerobic pathways to enhance ATP yield and to main¬
tain redox balance during low-oxygen conditions, and
mechanisms for minimizing metabolic acidosis often
associated with anaerobic metabolism. Nevertheless,
epigean crustaceans have been recognized as being of a
"modest anaerobic capacity" without special and effi¬
cient mechanisms of anaerobic metabolism (Hervant
et ah, 1996, and references therein).
ENCYCLOPEDIA OF CAVES
656
R
Examination of biochemical (i.e., intermediary and
energy metabolism) responses of hypogean crustaceans
N. virei, N. rhenorhodanensis, and S. virei during anoxic
stress show that, similarly to epigean crustaceans,
anaerobic metabolism does not lead to a high ATP pro¬
duction rate (Hervant et ah, 1996,1997). The five crusta¬
ceans studied respond to severe experimental anoxia
with a slight improvement of a classical anaerobic
metabolism that is characterized by a decrease in ATP
and phosphagen (arginine phosphate, representing an
immediate source of ATP), a coupled utilization of gly¬
cogen and some amino acids (mainly glutamate), and
the accumulation of lactate and alanine as end pro¬
ducts. The only difference is that both Niphargus spe¬
cies also accumulate a low proportion of succinate,
which slightly enhances ATP yield during anaerobic
respiration. Lactate is largely excreted by all five crusta¬
ceans. This excretion, which is unusual for crustaceans,
can be considered a simple way to fight against meta¬
bolic acidosis linked to anaerobic end-product (includ¬
ing H + ) accumulations (Hervant et ah, 1996).
There is a striking difference in the respective
amounts of glycogen and phosphagen stored by epigean
and hypogean crustaceans. Glycogen body reserves
(Table 1) are 1.5 to 3.2 times higher in the three hypo¬
gean crustaceans studied than in the epigean G. fossarum
and A. aquaticus (Hervant et ah, 1996, 1997) but are also
higher than those reported for all surface-dwelling crus¬
taceans, even those most tolerant of anoxia or hypoxia
(Malard and Hervant, 1999). High amounts of ferment¬
able fuels result in a more sustained supply for anaero¬
bic metabolism, thereby increasing survival time during
oxygen deprivation. Moreover, glycogen utilization rates
and lactate production rates are significantly lower in
hypogean crustaceans. This finding is probably linked
to lower SMRs and to the reduction of locomotory and
ventilatory activity during anoxia (Table 1) (Hervant
et ah, 1996,1998; Issartel et ah, 2009).
During a Postanoxic Recovery Phase
It is ecologically very important for organisms to
recover quickly and completely from hypoxic or anoxic
stress when oxygen is available once more. This recovery
implies a restoration of high energy compounds (mainly
ATP, phosphagen, and glycogen), as well as disposal of
anaerobic end products (mainly lactate, alanine, and suc¬
cinate). End products can be disposed of in three differ¬
ent ways during a postanoxic recovery phase: complete
oxidation, conversion back into storage products such as
glycogen (via the glyconeogenesis pathway, such as gly¬
cogen de novo synthesis from lactate, amino acids, and/
or glycerol), and excretion into the medium (Hervant
et ah, 1996). Excretion is an important mechanism for the
disposal of lactate during aerobic recovery in the epigean
G. fossarum and A. aquaticus. This is a costly strategy
because it implies a loss of energy-rich carbon chains. In
contrast, hypogean crustaceans preferentially use glyco¬
neogenesis to convert lactate into glycogen stores. The
existence of glyconeogenesis has already been demon¬
strated in several crustaceans, although the organ sites
for this metabolic pathway have not been identified
clearly. Experiments using injections of labeled glucose
and lactate (see Malard and Hervant, 1999) revealed that
the gluconeogenesis rate in N. virei during postanoxic
recovery was higher than any rate measured previously
for epigean crustaceans. Glycogen reserve restoration
was indeed 2.5 to 6.6 times greater in hypogean species
than in G. fossarum and A. aquaticus (Table 1). This ability
to quickly resynthesize during recovery periods the
body stores depleted during lack of oxygen allows
groundwater organisms to successfully fuel an ensuing
hypoxic or anoxic period; therefore, groundwater species
are well adapted to live in habitats showing frequent and
unpredictable alternations of normoxic and hypoxic/
anoxic phases.
In addition to the immediate harmful effect of
anaerobiosis, anoxia-tolerant animals must also cope
with postanoxic damage that occurs during reoxygena¬
tion: a burst of reactive oxygen species (ROS) occurs
during the recovery of activity of the mitochondrial
respiratory chain. Thus, because of the damaging effects
of ROS overproduction, animals that experience anoxia
have developed both enzymatic and non-enzymatic
antioxidant mechanisms (Issartel et ah, 2009, and
references therein). Although we demonstrated excep¬
tional anoxia tolerance in the cave-dwelling urodel
P. anguinus, this species paradoxically does not present
the "ordinary" antioxidant responses (showing no
increase in liver antioxidant enzyme activities and a
general low control of ROS damage; Issartel et ah, 2009).
These surprising results open new research perspectives
in the comprehension of anoxia tolerance.
IMBRICATION OF ADAPTIVE
STRATEGIES TO LOW FOOD RESOURCES
AND LOW OXYGEN SUPPLY
The occurrence of adaptive strategies in response to
low oxygen among animals living in an oligotrophic
environment may seem paradoxical. Notwithstanding
the fact that the supply of organic matter is typically
lower in groundwater than in surface water, field
ecological studies carried out for the past 30 years
have shown that several groundwater habitats (partic¬
ularly unconsolidated sediments) have reduced DO
concentrations. Although a few subterranean habitats
are known to be organic-matter rich (Salin et ah, 2010),
low DO concentration in many groundwater habitats
ENCYCLOPEDIA OF CAVES
RESPONSES TO LOW OXYGEN
657
is most likely attributable to the lack of oxygen pro¬
duction and low transport rate of DO than to elevated
concentrations of organic matter. If food availability
drives habitat selection in groundwater, hypogean ani¬
mal populations would preferentially occur in ground-
water biotopes receiving higher fluxes of organic
matter from the surface environment. These habitats
are also more likely to exhibit reduced DO concentra¬
tions because of increased respiration rates associated
with the input of organic matter. Thus, the selection of
habitats having increased food supply increases the
probability of facing hypoxic stress among hypogean
animals. Meantime, the development of behavioral,
respiratory, and metabolic strategies to low food sup¬
plies also selects for higher resistance to a low oxygen
supply. Clearly, the role of food availability and the
significance of low oxygen supply in determining the
development of adaptive strategies and distribution
patterns of animals in groundwater are overlapping
aspects that can hardly be treated independently.
Some metabolic pathways, therefore, that are specifi¬
cally linked to the response to long-term starvation
and/or adaptation to oxygen stress are associated with
energy-limited subterranean organisms. Hervant and
Renault (2002) and then Issartel et al. (2010) demon¬
strated that the groundwater crustacean S. virei and a
cave population of the urodel Calotriton asper preferen¬
tially utilizes lipids during food shortage, in order to
(1) save carbohydrates and phosphagen, the two main
fuels metabolized during oxygen deficiency, and (2) save
proteins (and therefore muscular mass) for as long as
possible. Thus, this species can (1) successfully withstand
a hypoxic period subsequent to (or associated with) an
initial nutritional stress, and (2) rapidly resume searching
for food during short-term, sporadic, nutrition events.
A general adaptation model for groundwater ani¬
mals involves their ability to withstand prolonged hyp¬
oxia and/or long-term starvation (Hervant et al., 2001;
Hervant and Renault, 2002) and to utilize in a very
efficient way the high-energy body stores. Because
the three hypogean crustaceans studied lack a high-
ATP-yielding anaerobic pathway (such as observed in
permanent anaerobic organisms), their higher survival
time in anoxia is mainly due to the combination of four
mechanisms: (1) high storage of fermentable fuels (gly¬
cogen and phosphagen); (2) low SMR in normoxia;
(3) further reduction in metabolic rate by lowering ener¬
getic expenditures linked to locomotion and ventilation
during hypoxia; and (4) high ability to resynthesize the
depleted body stores during subsequent recovery peri¬
ods. The ability to maintain and rapidly restore (with¬
out feeding) high amounts of fermentable fuels for use
during lack of oxygen can be considered an adaptation
to life in a patchy environment. Through their efficient
exploratory behavior in a moving mosaic of patches of
low and high DO concentration (Malard and Hervant,
1999), numerous groundwater animals probably experi¬
ence highly variable DO and/or food concentrations.
The behavioral, physiological, and metabolic responses
of numerous hypogean animals partly explain why
they occur in groundwater systems with a wide range
of DO (Malard and Hervant, 1999).
A high resistance to lack of oxygen (and/or to food
deprivation) is not universally found in subterranean
organisms but is probably more related to oxygen
availability and/or to the energetic state of each sub¬
terranean ecosystem, as demonstrated by Salin et al.
(2010) with hypogean fish living in energy-rich tropi¬
cal caves. Indeed, groundwater ecosystems are far
more complex and diverse than earlier presumed.
The aquatic amphipod Gammarus minus (Hervant
et al., 1999) and the tropical fish Astyanax fasciatus
(Hervant, unpublished data) showed no significant
difference in ecophysiological responses to experi¬
mental anoxia and subsequent recovery among a sur¬
face and a cave population. The ability to withstand
and recover from long periods of nutritional and/or
oxygen stress is clearly a critical adaptation for the
colonization of energy-limited subterranean environ¬
ments, but troglomorphism is not necessarily linked
to starvation capacities and/or hypoxia tolerance in
some energy-rich caves. Despite a strong tendency
toward morphological and ecophysiological conver¬
gence, subterranean organisms do not form a homo¬
geneous group.
Bibliography
Boulton, A. Valett, H. M., & Fisher, S. G. (1992). Spatial distribu¬
tion and taxonomic composition of the hyporheos of several
Sonoran Desert streams. Archiv fur Hydrobiologie, 125, 37—61.
Foulquier, A., Malard, F., Mermillod-Blondin, F., Datry, T., Simon, L.,
Montuelle, B., et al. (2010). Vertical change in dissolved organic car¬
bon and oxygen at the water table region of an aquifer recharged
with stormwater: Biological uptake or mixing? Biogeochemistry, 99,
31-47.
Flervant, F., Mathieu, ]., Garin, D., & Freminet, A. (1996). Behavioral,
ventilatory, and metabolic responses of the hypogean amphipod
Niphargus virei and the epigean isopod Asellus aquaticus to severe
hypoxia and subsequent recovery. Physiological Zoology, 69,
1277-1300.
Hervant, F., Mathieu, ]., & Messana, G. (1997). Locomotory, ventila¬
tory and metabolic responses of the subterranean Stenasellus virei
(Crustacea: Isopoda) to severe hypoxia and subsequent recovery.
Compte Rendus de I'Academie des Sciences de Paris, Life Sciences, 320,
139-148.
Hervant, F., Mathieu, ]., & Messana, G. (1998). Oxygen consumption
and ventilation in declining oxygen tension and posthypoxic
recovery in epigean and hypogean aquatic crustaceans. Journal of
Crustacean Biology, 18(4), 717—727.
Hervant, F., Mathieu, ]., & Culver, D. C. (1999). Comparative
responses to severe hypoxia and subsequent recovery in closely
related amphipod populations ( Gammarus minus) from cave and
surface habitats. Hydrobiologia, 391, 97—204.
ENCYCLOPEDIA OF CAVES
658
R
Hervant, F., Mathieu, ]., & Durand, J. P. (2001). Behavioural, physio¬
logical and metabolic responses to long-term starvation and
refeeding in a blind cave-dwelling salamander ( Proteus anguinus )
and a facultative cave-dwelling newt ( Euproctus asper). Journal of
Experimental Biology, 204, 269—281.
Hervant, F., & Renault, D. (2002). Long-term fasting and realimenta¬
tion in hypogean and epigean isopods: A proposed adaptive
strategy for groundwater organisms. Journal of Experimental
Biology, 205, 2079-2087.
Issartel, J., Hervant, F., De Fraipont, M., Clobert, J., & Voituron, Y.
(2009). High anoxia tolerance in the subterranean salamander
Proteus anguinus without oxidative stress nor activation of antioxi¬
dant defenses during reoxygenation. Journal of Comparative
Physiology B, 179, 543—551.
Issartel, J., Voituron, Y., Guillaume, O., Clobert, J., & Hervant, F.
(2010). Selection of physiological and metabolic adaptations to
food deprivation in the Pyrenean newt Calotriton asper during
cave colonisation. Comparative and Biochemical Physiology A, 155(1),
77-83.
Malard, F., & Hervant, F. (1999). Oxygen supply and the adaptations
of animals in groundwater. Freshwater Biology, 41(1), 1—30.
Salin, K., Voituron, Y., Mourin, J., & Hervant, F. (2010). Cave coloni¬
zation without fasting capacities: An example with the fish
Astyanax fasciatus mexicanus. Comparative and Biochemical
Physiology A, 156(4), 451—457.
Strayer, D. L., May, S. E., Nielsen, P., Wollheim, V., & Hausam, S.
(1997). Oxygen, organic matter, and sediment granulometry as
controls on hyporheic animal communities. Archiv fur
Hydrobiologie, 140, 131—144.
Winograd, I., & Robertson, F. (1982). Deep oxygenated groundwater.
Anomaly or common occurrence? Science, 216(4551), 1227—1229.
ROOT COMMUNITIES IN LAVA TUBES
F red D. Stone, Francis Q. Howarth , 1 Hannelore Hoch , 1
and Manfred Asche
University of Hawaii at Hilo, ^Bernice E Bishop Museum,
+ Museum fur Naturkunde, Germany
Plant roots were not considered an important food
resource in cave ecosystems until the discovery by
Francis Flowarth of a planthopper and other cave-
adapted animals on tree roots in a lava tube in Hawaii
Volcanoes National Park, Howarth (1973). Furthermore,
very few troglobites (obligate cave-dwellers) had been
reported from lava tubes or from tropical caves. In the
subsequent decades, cave-adapted species have been
discovered in many areas of the tropics and in lava tubes
as well as in other suitable subterranean habitats.
Advances in knowledge of tropical cave communities
and of the Adaptive Shift hypothesis for cave species
evolution are described by Hoch and Howarth (1989,
1993) elsewhere in this encyclopedia. The potential for
discovery of new cave species in the tropics is great, as
only a tiny part of the potential underground habitat has
been studied to date (Fig. 1).
HABITAT AVAILABLE FOR TROGLOBITE
EVOLUTION
Lava tubes are a small subset of the total habitat
available to troglobites; other habitats include:
• Mesocaverns, which are small voids, about
0.5—25 cm in width, beneath the surface in many
kinds of substrates. These mesocaverns include
fractures, cracks, and vesicles in many kinds of
rock; cracks, animal burrows, and tree root holes in
soil; spaces between rocks in talus slopes or other
types of rock piles; spaces in rocks and gravel from
alluvial deposits or along stream beds; and solution
cavities. Substrates, in which the mesocaverns are
interconnected to form a vast anastomosing system,
can provide ideal habitat for troglobitic species. The
mesocavernous environment is most likely the
predominant habitat for cave species because it has
such a large potential area and is near enough to the
surface to contain abundant energy resources from
root penetration, migrating surface organisms, and
water-transported organic matter.
• Caves formed in limestone or other soluble rock are
generally the best studied cave communities because
of their attraction to cavers and their accessibility,
but most, if not all, obligate cave species in caves
also live in the surrounding mesocaverns. Limestone
caves may also contain important root communities,
particularly in areas near the surface or with ready
access to the water table.
• Tree root habitats are far more widespread than
those found in lava tubes. Roots can penetrate
through soil layers and enter the mesocavernous
habitat in many kinds of surface materials. Lava
tube root areas are particularly important because
they allow access to the underground habitat by
scientists, who otherwise have a difficult time
gaining access to the cave species. Lava tubes have
the additional advantage that they form in lava
flows that flow along the surface of the ground, so
they generally tend to be fairly shallow throughout
their length.
THE CAVE ENVIRONMENT IN
RELATION TO CAVE SPECIES
Energy Sources
Tree root communities gain their energy from pho¬
tosynthesis on the ground surface and therefore have
the same trophic structure as surface communities
with primary producers, herbivores, carnivores, detriti-
vores, decomposers, and fungivores. Sugar produced
Encyclopedia of Caves.
© 2012 Elsevier Inc. All rights reserved.
ROOT COMMUNITIES IN LAVA TUBES
659
FIGURE 1 Lava tube passage with tree roots.
in the tree leaves is transported downward through
the trunk and into the roots, where it becomes the
energy source for root growth and for the species that
feed on the roots (Fig. 2). Species that live on and
around tree roots include many surface species, such
as ants, that venture underground but are not neces¬
sarily adapted to survive entirely in the deep cave
zone. Other species spend a portion of their life cycle
underground on or near tree roots, such as immature
stages of planthoppers (Hemiptera: Fulgoromorpha;
e.g., Cixiidae) and cicadas (Hemiptera: Cicadidae).
Cicada nymphs spend up to 17 years underground
feeding on tree roots until they crawl up the tree trunk
and emerge as surface-dwelling adults. In spite of
their long sojourn underground, no cicada species
has yet been discovered that spends its entire life
cycle in the subterranean environment. This adaptive
shift has occurred many times by planthopper species,
however.
Moisture/Humidity
Troglobites live in air saturated with water vapor
and have a number of adaptations to cope with high
humidity, such as loss of pigments and thinning of
their skin or cuticle. Species that feed directly on
the watery sap in tree roots can survive in some¬
what drier areas due to the constant supply of mois¬
ture through the tree roots. Examples include
nymphs of cave planthoppers in the family Cixiidae
(Table 1).
Air Flow
Restricted airflow is essential to survival of cave-
adapted species, as airflow is usually desiccating.
Certain types of air motion may actually introduce
moisture into caves in dry areas. For example, the inter¬
action of surface air with cave air often causes moisture
to condense in the zone of contact. This results in a
drip zone just inside the entrance of caves. Lava tubes
in dry areas of Hawaii sometimes act to remove mois¬
ture from air flowing through them, if they are suffi¬
ciently deep and have restrictions to prevent the loss of
water vapor. These cave passages also capture percolat¬
ing rain water. Hawaiians took advantage of these
water traps as sources of water for drinking and agri¬
culture. Tree roots follow the moisture, so often on dry
lava flows on the leeward side of Hawaii, native pio¬
neering trees, Metrosideros polymorpha ('ohi'a), survive
by sending their roots into these water-trap caves.
0 2 and C0 2
In the root zone deep in the soil, or in mesocaverns,
oxygen is depleted and carbon dioxide increased by res¬
piration of the roots and of soil organisms. Air exchange
with the surface is restricted in these zones by the small,
labyrinthine nature of the spaces, allowing the oxygen/
carbon dioxide imbalance to be maintained. Species liv¬
ing in this zone need to adapt to the low oxygen and
high carbon dioxide levels. In deep caves with restricted
airflow, a similar oxygen/carbon dioxide imbalance can
occur for the same reasons. Areas with animal guano or
ENCYCLOPEDIA OF CAVES
660
SOLAR ENERGY:
PHOTOSYNTHESIS
CO,
TREE LEAVES ON SU ^CE
v
WATER
MINERALS,
NUTRIENTS
LAVA SUE STRATE
ACCIDENTALS
=F
TREE ROOTS IN LAVA TUBE
Metrosideros polymorpha /
F
CAVE
MOTHS
Schrankia howarthi
CAVE
CENTIPEDES
Lithobius sp.
fungus, *
BACTERIA <
CAVE
PLANTHOPPERS:
Oliarus polyphemus
V
CAVE
SPIDERS:
Lycosa howarthi
"DETRITUS: DEAD ROOTS AND
-ANIMALS, WASTE PRODUCTS
FUNGI: Pistillaria sp. on moth cocoon
CAVE
MILLIPEDES
Nannolene sp.
CA
SPRINGTAILS
CAVE CRICKETS:
Caconemobius uuku
Inorganic energy and nutrient flow:
Organic energy and nutrient flow:
o
—►
FIGURE 2 Cave root food web.
ENCYCLOPEDIA OF CAVES
ROOT COMMUNITIES IN LAVA TUBES
661
TABLE 1 Cave Tree Root Communities in Hawai'i and Australia
Hawai'i Lava Tubes
Howarth (1981,1991)
Australia: Bayliss Lava Tubes
Howarth & Stone (1990)
Producers
Metrosideros polymorpha ('Ohi'a tree)
Brachychiton spp.. Eucalyptus spp.. Ficus spp.
Primary consumers
Insecta: Hemiptera: Cixiidae (cave
planthoppers)
Oliarus polyphemus and other Oliarus species
Solanaima baylissa
Insecta: Lepidoptera: Noctuidae (cave
moths)
Schrankia howarthi
Insecta: Coleoptera (beetles)
Curculionidae: Rhytirhininae
Secondary consumers
Aranea (spiders)
Lycosidae: Lycosa howarthi ; Linyphiide; Theridiidae
Pholcidae: Spermophora sp.; Nesticidae:
Nesticella sp.; Zodariidae: Storena sp.
Acari (mites)
Rhagidiidae: Foveaclieles sp.
Chilopoda (centipedes)
Lithodiidae: Lithobius sp.
Scutigeridae
Hemiptera
Reduviidae: Nesidolestes ana
Reduviidae: Micropolytoxus
Scavengers, detritivores, fungivores
Insecta: Dermaptera: Anisolabidae
Anisolabis howarthi
Insecta: Orthoptera (crickets)
Gryllidae: Caconemobius varius, C. unka,
Thaumatogryllus cavicola (cave tree cricket)
Insecta: Blattodea: Nocticolidae,
Blatellidae (cave cockroaches)
Nocticolidae: Nocticola sp.; Blatellidae:
Neotemnopten/x baylissensis
Insecta: Hemiptera: Mesoveliidae
Cavaticovelia aaa
Diplopoda (millipedes)
Cambalidae: Nannolene sp.
Polyxenidae; Polydesmidae; Cambalidae (all
undetermined species)
Crustacea: Isopoda
Littorophiloscia sp.
Oniscomorpha: undetermined species
Insecta: Collembola
Sinella yoshia; Protanura hawaiiensis
Entomobryidae: Pseudosinella sp.
Decomposers
Fungi
Pistillaria sp.
washed-in organic matter can have enhanced carbon
dioxide and lowered oxygen due to respiration by bac¬
teria and other decomposers. Carbon dioxide is denser
than air; thus, if no mixing occurs, CC^-laden air will
tend to flow downward and accumulate. With
restricted air flow, the air in caves may become strati¬
fied with higher carbon dioxide concentrations in lower
areas, Deharveng & Bedos (1986).
Influence of Cave Geomorphology on Ecological
Conditions
The shape of caves and lava tubes is critical for the
occurrence of habitat suitable for cave-adapted species.
The most important elements are the following: (1)
restriction of airflow to reduce the impact of surface
air, which can cool and dry the cave; (2) passages
that trap water vapor, such as low ceiling segments
followed by higher ceiling areas, upsloping, or upper-
level, dead-end passages; and (3) passages that serve
as carbon dioxide sumps ( e.g ., low floor areas or down-
sloping dead-end passages) Howarth (1981,1991).
CAVE ZONES IN RELATION
TO ROOT GROWTH
Tree Roots
Tree roots penetrate the substrate to obtain water and
nutrients for growth. In lava flows and limestone areas
ENCYCLOPEDIA OF CAVES
662
R
with fractured or cavernous substrate, surface water rap¬
idly percolates to the water table at some depth below the
surface. Some tree species are adapted to send their roots
deep beneath the surface to reach the water table. Areas
with fractured or cavernous rock often have little or no sur¬
face soil, due to its erosion into the underground habitat,
or, in the case of lava tubes, due to the recent lava flows.
Trees may also have adaptations that allow them to sur¬
vive on nutrients that leach from the cave rocks. Where the
water table is too far below the surface for roots to pene¬
trate, the voids in the substrate may accumulate water
vapor that the roots can utilize. In Hawaii, trogloxenes
such as bats and crickets do not occur and therefore do not
contribute nutrients in guano to the cave community. In
mainland areas, such as the Undara lava tubes in
Queensland, Australia, bats bring significant amounts of
guano into the lava tubes that provide nutrients for the
tree roots as well as the scavengers (guanobites).
Trees that can grow on bare rock have adaptations
for sending roots deep to a water source and may pene¬
trate into caves more than 30 m belowground, Jackson
et ah, (1999). This is especially true in the tropics, where
the high evapotranspiration rate imposes a severe stress
on trees. In the wet/dry tropics, these trees often stay
green while those growing in soil often lose their leaves
in the dry season. In Hawaii, varieties of 'ohi'a are
adapted to colonize bare rock in new lava flows. Roots
appear to have special adaptations for growing on bare
rock. The 'ohi'a tree is one of the first colonists on new
lava flows, and it also grows in wet soil areas in the
rainforest, Cordell et ah, (1998). Research in lava tubes
in Hawaii Volcanoes National Park showed that trees
in wetter areas had roots with thinner outer layers and
larger vessels in the xylem, McDowell (2002). This
adaptation would make the roots easier to penetrate by
the piercing mouthparts of the cave planthoppers. In
Australia, Brachychiton species and some Eucalyptus
species send roots into caves, and Ficus species do this
wherever they occur throughout the tropics.
Entrance
Entrance zones of lava tubes are often more densely
vegetated than are the surrounding areas. This is
partly because the entrance pits accumulate dust and
debris from the surface, providing nutrients, and
partly due to the ready access to moisture. A drip zone
often occurs just inside cave entrances, due to the mix¬
ing of surface and cave air which results in condensa¬
tion of water vapor. This can often be seen as a fog
zone when conditions are right. Cave-adapted species
that generally occur in the deep cave zone can some¬
times be found under rocks, in cracks, or on roots due
to the constant moisture in this zone. Tree roots that
hang from the ceiling in the entrance zone may be
more similar to aerial roots that occur on the surface,
due to their periodic exposure to desiccating surface
air. They tend to be bushier and have a thicker cortex
than roots in the deep cave environment.
Twilight
This is the zone just inside the entrance where some
light penetrates. It is strongly impacted by surface air,
so it is generally not suitable for cave-adapted species.
In cases where there is a restricted entrance or an
upsloping dead-end passage, the twilight zone may be
small or absent, and cave-adapted species can then be
present very near the entrance.
Transition/Mixing Zone
Beyond the zone of light penetration is an area that
still receives some impact from surface air. This zone is
drier than the deeper cave and has a more variable tem¬
perature. In many temperate caves, this zone expands in
the winter when the surface temperature remains below
cave temperature. This winter effect is due to cool sur¬
face air moving into the cave, where it is warmed by the
latent heat of the cave. The warming process evaporates
moisture from the cave surface, desiccating the passage.
In the tropics, surface temperatures generally rise above
and fall below cave temperatures nearly every day creat¬
ing a diurnal winter effect. This zone often has troglo-
philic and trogloxenic species that can cope with the
drier conditions and more variable temperature. A zone
of mixing occurs between the transition zone and the
deep cave air, which can sometimes be seen as a foggy
area due to condensation of water vapor into droplets.
Deep Cave (Saturated Humidity) Zone
This zone is beyond the effects of surface air.
Humidity remains near 100%, and the temperature
remains relatively constant. Air motion is reduced.
Due to the constant moisture and lack of desiccation,
tree roots proliferate in this zone and provide the basis
for some of the most diverse communities of cave-
adapted species.
Bad Air (Saturated Humidity Plus High Carbon
Dioxide/Low Oxygen) Zone
In areas with low dead-end passages, virtually no air
motion, and an energy source for respiration, carbon
dioxide can increase and oxygen decrease. Carbon diox¬
ide readily mixes with the air, so there is no buildup
where there is air exchange with other zones; however,
where air is stagnant and biogenic carbon dioxide is
ENCYCLOPEDIA OF CAVES
ROOT COMMUNITIES IN LAVA TUBES
663
being produced, cave air can stratify, resulting in higher
carbon dioxide levels in low areas. If these areas also
have high humidity and food energy sources, they are
home to the most highly cave-adapted species. Bayliss
Cave is a lava tube in Australia with a large bad air
zone, high humidity, and abundant tree roots penetrat¬
ing the cave ceiling and growing through the soil on the
floor. It is home to 24 highly cave-adapted species,
among the highest diversity of any cave community,
Atkinson et al., (1976), Howarth & Stone (1990).
MORPHOLOGY AND ADAPTATIONS
OF CAVE SPECIES
Obligate cave species respond to cave conditions of
total lack of light, high humidity, constantly wet substrate,
and potentially high carbon dioxide and low oxygen by
adaptive alteration of their morphology, physiology, and
behavior. Morphological changes include loss of eyes and
pigment; flightlessness; enhanced senses of smell, hearing,
and touch and ability to detect air motion; elongation of
appendages such as legs and antennae; loss of pulvilli
and lengthening of claws; and rotation of legs to keep the
body upright. Physiologically, cave species lose their circa¬
dian rhythms and diurnal activity patterns; they often
have fewer but larger offspring and a slower metabolic
rate than surface species. Though energy sources may be
abundant in caves, they are often more dispersed than on
the surface and more difficult to find in the dark, three-
dimensional maze, so cave species have adapted to sur¬
vive long periods without feeding.
Behaviorally, cave species must develop special
adaptations for finding food, finding a mate, and
avoiding predators. In the labyrinthine spaces of caves
and mesocaverns, cave species may resort to a constant
random search for food. Tree roots can assist the
search by providing a pathway that cave species can
follow. Because the roots are also an energy source,
predators and scavengers may use the root pathways
to assist in locating prey. Tree roots also provide hid¬
ing places for predators such as the small-eyed/big-
eyed hunting spiders.
Cave planthoppers use tree roots for their food
source as well as to assist them in finding mates.
Female planthoppers sit on the roots and produce a
substrate-borne sound that travels readily along the
roots. Male planthoppers move randomly over the cave
walls until they find a root. When a male detects a sig¬
nal from a female, he returns the call and the pair
"duet" until the male locates the female. The female
lays its eggs on the roots, and the nymphs insert
their mouthparts into the xylem to feed. Planthopper
nymphs produce wax filaments from a gland on their
abdomen and make a loose wax "cocoon" that protects
them from predators while they feed, Hoch & Howarth
(1989,1993)
Cave moths use a different strategy for finding their
mates. Female moths sit on the roots and often have
reduced wings and are flightless. Male moths fly errati¬
cally through the cave with their legs held out in front.
They have extremely sensitive touch and can feel a
small root and even alight on a strand of spider web
without alerting the spider. The female moth releases a
pheromone, and when the male detects the chemical it
flies upwind until it reaches the female. The female
lays its eggs on the root, and the inchworm caterpillar
feeds on the small tree roots. It also uses the roots to
build a hanging cocoon to protect it from predators
while it pupates.
Male cave crickets and cockroaches have tergal
glands that release pheromones to attract the females,
as do their surface relatives. The female crawls onto
the male to the tergal gland, and the male attaches to
her with a special genital hook. In Thai species of
the cockroach family Nocticolidae, the tergal glands
are highly modified into large horn-like structures,
whose function has not been studied.
CONSERVATION OF CAVE
ROOT COMMUNITIES
As with other cave communities, cave root commu¬
nities include many rare and highly specialized species
that may be threatened by impacts to their habitat.
However, their survival requires more than just protec¬
tion of caves or cave passages. It is dependent on pro¬
tection of the plant roots that are the main energy
source for the community. This requires protection of
the overlying surface ecosystems. Identification of the
plant species that supply the roots in the caves is impor¬
tant for a management plan that will protect the critical
plant species and allow their re-introduction in adjacent
areas where they have been lost. Invasive deep-rooted
species that compete with the native plant species and
alter the cave root composition also need to be identi¬
fied. Roots deep in caves are difficult to identify visu¬
ally, as their growth form differs significantly from that
of roots near entrances or in soil. Recently, use of DNA
analysis has become an important conservation and
management tool by making accurate root identifica¬
tions possible, Howarth et al., (2007).
See Also the Following Articles
Adaptive Shifts
Food Sources
Volcanic Rock Caves
ENCYCLOPEDIA OF CAVES
664
Bibliography
Atkinson, A., Griffin, J. J., & Stephenson, P. J. (1976). A major lava
tube system from Undara Volcano, North Queensland. Bulletin of
Vulcanology, 39, 1—28.
Cordell, S., Goldstein, G., Meuller-Dombois, D., Webb, D., &
Vitousek, P. M. (1998). Physiological and morphological variation
in Metrosideros polymorpha, a dominant Hawaiian tree species,
along an altitudinal gradient: The role of phenotypic plasticity.
Oecologia, 113, 118-196.
Deharveng, L., & Bedos, A. (1986). Gaz carbonique (pp. 144—152).
Thai-Maros 85: Rapport Speleologique et Scientifique to Thailand
and Sulawesi. Toulouse, France: Association Pyreneenne de
Speleologie.
Hoch, H., & Howarth, F. G. (1989). Six new cavernicolous cixiid
planthoppers in the genus Solonaima from Australia (Homoptera:
Fulgoroidea). Systematic Entomology, 14, 377—402.
Hoch, H., & Howarth, F. G. (1993). Evolutionary dynamics of
behavioral divergence among populations of the Hawaiian
cave-dwelling planthopper Oliarus polyphemus (Homoptera:
Fulgoroidea: Cixiidae). Pacific Science, 47, 303—318.
Howarth, F. G. (1973). The cavernicolous fauna of Hawaiian lava
tubes. 1. Introduction. Pacific Insects, 15, 139—151.
Howarth, F. G. (1981). Community structure and niche
differentiation in Hawaiian lava tubes. In D. Mueller-Dombois,
K. W. Bridges & H. L. Carson (Eds.), Island ecosystems: Biological
organization in selected Hawaiian communities (Vol. 15, US/IBP
Synthesis Series (pp. 318—336). Stroudsburg, PA: Hutchinson
Ross Publishing.
Howarth, F. G. (1991). Hawaiian cave faunas: Macroevolution
on young islands. In E. C. Dudley (Ed.), The unity of evolutionary
biology (Vol. 1). Portland, OR: Dioscorides Press.
Howarth, F. G., James, S. A., McDowell, W., Preston, D. J., & Imada,
C. T. (2007). Identification of roots in lava tube caves using molec¬
ular techniques: Implications for conservation of cave arthropod
faunas. Journal of Insect Conservation, 11, 251—261.
Howarth, F. G., & Stone, F. D. (1990). Elevated carbon dioxide levels
in Bayliss Cave, Australia: Implications for the evolution of obli¬
gate cave species. Pacific Science, 44, 207—218.
Jackson, R. B., Moore, L. A., Hoffman, W. A., Pockman, W. T., &
Linder, C. R. (1999). Ecosystem rooting depth determined with
caves and DNA. Proceedings of the National Academy of Science
USA, 96, 11387-11392.
McDowell, W. M. (2002). An ecological study of Metrosideros polymor¬
pha Gaud. (Myrtaceae) roots in lava tubes. Master's thesis.
University of Hawai'i, 91 pp.
ENCYCLOPEDIA OF CAVES
SALAMANDERS
Spela Qoricki, Matthew L. Niemiller , 1 and
Dante B. Fenolio
University of Maryland, ^University of Tennessee,
*Atlanta Botanical Garden
Salamanders are a diverse group of vertebrates, exploit¬
ing moist cool habitats in a variety of ways. Several
lineages have colonized subterranean habitats, particu¬
larly in regions of climatic extremes. Salamanders associ¬
ated with karst exhibit differences in the amount of time
they spend in subterranean habitats, their dependence
on these resources, and morphological and behavioral
adaptations to life underground (troglomorphisms).
Differences are manifested within and across taxonomic
groups as well as between geographical regions. North
America and Europe host the greatest number of known
cave-dwelling species. Cave-dwelling salamanders may
inhabit subterranean environments for significant por¬
tions of their life cycle but not all of it, others migrate to
the surface periodically, and still others are exclusively
found underground. Sources of general information on
salamanders are Amphibia Web (2011), Grossenbacher
and Thiesmeier (1999), and Lannoo (2005).
Salamanders are the only tetrapods that exhibit an
exclusive subterranean existence. Such troglobitic sala¬
manders belong to the families Plethodontidae (in North
America) and Proteidae (in Europe). Troglomorphisms
seen in troglobitic salamanders vary from species to spe¬
cies and even from locality to locality, but they frequently
include reduced eyes and pigmentation, reduced number
of trunk vertebrae, elongated appendages, and modified
head shape with increased dentition. Salamanders are
opportunistic predators, feeding on a variety of small
invertebrates and will often also consume their own kind.
There are records of notable dietary oddities in troglobitic
salamanders, including feeding on bat guano and silt.
Troglobitic salamanders are usually found in caves where
fishes do not occur, and function as top predators in
those subterranean waters. Because of the constant water
temperatures in their habitat, they are likely active year-
round. For orientation, feeding, and mating, they rely on
mechano- and chemosensory cues, which corresponds to
progressive development of extraoptic sensory systems.
Salamanders from subterranean environments usually
have lower metabolism, efficient energy (fat) storage, and
a longer life span than their surface counterparts. They
have fewer, larger eggs and offspring compared to
related surface-dwelling species. The rate of development
in amphibians strongly depends on environmental fac¬
tors, most notably ambient temperature. When compared
with related epigean species that reproduce at the same
water temperature, troglobitic salamanders usually
develop more slowly. All begin their life inside a gelati¬
nous egg capsule, deposited in water. Eggs hatch into free-
swimming aquatic larvae, with visible bushy gills and a
tailfin. Instead of metamorphosing into mature adults,
however, most troglobitic salamanders attain sexual matu¬
rity while retaining these and other larval characteristics—
a condition known as neoteny or paedomorphism.
We can only speculate about the origins and time
required for the evolution of obligate subterranean exis¬
tence. While the ancestors of some troglobitic species may
have colonized caves as early as in the Miocene, it is more
likely that subterranean colonization occurred much later,
even as late as during the Pleistocene. Salamander
fossils are rare and the oldest found that may belong to a
cave-dwelling species date to the Pleistocene.
The following sections portray all ten currently
described troglobitic salamanders. Accounts also are
provided for their close relatives that show affinities to
subterranean habitats.
OLMS AND WATERDOGS (PROTEIDAE)
The family includes six species of surface-dwelling
mudpuppies or waterdogs (Necturus sp.) from eastern
and central North America and the troglobitic olm
(Proteus anguinus) found in southeastern Europe. These
Encyclopedia of Caves.
665
© 2012 Elsevier Inc. All rights reserved.
666
s
FIGURE 1 The olm ( Proteus anguinus). (A) Described in 1768, this
blind salamander was the first scientifically documented cave-dwell¬
ing animal, but its existence had already been known long before. Its
earliest representation may be a Venetian stone carving from the
tenth or eleventh century. (B) A rare, pigmented individual with
eyes, from southeastern Slovenia. This unique population was dis¬
covered toward the end of the twentieth century. (C) Female protect¬
ing her clutch of eggs from predation by conspecifics. Photos courtesy
of G. Aljancic, Tutor Cave Laboratory. Used ivith permission.
salamanders are aquatic throughout their life, retaining
external gills, but are also able to breathe air through
their skin and with their sac-like lungs. All representa¬
tives of this family are obligate paedomorphs; paedo-
morphism in Proteus is thus seemingly not a result of
conditions particular to the subterranean environment.
Thyroid glands, which control metamorphosis, are
functional in all species of this family. Metamorphosis
cannot be induced artificially, although changes in the
skin structure toward a metamorphosed form have
been observed in Proteus. Furthermore, it has been
shown that Necturus possesses functional thyroid hor¬
mone receptors and its tissues are not generally irre¬
sponsive to the hormone (Safi et al., 2006). Rather, the
absence of metamorphosis may be due to loss of func¬
tion of thyroid hormone-dependent genes required for
tissue transformation.
Olm (Proteus anguinus)
The olm ( Proteus anguinus) is the only European obli¬
gate cave-dwelling vertebrate. It has been found in
over 250 springs and caves in the Dinaric karst on the
western Balkan Peninsula, between the rivers Isonzo/
Soca (northeastern Italy) and Trebisnjica (Bosnia and
Herzegovina). Several populations, however, have been
destroyed or severely diminished by dam construction,
earth fill, or pollution. The salamander inhabits karst
waters from close to the surface to deep in fissures.
During flooding or at night it may be found in cave
entrances or short distances away from springs.
This elongated salamander reaches 25 cm in total
length, although 40-cm-long individuals have also
been found. The salamander's predominant mode of
locomotion is swimming with eel-like undulations of
the body ending in a short, finned, laterally com¬
pressed tail (Fig. 1A). The very short limbs, with digits
reduced in number to three on the forelegs and two on
the hindlegs, are used for crawling on the water bot¬
tom and occasionally outside the water. Its muscular
head ends with a flattened, blunt snout.
The species is currently divided into two subspecies
based on the extent of troglomorphism. Analyses of
molecular data suggest, however, that P. anguinus may
be a complex of several species that independently
evolved troglomorphic traits. Mitochondrial DNA data
suggest that P. anguinus is divided into six divergent
lineages. These are about 5 to about 15 million years
old, that is, they originated in the Pliocene or Miocene.
Conversely, it has been hypothesized that troglo¬
morphism might have evolved in less than 500,000
years (Trontelj et al., 2007).
Morphologically, the extremely rare subspecies
Proteus anguinus parkelj, found only in the springs of
two streams near Jelsevnik in Bela Krajina (southeast¬
ern Slovenia), is believed to resemble the supposed
epigean ancestor of both forms. Despite having an
entirely subterranean life, this salamander is darkly
pigmented (Fig. IB). Its small but functional eyes are
covered with a transparent skin and have well-differ¬
entiated lens and retina. Conversely, the pineal gland,
which in amphibians controls circadian rhythms, gam¬
ete development, and pigmentation changes, is greatly
reduced.
In contrast, the widespread, troglomorphic Proteus
anguinus anguinus is characterized by a yellowish to
pinkish-white skin, with few scattered melanophores
(pigment cells) invisible to the naked eye (Fig. 1A).
Melanin synthesis can be light-induced, and after a
prolonged exposure to light, the animals turn dark.
The skin is thinner than in the pigmented form,
and contains fewer multicellular glands. The head and
snout of P. a. anguinus are elongated and flattened, and
ENCYCLOPEDIA OF CAVES
SALAMANDERS
667
the cranial bones are longer and the teeth more numer¬
ous than in the pigmented form. Cervical vertebrae are
often elongated as well. Compared to the pigmented
form, P. a. anguinus has a shorter trunk, but a propor¬
tionately longer tail and extremities. Morphometric
variability among genetically and hydrologically iso¬
lated populations can be quite high.
The eyes of this salamander are greatly reduced in
size and structure, and lie embedded deep in the
hypodermal tissue. While vision is lost, the eye is still
capable of detecting light. The pineal gland is also
greatly reduced, but may still be capable of light detec¬
tion and hormonal activity. As is common in salaman¬
ders, the skin and midbrain are also photosensitive.
Extraoptic sensory systems, including the lateral line,
inner ear, ampullary electroreceptors, olfactory epithe¬
lium, and taste buds, are well developed and believed
to play a crucial role in prey detection and communi¬
cation among individuals. Furthermore, behavioral
experiments suggest that the animals may use the
Earth's magnetic field to orient themselves.
Because of its very low metabolic rate and efficient
fat storage, predominantly in the liver and tail,
P. anguinus can survive prolonged periods of food
deprivation. When food in the form of small subterra¬
nean crustaceans, mollusks, or surface insect larvae
is abundant, it becomes a voracious and efficient
predator.
P. anguinus is a social animal, recognizing and com¬
municating with conspecifics through scent. Both
males and females use water-borne chemical signals to
locate each other, but in order to recognize sex and
reproductive state, they need to come into direct con¬
tact. Outside of the breeding period, aggressive behav¬
ior is reduced and the animals aggregate at communal
resting places under stones and in cracks. Long peri¬
ods of inactivity are punctuated by sessions of foraging
or exploratory behavior. The animals show no appar¬
ent daily rhythm of activity and resting.
Little is known about timing of reproduction. A
female is thought to reproduce only every 12 years on
average. Up to 70 eggs are attached under rocks, and
are guarded by the female until hatching (Fig. 1C).
Larvae hatch after 4 to 6 months. The newly hatched
larvae are 2.5 cm long and have incompletely formed
limbs, but well-differentiated eyes and scattered, visi¬
ble melanophores in the skin. During early larval
development, the eyes only slightly increase in size
without further differentiation, and then gradually
sink into the surrounding tissue. After four months,
major degenerative changes appear which cannot be
prevented by illumination. Sexual maturity is reached
at a size of about 20 cm, which is attained after about
15 years. P. anguinus is thought to have the longest life
span of all amphibians, living for 70 years or more.
LUNGLESS SALAMANDERS
(PLETHODONTIDAE)
Salamanders of the family Plethodontidae are widely
distributed in North and Central America, with popula¬
tions also in South America, southern Europe, and on
the Korean Peninsula. Lacking lungs, these salamanders
breathe primarily through their skin (cutaneous respira¬
tion). Pheromones play an important role in communica¬
tion, and the characteristic nasolabial grooves present in
metamorphosed adults allow fine-tuning in chemorecep-
tion. A great diversity of lifestyles has evolved in this
group, from terrestrial, arboreal, to semi-aquatic and
aquatic; life histories range from biphasic and larval
reproduction to direct development and possibly even
vivipary. It is no surprise to find in this family species
associated with caves. The family includes nine
described troglobitic species, all found in central and
eastern North America. They are grouped in the genera
Eurycea, Haideotriton, and Gyrinophilus. Formerly used
genera Typhlotriton and Typhlomolge have been synony-
mized with Eurycea.
Brook Salamanders ( Eurycea)
This is a diverse genus of 26 recognized species, five
of which are troglobites. Except for Eurycea spelaea, the
troglobitic species inhabit subterranean waters of the
Edwards Plateau and Balcones Escarpment in central
Texas, in the United States. Many species of Eurycea in
this region are spring-dwelling troglophiles, and addi¬
tional troglophilic Eurycea are encountered in caves
outside the region. Paedomorphism is common in both
cave- and spring-dwelling salamanders. Eurycea spelaea
is the only metamorphosing troglobitic member of the
genus.
Texas Cave and Spring-Dwelling Eurycea
All species of Eurycea from the Edwards Plateau in
Texas share a common ancestor. It has been hypothe¬
sized that about 15 million years ago a split into two
lineages occurred, corresponding to the geographic
divide imposed by the Colorado River. The resulting
lineage south of the divide later gave rise to a group of
extremely troglomorphic cave-dwelling species collec¬
tively called Typhlomolge and its sister lineage com¬
prised of less troglomorphic cave- and spring-dwelling
species called Blepsimolge. For the spring-dwelling
species north of the Colorado River the name
Septentriomolge has been coined. Each troglobitic spe¬
cies has a surface counterpart inhabiting the spring
outflows of the same watershed or geological forma¬
tion. The complex interspecies relationships along with
ENCYCLOPEDIA OF CAVES
668
s
FIGURE 2 Brook salamanders (Eurycea). Troglobitic
E. rathbuni (A) and (B), E. waterlooensis (C), and
E. tridentifera (D) from the Edwards Plateau have
degenerate eyes and reduced pigmentation compared
to spring-dwelling populations of E. sosorum (E)
and E. Intitans (F). Adult (G) and larva (H) of Eurycea
spelaea from the Ozark Plateau. This species is the only
troglobitic brook salamander to readily undergo meta¬
morphosis in nature. Photos by D. Fenolio.
their geographic and ecological segregation indicate
that an ancient speciation event giving rise to the sub¬
terranean and largely epigean species was followed by
a rapid adaptive radiation and more recent coloniza¬
tion of subterranean habitats by still more species.
Evidence of intermittent gene flow between species
that diverged millions of years ago has been detected.
Conversely, in some of the species with relatively
broad ranges, there is generally very little migration
between geographically distant populations, with "dis¬
tant" sometimes referring to as little as a few hundred
meters (Chippindale et al., 2009; Lucas et al., 2009).
Typhlomolge (Blind Salamanders)
All species are neotenic troglobites with very long
and thin limbs, short trunk, broad and flattened snout,
and a virtually pigmentless and translucent skin
(Fig. 2A). They have a shimmering white appearance
due to the reflective connective tissue that lies
beneath the skin. The rudimentary eyes may be visi¬
ble as tiny dark spots under the skin, and possibly
still detect light. These salamanders are moderately
small with the total length of adults ranging from
7—14 cm. Their diet consists of a variety of small
aquatic subterranean invertebrates, including crusta¬
ceans and snails. When feeding, the salamanders usu¬
ally probe the bottom using lateral movements of
their spatulate head.
The Texas blind salamander (Eurycea rathbuni) has
been found in caves, wells, and pipes that intersect the
San Marcos Pool of the Edwards Aquifer, although it
may have a wider geographic range than previously
thought. Molecular analyses revealed high levels of
genetic variation and the possibility that two distinct
species occur within E. rathbuni. In springs immedi¬
ately above the subterranean waters occupied by
E. rathbuni, E. nana (see below) is locally abundant.
A captive breeding program has been established for
both species.
ENCYCLOPEDIA OF CAVES
SALAMANDERS
669
Artificial induction of transformation in E. rathbuni
resulted in only partial metamorphosis, with some tissues
{eg., skin) continuing to show only larval features. This
salamander has an extremely depressed and broad ante¬
rior part of the head (Fig. 2B), with elongated anterior
cranial bones, and the longest limbs of all troglomorphic
Eurycea. The eyespots are barely visible. Juveniles have
darker pigmentation and proportionally larger eyes.
In contrast, the olfactory system is well developed
and plays a very important role in the social behavior
of this salamander (Epp et al., 2010). Nothing is known
about social interactions or breeding of this species in
nature; however, they have been observed in captivity.
These salamanders rarely show aggressive or social
behavior. The male of E. rathbuni lacks mental (chin)
and caudal hedonic glands used during courtship in
other Eurycea salamanders. Nevertheless, the species
has an elaborate tail-straddling walk similar to that
which has been observed in other plethodontid sala¬
manders, albeit somewhat simplified. In captivity,
these salamanders can live for over 10 years.
The Blanco blind salamander ( Eurycea robusta) was
described based on a single specimen collected from a
well drilled in the bed of the Blanco River east of San
Marcos. More specimens had been collected at the time,
but birds consumed two individuals while they were in
a bucket and one preserved specimen has vanished.
Molecular data are not available for this species, but the
geological formation in which it occurs is hydrologically
isolated from that in which the geographically proximal
E. rathbuni is found. The epigean salamander found in
the springs of the Blanco River drainage is E. pterophila
(see below). Compared to E. rathbuni, the body of
E. robusta is stockier, with longer trunk, and shorter, rel¬
atively robust limbs. The head also is wider posteriorly
and more massive. The eyes are extremely reduced and
invisible through the skin.
The Austin blind salamander ( Eurycea waterlooensis) is
known only from the outflows of Barton Springs in
Austin, where individuals presumably are washed out
from the Barton Springs segment of the Edwards
Aquifer. This species is much more rarely encountered
than the sympatric surface-dwelling E. sosorum (see
below), both of which are included in a captive breeding
program. This species is similar to E. rathbuni, but has a
wider head (Fig. 2C). The eyespots are visible and super¬
ficially resemble those of E. rathbuni. The limbs are
shorter and the skin is slightly more pigmented than in
E. rathbuni and E. robusta. The tail fin is weakly developed.
Blepsimolge (Sighted Salamanders)
A large majority of these salamanders are neotenic
and completely aquatic. Besides those listed below, the
group includes several populations of yet undescribed
species. The salamanders of this group are primarily
surface-dwelling, found in the immediate vicinity of
spring outflows, but they also inhabit caves, fissures,
and sinkholes. Various species have independently
evolved cave-associated morphological traits, which
has become more apparent as new populations are dis¬
covered and genetically analyzed (Bendik et al., 2009).
These small salamanders (6—10 cm long) usually
have a dark, finely patterned dorsal coloration and
translucent ventral parts (Figs. 2E,F). The intensity and
pattern of dorsal coloration may vary due to an irregu¬
lar pattern of melanophores and reflective white irido-
phores. Single or double dorsolateral rows of light
spots extend along the body. The species differ in eye
size, head shape, tail fin shape, and color pattern on
the tail and flanks. Certain cave-dwelling populations
exhibit marked reduction of eyes and pigmentation.
The skin of the troglomorphic animals has a light
yellowish color, with diffuse gray or brown spots dor-
sally and lighter spots laterally (Fig. 2D). Juveniles are
darker than adults. The snout is flattened but trun¬
cated. The eyes are regressed, sometimes lacking a
lens, but are usually still visible through the skin as
dark spots; they are larger than the rudimentary eyes
of E. rathbuni and other members of the Typhlomolge
group. The trunk is short, containing fewer vertebrae,
and the limbs are somewhat elongated.
The life history of most species is poorly known and
based predominantly on observations of E. nana and
E. sosorum in captive breeding programs (e.g., Najvar
et al., 2007). The spring-dwelling Eurycea are fairly sed¬
entary, although they may seasonally move between
surface and cave habitats. On the surface, these secre¬
tive salamanders usually hide under rocks, in gravel
substrate, or among aquatic plants. They can survive
temporary drying of ephemeral springs, presumably
by retreating to subsurface refugia. They are active
throughout the year and appear to reproduce during
all seasons. Up to 70 eggs are laid singly and do not
receive parental care. In captivity, the salamanders
usually live for about 4 years, but an individual of
E. sosorum survived to at least 12 years.
Historically, many spring-dwelling populations from
throughout the Edwards Plateau were assigned to the
Texas salamander ( Eurycea neotenes) based on morpho¬
logical similarity, but they were found to differ geneti¬
cally. This species is now restricted to the springs in the
vicinity of Helotes, near San Antonio. A disjunct and
genetically divergent population inhabits the area of
Comal Springs in New Braunfels.
One salamander that has been elevated to species
from E. neotenes is the Barton Springs salamander
{Eurycea sosorum), known from the Barton Springs pool
and adjacent springs in Austin. This species is
ENCYCLOPEDIA OF CAVES
670
s
primarily surface-dwelling, but appears to reproduce
in subterranean conduits. Skin pigmentation in
E. sosorum varies considerably, but all individuals have
very small eyes and somewhat elongated limbs. The
flattened, slightly elongated head ends with a trun¬
cated snout (Fig. 2E). The subterranean portion of the
Barton Springs Aquifer is inhabited by the troglobitic
E. ivaterlooensis (see above).
The San Marcos salamander ( Eurycea mm) is abun¬
dant in the Spring Lake pool at the source of San Marcos
River, the only site where it occurs. Metamorphosis has
been induced artificially in this species.
The Fern Bank salamander ( Eurycea pterophila) inha¬
bits springs and caves of the Blanco and Guadalupe
River drainages. Most populations are morphologically
similar to E. neotenes, to which they were formerly
assigned, but cave-dwelling individuals with reduced
eyes and pigmentation have also been identified. This
species is genetically different from E. neotenes, but
similar to, perhaps even conspecific with, E. latitans
and/or E. tridentifera (see below).
The Cascade Caverns salamander ( Eurycea latitans)
is possibly a complex of several species found in caves
and springs of the Cibolo Creek basin and south of the
Guadalupe River in the southeastern part of the
Edwards Plateau. For some time before initial molecu¬
lar analyses were completed, most populations were
regarded as E. neotenes, and the population at the
type locality (Cascade Caverns) was considered to be
hybrid between E. neotenes and E. tridentifera (see
below). This population includes individuals with a
spectrum of morphological features, ranging from
highly troglomorphic, most similar to those of
E. tridentifera, to surface-like (Fig. 2F), most similar to
what was historically considered E. neotenes. Recent
molecular analyses suggest that at least part of
E. latitans, including the Cascade Caverns population,
is conspecific with E. tridentifera. Furthermore, a few
cave populations were determined genetically to be
hybrids between E. latitans and E. neotenes, and in two
cases individuals of both species were located within
the same cave.
As currently recognized, the Comal blind salaman¬
der ( Eurycea tridentifera) is a troglobitic species. It exhi¬
bits morphological modifications similar to the blind
salamanders of the Typhlomolge group, although gen¬
erally not to the same extreme extent (Fig. 2D). The sal¬
amander is found in caves of the Cibolo Creek basin
and south of the Guadalupe River, in the southeastern
part of the Edwards Plateau. These populations may
actually be cave populations of E. latitans that have
become troglomorphic. At the entrance to the type
locality (Floney Creek Cave), individuals intermediate
in morphology between E. tridentifera and the surface
species have been found.
The Valdina Farms salamander ( Eurycea troglodytes)
is likely a complex of several species found in the
southwestern part of the Edwards Plateau. Most popu¬
lations in this species complex are neotenic. However,
natural metamorphosis has been observed in popula¬
tions from several springs and caves. Two groups of
populations within this complex are troglomorphic,
with enlarged head, and reduced eyes and pigmenta¬
tion. The morphologically variable population from
the type locality (Valdina Farms Sinkhole) was for
some time considered a hybrid between E. neotenes
and E. tridentifera. This population is now extinct,
but analyses of others have revealed that E. troglodytes
is genetically distinct from both E. neotenes and
E. tridentifera.
Septentriomolge (Northern Salamanders)
Most populations of these spring-dwelling species
from north of the Colorado River were discovered after
1995. The few known prior to 2000, when these species
were formally described, had been considered periph¬
eral isolates of E. neotenes. Morphologically they resem¬
ble spring-dwelling species south of the Colorado River
(Blepsimolge), described above. They can be seen year-
round, most easily during the spring and summer
months, but small juveniles are rarely observed.
The San Gabriel Springs salamander ( Eurycea naufra-
gia) is known from springs and caves in the drainage
of the San Gabriel River, in the vicinity of Georgetown.
The adults have prominent eyes with melanophores
concentrated around them. The range of the Jollyville
Plateau salamander ( Eurycea tonkawae) is limited to
a few drainages on the Jollyville Plateau segment
of the Edwards Aquifer. This species makes
extensive use of subterranean aquatic habitat, espe¬
cially when surface spring flow decreases. Adults have
well-developed eyes, broad jaws, and blunt snouts.
The rare Chisholm Trail salamander ( Eurycea chishol-
mensis) is known only from springs at Salado.
Compared to E. tonkawae and E. naufragia, the eyes of
this salamander are reduced, and the head is flattened
and elongated.
Other Troglobitic and Troglophilic Eurycea
Several species of Eurycea with overlapping ranges
occur syntopically in caves and spring-fed headwaters
within the Ozark and Appalachian Highlands and the
Interior Lowlands of North America. These include
E. tynerensis, E. lucifuga, E. longicauda, and the troglobi¬
tic E. spelaea.
The grotto salamander ( Eurycea spelaea) is restricted
to the Springfield and Salem Plateaus in the Ozark
ENCYCLOPEDIA OF CAVES
SALAMANDERS
671
region of southern Missouri, extreme southeastern
Kansas, and adjacent areas in Arkansas and Oklahoma
(U.S.A.). DNA sequence divergence within E. spelaea is
relatively large, and some of the populations may rep¬
resent distinct species. This is the only known troglobi-
tic member of the genus that undergoes complete
metamorphosis. Adults are known only from caves,
but a large fraction of the larval population develops
in surface springs and streams. Here they displace the
larvae of sympatric Eurycea species to more ephemeral
parts of the stream and where predatory fishes and
crayfish are present. Both adults and larvae are sensi¬
tive to floods, which can extirpate a significant propor¬
tion of a population.
Adults grow to 7—14 cm in length and are white,
pinkish white, or light brown in color (Fig. 2G). The
eyes are reduced in size and structure, visible as raised
dark spots through the fused eyelids. The larvae are
tan dorsally, often spotted or mottled; those develop¬
ing inside caves can be pale to pink (Fig. 2H). They
have relatively small but functional eyes, which degen¬
erate in old larvae or during metamorphosis. The
extent of retinal degeneration in adults is related to
postmetamorphic age but there is variability in each
age group. In its terminal stage, reduction encom¬
passes the entire area of the retina and vision is lost.
Individuals raised in light maintain some pigmentation
or become pigmented. The larvae may retain vision
longer than those kept in darkness and the eyelids in
adults do not fuse.
Adults congregate in the main caverns where colonial
bats roost, most likely for the purposes of feeding and
breeding. They are most active during spring and sum¬
mer, when moisture levels are high and food is abun¬
dant. They may be found climbing moist vertical rock
walls outside water, but commonly return to the water
to hunt aquatic invertebrates. Besides aquatic and terres¬
trial invertebrates they also feed on bat guano (Fenolio
et al., 2006). In caves where colonial bats have declined,
emaciated individuals have been observed. Mating in
late spring is followed by oviposition through summer
and fall. Larvae metamorphose 2—3 years after hatching,
at the size of 5—12 cm, and mature soon afterward.
The sister species of E. spelaea, the troglophilic
Oklahoma salamander ( Eurycea tynerensis), is also
found on the Ozark Plateau. The species comprises
both strictly aquatic neotenic populations and more
terrestrial metamorphic populations. Neotenic animals
that inhabit caves are often pale, whereas those on the
surface retain normal larval pigment patterns. The
metamorphs rarely venture far away from water.
Usually they inhabit cool, moist habitats both on the
surface and in the twilight zone of caves. Larvae may
use the karst cavities to move within or between
streams.
The troglophilic cave salamander ( Eurycea lucifuga)
occurs in the broad region between southern Indiana,
northwestern Virginia, northern Alabama, and north¬
eastern Oklahoma (U.S.A.); this includes the karst
areas of the Ozarks, Interior Low Plateau, and the
Appalachians. The twilight zone of caves is the pre¬
ferred habitat of this species, although it can also be
found deeper into caves and, despite its name, outside
of caves in forested limestone ravines, springs, and
spring-fed streams. The species is broadly sympatric
with the long-tailed salamander ( Eurycea longicauda),
which is distributed throughout a similar area. This
species is also frequently associated with caves and
mines. These slender salamanders have very long tails
and grow to 18—19 cm. They are very conspicuous,
colored yellow or orange, and covered with
black spots (Figs. 3A,B). E. lucifuga has long limbs and
a prehensile tail used in climbing rock walls and
ledges. The eyes of this salamander are large and capa¬
ble of twilight vision. Adults can be found in caves
year-round, but their numbers fluctuate seasonally.
They aggregate there between spring and fall, when
surface temperatures are high. During this time the
females deposit their eggs in subterranean waters.
Larvae hatch after rains, when water flow increases
and more food becomes available (Ringia and Lips,
2007).
Georgia Blind Salamander (Haideotriton
wallacei)
The geographically isolated troglobitic Georgia blind
salamander ( Haideotriton wallacei ) has been found in a
few caves pertaining to the Floridan Aquifer below
Georgia, Florida, and Alabama (U.S.A.). At least two
historic localities have been destroyed by human activi¬
ties, but cave divers have since spotted this species in
the corresponding portion of the aquifer. This species is
sometimes included in the genus Eurycea.
Adults of this neotenic salamander measure 5—8 cm
in length and have little pigmentation, being mostly
pink or whitish with scattered melanophores on the
back and sides. The head is broad and the snout is
long, but not flattened. The limbs are slender and the
eyes are almost invisible, embedded in a mass of adi¬
pose tissue below the skin. Juveniles are slightly more
pigmented and the eyespots are visible.
Haideotriton wallacei can be seen in cave pools and
streams especially in caves where bats defecate over or
near the water. Salamanders move about slowly, rest¬
ing on bottom sediments or climbing on submerged
limestone sidewalls and ledges. Deeper in subterranean
tunnels the salamanders are much less common, pre¬
sumably because their food (benthic invertebrates) is
ENCYCLOPEDIA OF CAVES
672
s
FIGURE 3 A few salamanders frequently found in
caves. In eastern North America, E nrycea lucifuga (A)
and E. longicauda (B) frequently inhabit the same cave.
(C) Plethodon petraeus is associated with karst and
caves on Pigeon Mountain in Georgia. (D) Pseudotriton
ruber also frequently uses caves for reproduction
(Miller et at., 2008). (E) This undescribed species of
Paramesotriton inhabits pools in the twilight zone of
caves in China. Photos (A)—(C) by D. Fenolio ; (D)—(E)
by M. Niemilier.
scarce. Cave stream sediments have been found in the
digestive tracts of H. wallacei. One hypothesis attempting
to explain silt feeding holds that Haideotriton inten¬
tionally ingests the material to digest biofilm and micro¬
organisms in the sediments (the same has been
observed in young Proteus anguinus). Another hypothe¬
sis posits that the silt represents failed feeding attempts.
Spring Salamanders (Qyrinophilus)
Four North American species are currently recog¬
nized. Three species are troglobites, found in caves of
southern and central Appalachian Highlands (U.S.A.).
In Gyrinophilus, paedomorphosis is not universal and
has appeared after colonization of caves, although the
metamorphosing nontroglobitic species already has an
extremely long larval period.
The troglophilic spring salamander (Gyrinophilus
porphyriticus) is common in and around small streams and
springs in eastern North America. Next to Eurycea lucifuga
this is the most frequently encountered salamander in
caves. Both larvae and adults can be found in caves
throughout the Appalachian Highlands, even considerable
distances from the entrance. These large salamanders,
which can grow over 20 cm in total length, are notorious
for supplementing their invertebrate diet with conspecifics
and other salamanders. Metamorphosed animals are usu¬
ally orangish-red to salmon and turn darker with age
(Fig. 4B), but some individuals of cave populations are
pale. The eyes of cave-dwelling individuals do not differ
from those from surface populations. A much higher
occurrence of ingested debris found in cave-dwelling
individuals compared to individuals from surface
streams may indicate a lower feeding efficiency in the
dark. Larvae, which are also often pale in color, can
grow up to 16 cm. They metamorphose after 3—5 years
in surface populations but perhaps after 10 years or
more in cave populations.
ENCYCLOPEDIA OF CAVES
SALAMANDERS
673
FIGURE 4 Spring salamanders ( Gyrinophilus).
Metamorphosed adults of G. patteucus (A) and left in (B)
are pale and have smaller, degenerate eyes compared to
the troglophile G. porpliyriticus (right in (B)). The larvae
of G. subterraneus (C), G. palleucus palleucus (D), G. palleu¬
cus necturoides (E), and G. gulolinmtus (F) have small but
functional eyes. In addition, the numerous neuromast
mechanoreceptors on the head and flanks enable them
to detect vibrations in the water. Photos (A)—(B) and
(D)—(F) by M. Niemiller; photo (C) by D. Fenolio.
The three troglobitic species are believed to have arisen
independently from a single epigean ancestor similar to
G. porphyriticus as early as 2.5 million (Pliocene) to as
recently as 60,000 years ago (Pleistocene), but precise phy¬
logenetic relationships are obscured due to recent specia-
tion (Niemiller et ai, 2008, 2009). Nonetheless, other
evidence supports their independent origins and DNA
data indicate that divergence has occurred in the presence
of continuous or periodic gene flow between subterra¬
nean populations and their surface-dwelling progenitor.
The Tennessee cave salamander ( Gyrinophilus palleu¬
cus) is found in caves in the Central Basin, Highland Rim,
and Cumberland Plateau of central Tennessee, northern
Alabama, and extreme northwestern Georgia. Its range is
on the periphery of that of G. porphyriticus. Two subspe¬
cies of G. palleucus are recognized: the pale salamander
(Gyrinophilus palleucus palleucus; Fig. 4D), which inhabits
caves in the Lower Tennessee River watershed in
Alabama and Tennessee, and the big mouth cave sala¬
mander ( Gyrinophilus palleucus necturoides), found in the
Collins, Duck, Elk, and Stones River drainages in
Tennessee. This form is tan with dark spots dorsally and
laterally (Fig. 4E). The subspecies also differ slightly in
head width, leg length, eye size, and the number of trunk
vertebrae. Except for the lost ability for eye accommoda¬
tion (focusing), vision is not greatly impaired in G. palleu¬
cus (Besharse and Brandon, 1973). Exposure to light has
no effect on eye development and does not induce any
change in skin pigmentation.
G. palleucus inhabits sinkhole-type caves that are
rich in nutrients, which support its invertebrate prey
base. Besides cave streams, where it is easily observed,
the salamander is believed to inhabit subterranean
waters inaccessible to humans. It has occasionally been
found in springs outside caves. Little is known about
the life history of this species. Breeding is most likely
seasonal, occurring in late autumn or early winter.
This salamander is typically paedomorphic, although
metamorphosed individuals have been encountered
in at least four caves. Metamorphosed individuals of
ENCYCLOPEDIA OF CAVES
674
s
G. palleucus differ from adult G. porphyriticus in their
pale skin, gaunt appearance, longer and narrower
snout and smaller eyes (Fig. 4A,B) and in retaining the
larval characteristic of an undivided premaxillary bone.
The Berry Cave salamander ( Gyrinophilus gulolineatus)
is known from caves in the Ridge and Valley of eastern
Tennessee. Its range is entirely contained within the range
of G. porphyriticus and the two species sometimes occur in
the same cave (Miller and Niemiller, 2008). Like G. palleu¬
cus, G. gulolineatus is usually paedomorphic and generally
resembles the former (Fig. 4F), but can be distinguished
from it by having a darker pigmentation, wider head,
more spatulate snout, and by attaining a greater adult
size. The eyes are small and degenerate, comparable in
size to the eyes of G. palleucus. Unlike G. palleucus, how¬
ever, transformed individuals have fully divided premax¬
illae, in which they resemble G. porphyriticus.
The West Virginia spring salamander ( Gyrinophilus
subterraneus) is known only from General Davis
Cave in West Virginia where it co-occurs with
G. porphyriticus. Unlike G. palleucus and G. gulolineatus,
G. subterraneus regularly metamorphoses, although at
an exceedingly large size (up to 18 cm). Large larvae
have fully developed, mature gonads. Both larvae and
adults have a light brownish-pink skin color, overlaid
dorsally and laterally by a darker reticulate pattern
(Fig. 4C). The head of larval G. subterraneus is broader
than the heads of either G. porphyriticus or G. palleucus,
but does not have the spatulate snout typical of
G. gulolineatus. Metamorphosed adults retain fused
premaxillae (as in G. palleucus). Both larvae and adults
have smaller eyes than G. porphyriticus and, unlike
G. palleucus, only weak visual perception.
Although G. porphyriticus also occurs in General
Davis Cave, only adults of this species have been found
(Niemiller et al., 2010). G. subterraneus has been observed
as far as 2 km into the cave; however, in the first 300 m
of the stream passage, predominantly larvae have been
encountered, suggesting that its main habitat may be
located deeper into the cave. The larvae prefer shallow,
calm pools and are found in water depths of 1—20 cm.
Adults can be found in shallow water as well, whereas
G. porphyriticus almost never enters the water. The
banks of the cave stream contain thick deposits of
decaying leaf litter, washed into the cave by floods. This
leaf litter is the source of nutrients for cave invertebrates
that adults of both species putatively prey on.
DIVERSITY PATTERNS OF
SALAMANDERS FOUND IN CAVES
Including the salamanders described above, over 90
species (approximately 15% of all known salamander
species) belonging to five families have been reported
from natural or manmade subterranean environments.
Their dependence on the cave environment varies not
only between species, but also among populations of
the same species.
Over half of all species found in caves are reported
from eastern North America. Several woodland sala¬
manders ( Plethodon sp.), especially Plethodon albagula,
P. glutinosus, P. dorsalis, and P. petraeus (Fig. 3C) are fre¬
quently associated with caves. Also reliant on caves are
the red salamander ( Pseudotriton ruber-, Fig. 3D), a few
dusky salamander species ( Desmognathus sp.), and the
green salamander ( Aneides aeneus). To the south, in
the Mexican Sierra del Madre Oriental and Sierra Madre
del Sur, two splayfoot salamanders ( Chiropterotriton sp.)
and two false brook salamanders ( Pseudoeurycea sp.)
occur in caves. Of the salamanders that occur in western
North America, the web-toed salamanders ( Hydromantes
sp.) in California, especially the shasta salamander
(Hydromantes shastae), are often found in caves.
Apart from the plethodontids, the mole salamanders
(Ambystoma sp.) of the family Ambystomatidae are reg¬
ularly encountered in caves, particularly in the region
from Arizona to Alabama and Tennessee. Neotenic
cave-dwelling populations of the barred tiger salaman¬
der (Ambystoma mavortium) have been reported from
New Mexico and in gypsum caves of northeastern
Oklahoma.
Europe's salamander fauna is dominated by the
family Salamandridae, but seven species of the pletho¬
dontids also occur here. The European cave salaman¬
ders (Speleomantes sp. and Atylodes sp.) of Sardinia,
northwestern mainland Italy, and southern France are
closely related to the Californian Hydromantes and
share many life history features with them. They are
strongly associated with caves, and will aggregate in
the twilight zone. Although they have prominent eyes,
they can also detect prey in darkness using olfactory
information. Several surface species of the family
Salamandridae frequent natural or manmade subterra¬
nean refugia in Europe, especially in the arid circum-
Mediterranean region. The widespread fire salamander
(Salamandra salamandra) not only seeks refuge in caves
but also frequently favors them as reproduction sites
even when surface water is available nearby. The
Pyrenean mountain newt (Calotriton asper) inhabits
streams, lakes, and ponds along the Pyrenees Mountains
of France and Spain, but is also found in subterranean
waters, especially on the periphery of its range. Except
for one pale-colored cave population, no morphological
modifications have been described in this species.
Instead, modifications related to cave life involve the
reproductive cycle and its periodicity, including pro¬
longed or continuous gamete production, slower egg
deposition, and delayed metamorphosis. Neotenic indi¬
viduals have also been found.
ENCYCLOPEDIA OF CAVES
SALAMANDERS
675
Despite extensive karst areas in Central America
and Asia, and great salamander diversity (especially in
Central America), the number of salamanders that
inhabit caves and other subterranean habitats is low.
The caves on the Yucatan Peninsula in Mexico and
adjacent Guatemala are the temporal habitat of the
plethodontid Yucatan salamander ( Bolitoglossa yucata-
na ), (perhaps) the nimble long-limbed salamander
(Nyctanolis penix), and potentially two more species.
In western Asia, only one species, of the family
Hynobiidae, has been found in a cave. The almost fully
aquatic Gorgan mountain salamander ( Pamdadylodon
gorganensis) is known from two localities in the Alborz
Mountains of northern Iran, one of which is a cave. In
eastern Asia, another hynobiid species, the oki salaman¬
der ( Hynobius okiensis) from the Dogo Island of Japan, is
believed to reproduce in hypogean waters. The salaman-
drid wart newts ( Paramesotriton sp.) have been observed
in the twilight zone of caves in Guizhou Province, China.
CONCLUSION
Very few salamanders are obligate cave-dwellers,
which is in contrast to the great number of species that
temporarily inhabit caves. Ranges of troglobitic salaman¬
ders are usually small, limited by hydrological or geo¬
logical barriers. Due to the limitations such as
inaccessibility or rarity, only the more common species
have been studied in detail. A fair amount of attention
has been given to their phylogenetic relationships, obser¬
vations on regressive morphological modifications asso¬
ciated with cave life, development, and behavior, but
limited knowledge exists of their life histories and ecol¬
ogy. Continuing research in these areas is essential for
gaining insight into the origins and processes involved
in evolution of obligate subterranean existence.
Because of suspected small population sizes, limited
distribution, and high specialization, troglobitic sala¬
manders are threatened by habitat and water quality
degradation, caused mainly by urbanization, agricul¬
ture, and deforestation. Their position in the food web
and the tendency to accumulate energy reserves in
their bodies make them particularly vulnerable to
the effects of toxic chemicals in contaminated water.
Most troglobitic salamanders are legally protected.
However, in the face of increasing demands for
groundwater, pollution, and alterations to the land¬
scape above cave systems, their numbers continue to
decline. Not only troglobitic species but also all sala¬
manders that use caves during part of their life history,
are important components of subterranean ecosystems
and greatly depend on the preservation of these
special habitats.
Bibliography
AmphibiaWeb: Information on amphibian biology and conservation
(2011). Berkeley, CA: AmphibiaWeb. Available: <http://amphi-
biaweb.org>.
Bendik, N. F., Gluesenkamp, A. G., & Chippindale, P. T. (2009). The
biogeography and rapid radiation of central Texas neotenic sala¬
manders. Proceedings of the 15th International Congress of Speleology,
1, 219-225.
Besharse, J. C., & Brandon, R. A. (1973). Optomotor response and eye
structure of the troglobitic salamander Gyrinophilus palleucus.
American Midland Naturalist, 89, 463—467.
Chippindale, P. T., Gluesenkamp, A. G., & Bendik, N. F. (2009).
Texas cave and spring salamanders ( Eurycea ): New discoveries
and new surprises. Proceedings of the 15th International Congress of
Speleology, 1, 227.
Epp, K. J., Gonzales, R., Jr., & Gabor, C. R. (2010). The role of water¬
borne chemical cues in mediating social interactions of the Texas
blind salamander, Eurycea rathbuni. Amphibia-Reptilia, 31(2),
294-298.
Fenolio, D. B., Graening, G. O., & Stout, J. F. (2006). Coprophagy in a
cave-adapted salamander; the importance of bat guano examined
through stable isotope and nutritional analyses. Proceedings of the
Royal Society B, 273(1585), 439-443.
Grossenbacher, K., & Thiesmeier, B. (Eds.), (1999). Handbuch der
Reptilien und Amphibien Europas. Band 4/1: Schwanzlurche (Urodela)
I. Wiesbaden: AULA (in German).
Lannoo, M. (Ed.), (2005). Amphibian declines: The conservation status of
United States species. Berkeley: University of California Press.
Lucas, L. K., Gompert, Z., Ott, J. R., & Nice, C. C. (2009). Geographic
and genetic isolation in spring-associated Eun/cea salamanders
endemic to the Edwards Plateau region of Texas. Conservation
Genetics, 10, 1309—1319.
Miller, B. T., & Niemiller, M. L. (2008). Distribution and relative
abundance of Tennessee cave salamanders ( Gyrinophilus palleucus
and G. gulolineatus ) with an emphasis on Tennessee populations.
Herpetology Conservation and Biology, 3(1), 1—20.
Miller, B. T., Niemiller, M. L., & Reynolds, R. G. (2008). Observations
on egg-laying behavior and interactions among attending female
red salamanders ( Pseudotriton ruber ) with comments on the use of
caves by this species. Herpetological Conservation and Biology, 3(2),
203-210.
Najvar, P. A., Fries, J. N., & Baccus, J. T. (2007). Fecundity of San
Marcos salamanders in captivity. Southwestern Naturalist, 52(1),
145-147.
Niemiller, M. L., Fitzpatrick, B. M., & Miller, B. T. (2008). Recent
divergence-with-gene-flow in Tennessee cave salamanders
(Plethodontidae: Gyrinophilus) inferred from gene genealogies.
Molecular Ecology, 17(9), 2258-2275.
Niemiller, M. L., Miller, B. T., & Fitzpatrick, B. M. (2009). Systematics
and evolutionary history of subterranean Gyrinophilus salaman¬
ders. Proceedings of the 15tli International Congress of Speleology, 1,
242-248.
Niemiller, M. L., Osbourn, M. S., Fenolio, D. B., Pauley, T. K.,
Miller, B. T., & Flolsinger, J. R. (2010). Conservation status and
habitat use of the West Virginia spring salamander ( Gyrinophilus
subterraneus) and spring salamander (G. porphyriticus) in General
Davis Cave, Greenbrier Co., West Virginia. Herpetological
Conservation and Biology, 5(1), 32—43.
Ringia, A. M., & Lips, K. R. (2007). Oviposition, early development
and growth of the cave salamander, Eurycea lucifuga: Surface and
subterranean influences on a troglophilic species. Herpetologica, 63
(3), 258-268.
Safi, R., Vlaeminck-Guillem, V., Duffraisse, M., Seugnet, I.,
Michelina Plateroti, M., Margotat, A., et al. (2006).
ENCYCLOPEDIA OF CAVES
676
s
Pedomorphosis revisited: Thyroid hormone receptors are functional
in Necturus maculosus. Evolution and Development, 8(3), 284—292.
Trontelj, P., Goricki, S., Polak, S., Verovnik, R., Zaksek, V., & Sket, B.
(2007). Age estimates for some subterranean taxa and lineages in
the Dinaric Karst. Acta Carsologica, 36, 183—189.
SALTPETRE MINING
David A. Hubbard Jr.
Virginia Speleological Survey and Virginia Department of
Mines, Minerals and Energy
DEFINITIONS AND IMPORTANCE
Historically, saltpetre is one of the most strategic of com¬
modities. It occurs naturally in caves and rockshelters,
but it is rare. A suite of related nitrates occurs in many
caves. The mining and processing of cave nitrate-
enriched sediments is a relatively simple endeavor,
although labor intensive. The tendency of these sedi¬
ments to contain a suite of nitrates rather than just potas¬
sium nitrate is one reason the archaic spelling saltpetre is
used in reference to the mining of cave nitrates and the
caves in which they occur. This convention is followed
throughout this article.
The invention of gunpowder revolutionized weap¬
onry and warfare. Gunpowder, also referred to as black
powder, consisted of a mixture of saltpetre, sulfur, and
charcoal. Although saltpetre was used in the preserva¬
tion of meats, the greatest historic demand for saltpetre
was during times of insurrection and war. Nowhere
has the quest for saltpetre contributed to historic events
more than in the United States, where this commodity
contributed to both the formation of a country and
almost its destruction.
The mineral niter (synonym, saltpeter) is potassium
nitrate (KNO 3 ). Like many other nitrate compounds,
niter is deliquescent; that is, it has a natural tendency
to draw water to itself and dissolve into a solution.
Although deliquescent minerals can absorb moisture
from humid air, they occur naturally in sheltered loca¬
tions under conditions of low humidity or during peri¬
ods of reduced humidity. The deliquescent nature of
saltpeter is the reason for the old warning of soldiers
and frontiersmen, who depended on their firearms for
survival, to "keep your powder dry!"
Caves and rockshelters, also termed rockcastles, are
locations where nitrates may accumulate. Analyses of
cave sediments, which were mined historically for salt¬
petre, commonly reveal no nitrate minerals. The reason
is that most of the classical saltpetre caves are in regions
where the humidity typically is too high for niter and
the even more deliquescent minerals nitromagnesite,
Mg(N0 3 ) 2 • 6H 2 0, and nitrocalcite, Ca(N0 3 ) 2 ■ 4H 2 0, to
crystallize into their solid mineral forms. Instead, the
saltpetre-rich sediments, historically termed petre dirt,
contain concentrated viscose nitrate solutions in the
form of sediment moisture.
Because nitrate minerals rarely crystallize in most of
the known saltpetre caves, other clues to the accumula¬
tion of nitrates in cave sediments are important. In the
absence of niter, the foremost evidence of significant
nitrate concentrations in cave sediment is the presence
of efflorescent crusts. These white or light-colored pow¬
dery crusts commonly are composed of a mixture of
soluble salts and minerals, such as gypsum and calcite,
that accumulate on cave sediment and rock surfaces as
a result of evaporation. Efflorescent incrustations sig¬
nify locations where periodic atmospheric conditions
allow evaporation and the concentration of the minute
amounts of dissolved solids in interstitial soil and rock
moisture. Precipitation and concentration drive the
wicking action of the dissolved solids through soil and
rock pores from their respective remote sources. The
sources of most saltpetre cave nitrates are the surface
ecosystems overlying saltpetre caves.
Recent microbiological work in caves and karst has
shown that bacteria are important in cave development
(s peleogenesis) and in the development of the secondary
cave mineral forms ( speleothems ) that were thought to
be the result of physiochemical reactions (Taylor,
1999). The importance of nitrifying and other bacteria
in the accumulation of efflorescent crusts and nitrate
accumulations in sediments (petre dirt) is unknown
but probably is not trivial.
SALTPETRE MINING
The mineral niter (KN0 3 ) is rarely found in caves,
but when observed it occurs as clear to white lint-like
fibers, acicular (needle-like) crystals, powder, crusts,
coralloid, or flowstone forms. The most extensive form
observed in saltpetre caves is the lint-like fibers that
occur in dense carpets on bedrock walls (Fig. 1) and
sediment-covered walls and floors. These niter fiber
occurrences can be harvested with the use of a thin
wooden spatula or paddle-like scraper, leaving little or
no evidence of extraction. Such a wooden scraper
was observed high on a Virginia saltpetre cave ledge
before this author had observed an efflorescent niter
occurrence.
The majority of the documented saltpetre caves
does not normally contain crystalline niter. At the
humidities typically found in these caves, the deliques¬
cent nitrate accumulations occur as viscose nitrate
solutions in efflorescent crusts on rock and sediment
Encyclopedia of Caves.
© 2012 Elsevier Inc. All rights reserved.
SALTPETRE MINING
677
(B)
FIGURE 1 (A) Fibrous form of the mineral saltpeter carpeting
cave walls in a Virginia saltpetre cave (photograph by John C. Taylor;
image width approximately 12 cm). (B) Fibrous form of the mineral
saltpeter carpeting cave surfaces in a Virginia saltpetre cave; scale is
in centimeters.
surfaces and within sediments and bedrock proximal
to the cave.
The earliest mining of saltpetre may have been in
China, dating from the development of gunpowder
approximately 1000 years ago. Although the discovery
of the effects of niter in the preservation of meat may
predate gunpowder, no evidence is known that niter
was extracted from caves for preservation of meat
remote from caves. The importance of gunpowder in
revolutionizing armed conflict resulted in intense
periods of demand for niter. In addition to mining in
rockshelters and caves, saltpetre was obtained from
artificial niter beds and from soils collected under
buildings. In Western societies, the earliest docu¬
mented search for niter is the 1490 quest of Hans Breu
for saltpetre in Sophienhohle (Sophie's Cave) in
Germany. The French Revolution and Napoleonic
wars of Europe (1792—1815) and the Revolutionary
(1775-1783), 1812 (1812-1815), and Civil (1861-1865)
wars of North America were periods of intense
demand for saltpetre. The written record of actual salt¬
petre mining processes is sparse, as is typical for many
mundane tasks. The most extensive known distribu¬
tion of saltpetre caves and the best-preserved evidence
of mining can be found in the southeastern United
States. Although the mining evidence in each saltpetre
cave is as different as caves are from one another,
similar patterns of marks, disturbances, and artifacts
convey some information about basic mining and pro¬
cessing methodologies.
Tangible evidence of saltpetre mining includes the
principal physical evidence of wall and floor sediment
removal, the secondary evidence of lighting and names
and artifacts used in mining or as bracing that date to
the mining era, evidence of simple sediment proces¬
sing and separation of rocks and clay balls, and modi¬
fied pathways and conveyances of miners and mined
sediment. Nitrate processing evidence is important in
distinguishing saltpetre mining from sediment extrac¬
tion for other purposes and is discussed in the section
on saltpetre processing.
Principal Physical Evidence of Mining
Physical evidence of sediment removal includes tool
marks, wall discolor marks of former sediment levels,
and tunnels and pits within sediments. Tool marks
on worked sediment faces are of wide- and narrow-
bladed tools. Sediment mining tool marks predating
the 1860s are typically hoe-like and wide bladed (10 to
13 cm), while many of those of the American Civil
War era are narrow bladed (5 to 7 cm) and attributed
to mattocks. The blades of these tools were metal, and
rarely were such tools left in caves, while pointed dig¬
ging and prying sticks and paddle-like spatulas are
more commonly found. Pick and mattock impact
marks on rock walls may relate to efflorescent wall
crust or sediment removal. The discoloration marks of
old sediment levels on bedrock walls provide an
ENCYCLOPEDIA OF CAVES
678
s
indication of the thickness of mined nitrate-rich sedi¬
ment. An important distinction is that not all sediment
in a saltpetre cave is equally enriched or desirable.
Slightly worked deposits of difficult to process clay
sediments with well-preserved mattock marks and
well-developed white efflorescent crusts are evidence
that the difficulty in extracting nitrates from clay out¬
weighs the effort. Similarly, pits and tunnels through
thick sediments may indicate low yield deposits, which
were partially exploited, but such excavations may have
had more utility as passage and haulage routes.
Secondary Evidence of Mining
Indirect evidence of mining includes soot stains
above torch perches and the stubs of pine faggots
(torches) at worked sediment faces and along well-
traveled routes between entrance and mining areas.
Rock piles in wall niches or backfilled in small pas¬
sages are evidence of hand-sorting of rock from the
sediment. Less common are piles of small rocks and
clay balls from sieving near the site of sediment min¬
ing or near locations where sediment was transported
in cloth or burlap sacks. The combination of old sedi¬
ment level stains and names dating to the mining era
stranded high on walls, as well as remote clusters of
tally and torch marks, may differentiate periods of
mining. Mining artifacts include pointed wooden dig¬
ging and prying sticks, wooden scrapers, and wood
hoops for holding sacks open during filling. Stone and
wood bracing stabilize undermined large rocks in
some saltpetre caves. In breakdown mazes, the use of
rock for slab and ceiling stabilization may not be obvi¬
ous. The breakup of rock slabs to access the underlying
sediment has resulted in labor-intensive passage
enlargement and rock walls and fills that may appear
as having other significance. Other modified pathways
and conveyances include cut steps in steep sediment
banks, stone steps, wooden stairs, boardwalks, and
plank ramps across chasms and canyons; various types
of ladders; and windlasses.
Perhaps the most rewarding evidence of saltpetre
mining is the correlation of names etched into cave
walls with written mine-era records of payrolls, equip¬
ment receipts, or saltpetre production or sales receipts.
Such corroboration sometimes allows the matching of
different historic cave names with present-day names
as well as verification of the nature of the sediment
mining.
SALTPETRE PROCESSING
In some saltpetre caves, evidence of the processing
is intermingled with the mining evidence. Large
hoppers or smaller vats were commonly used in salt¬
petre processing, but barrels and sections of hollow
logs also were employed. Two of the best-preserved
examples of sediment processing equipment are dis¬
played on cave tours: rectangular hoppers in
Mammoth Cave in Kentucky and V- or wedge-shaped
vats in Organ Cave in West Virginia. Intact hoppers
and vats are rare in most present-day saltpetre caves.
All that remains in many saltpetre caves are the inter¬
nal casts of nitrate-depleted mine spoils. Surrounding
such casts occur the discarded piles of leached cave
sediments, discarded from the vats prior to their last
loads of sediment. The wooden remains of many hop¬
pers and vats have rotted or been destroyed by
vandals.
Processing requires a source of water, which is not
usually located close to the nitrate-rich cave sediments
that require at least a periodic reduction in the
humidity to accumulate. Processing of the majority of
saltpetre cave sediment has entailed the puddling of
the nitrate-rich sediments in a hopper or vat. After the
nitrates are leached from the cave sediments, the
liquid, termed beer, is decanted. To this beer another
leachate, from wood ashes (potash), is added until a
white precipitate ceases to form from the mixing of the
two solutions. This process exchanges the potassium
cation of the potash solution for the calcium and mag¬
nesium cations of the cave nitrate solution. The result¬
ing potassium-rich nitrate solution is transported to
large iron kettles for further processing and eventual
fractional crystallization of the niter crystals. The ket¬
tles are usually hemispherical and without legs or
other attachment points that would serve as heat sinks,
resulting in differential heating within the kettle.
Most kettles have a wide lip or rim by which they are
supported over a heat source. The fractional crystalli¬
zation process enables workers to selectively crystallize
specific soluble salts, while leaving more soluble salts
in solution.
The processing of sediment, especially for smaller
caves, occurred outside of many saltpetre caves for
lack of a suitable water source. Some small caves were
worked as satellite locations to a nearby major saltpe¬
tre works. No evidence of kettle processing is known
from within any saltpetre cave, although kettles are
known to have been hidden in and recovered from
caves.
Perhaps the most impressive examples of evidence
of in-cave saltpetre processing are preserved at
Mammoth Cave in Kentucky. Some of these War of
1812 era workings may be observed on tours
and include large hoppers and a complex system of
hollow log plumbing, whereby freshwater was
pumped into and saltpetre leachate was pumped out
of the cave.
ENCYCLOPEDIA OF CAVES
SCALLOPS
679
The processing of the nitrate deposits of the sand¬
stone rockcastles of eastern Kentucky is abbreviated.
The nitrates from these shelter caves are typically
potassium nitrate and do not require the ionic
exchange step required for the calcium- and magne¬
sium-rich nitrates of the carbonate solutional caves.
Evidence of saltpetre processing in caves is impor¬
tant in establishing that the mining evidence is of salt¬
petre mining. Cave sediments have been excavated
from caves worldwide for uses other than saltpetre
extraction.
CONTRASTING SALTPETRE MINING
WITH OTHER CAVE SEDIMENT
EXTRACTION
The cave sediments of many European caves are
enriched not only in nitrates but also phosphate as a
result of the bones of mega fauna, which used these
caves during the Ice Ages of the Pleistocene. Locally,
nitrate- and phosphate-enriched sediments have been
extracted from European caves for fertilizing gardens.
In China, cave sediments have historically been
worked for fertilizer and for bones to be used in the
apothecary trade. Marketed as "dragon bones," these
fossil remains have been processed for use in folk rem¬
edies and as aphrodisiacs. Bat guano has been utilized
as fertilizer in numerous areas within the United
States, but the extraction of other nitrate-rich cave
sediments for garden use is also known from at least
one U.S. locality. It is likely that European immigrants
to the United States continued the time-honored
traditional exploitation of enriched cave sediments as
garden fertilizer at other southeastern U.S. sites.
Evidence of the extraction of cave sediments without
associated saltpetre processing evidence, written his¬
toric records of saltpetre mining, or local saltpetre min¬
ing lore may represent other uses of cave sediments,
such as for garden fertilizer, chinking for a log home,
ceramics, fossil or artifact pilferage, or other usage.
In summary, saltpetre mining is a historic extractive
industry tied to the development and usage of black-
powder-charged firearms. The archaic spelling of the
mineral, saltpetre, is retained as a descriptor because,
in most cases, this mineral has only rarely been
encountered in saltpetre caves; rather, a suite of the
deliquescent nitrates typically occurs as viscous solu¬
tions within cave sediments, which were mined and
the nitrates extracted by leaching and chemically con¬
verted to the crystalline commodity, potassium nitrate
(saltpeter), the major constituent of black powder.
Bibliography
DePaepe, D. (1985). Gunpowder from Mammoth Cave: The saga of
saltpetre mining before and during the War of 1812. Hayes, KS: Cave
Pearl Press.
Duncan, M. S. (1997). Examining early nineteenth century saltpeter
caves: An archaeological perspective. Journal of Cave and Karst
Studies, 59(2), 91-94.
Faust, B. (1964). Saltpetre caves and Virginia history. In H. H.
Douglas (Ed.), Caves of Virginia (pp. 31—56). Falls Church, VA:
Virginia Cave Survey.
Hill, C. (Ed.), (1981). Saltpeter. National Speleological Society Bidletin,
43(4), 83-133.
Hill, C., & Forti, P. (1997). Cave minerals of the world (2nd ed.).
Huntsville, AL: National Speleological Society.
Shaw, T. R. (1992). History of cave science. Sydney, Australia: Sydney
Speleological Society.
Smith, M. O. (1990). Saltpeter mining in East Tennessee. Maryville, TN:
Byron's Graphic Arts.
Taylor, M. R. (1999). Dark life. New York: Scribner.
SCALLOPS
Phillip J. Murphy
University of Leeds, UK
INTRODUCTION
Studies of the solutional sculpturing of cave walls can
provide information on both the direction and dis¬
charge of water flow in a cave passage. Scallops are the
most well-studied type of solutional sculpturing. Their
asymmetry indicates the direction of groundwater flow
and their wavelength are inversely proportional to the
flow velocity. Laboratory and field investigations have
enabled the calculation of mean flow velocity from
scallop wavelength data and thus the calculation of dis¬
charge at the time of scallop formation.
SCALLOPS AND FLUTES
Scallops are asymmetrical, cuspate, oyster-shell¬
shaped dissolution depressions in cave walls. The term
scallop was first proposed by Coleman (1949) to replace
the term flute used prior to this. The term flute has
been used for the elongate, nearly parallel crested
forms seen in some cave passages. They are scallops of
infinite width and are much rarer than true scallops.
Vertical grooves seen in vadose shafts formed by water
streaming down in thin sheets have also been called
flutes.
Encyclopedia of Caves.
© 2012 Elsevier Inc. All rights reserved.
680
s
GROUNDWATER FLOW DIRECTION
Scallops are asymmetrical in the direction of flow,
with a smooth slope on the downstream side of
the scallop and a steep cusp on the upstream side.
Coleman (1949) realized they could be used as a quick
and simple indicator of the direction of flow in a con¬
duit. They occur in packed patterns on cave floors,
walls, and ceilings. They are measured from cusp to
cusp and can vary from 1 cm to several meters in
length (where they become difficult to distinguish
from bends in the passage walls).
The simplest way to ascertain the direction of asym¬
metry in a scallop population is to shine a light along
the scalloped surface and then to position yourself so
that you are looking directly toward the scallops. If the
light is shining upstream, the steep slopes of the scal¬
lops are brightly illuminated and the shallow slopes
are in darkness, resulting in the majority of each scal¬
lop not being illuminated (Fig. 1). If the light is shining
downstream, the steep slopes are in darkness and the
shallow slopes are brightly illuminated, resulting
in the majority of each scallop being illuminated. If
the scallop asymmetry is not sufficiently distinct,
their profiles may be transferred to paper using a pro¬
file template and analyzed graphically. A large
sample of scallops can then be analyzed statistically
for asymmetry.
GROUNDWATER FLOW VELOCITIES
The size of the scallops is inversely proportional to
the flow velocity of the water that formed them.
Scallop populations usually form unimodal well-
sorted, log-normal distributions. In the right
FIGURE 1 Scallops illuminated with the light source pointing
upstream. Joint Hole, North Yorkshire, UK. Photograph taken by the
author.
FIGURE 2 Fluid motion in the vicinity of a scallop: (A) detach¬
ment of the laminar boundary, (B) transition to turbulence, (C) the
lee eddy, and (D) flow reattachment. After Curl (1974).
conditions, patterns of scallops of characteristic length
cover all available surfaces. They are the stable form of
these surfaces in the prevailing conditions and can be
considered as solutional analogs to current ripples in
unconsolidated sediments.
Curl (1974) studied the process of scallop and flute
formation and was able to show that they are a
hydraulic phenomenon. Curl proposed that scallops
are formed where there is detachment of the saturated
boundary layer at a specified Reynolds number, occur¬
ring in the subcritical turbulent flow regime. Where
the main flow separates, it passes above an area of
slower, recirculating flow. Within a short distance the
fast flow becomes irregular and turbulent. Because the
turbulence produced causes mixing between the fast
stream and the lee eddy, fluid is entrained out of the
lee eddy, causing the jet to turn toward the surface
and reattach (Fig. 2). Detachment allows aggressive
waters to come into contact with the bedrock, increas¬
ing the rate of direct erosion. The rate of solution is the
highest in the area of reattachment where the turbulent
flow impinged the most on the surface. This means
scallops migrate downstream. The more the frequency
of detachment, the more the flow velocity, hence
reducing the erosion length available to each individ¬
ual scallop.
Being a hydraulic phenomenon, scallop formation
and the conduit flow conditions that gave rise to them
are described by fluid dynamic equations. Previous
work has shown that scallops form at a stable scallop
Reynolds number (Re*) of ~2200 where Re* is related
to the mean boundary shear velocity, u*, the mean
scallop wavelength A, fluid density p t -, and fluid
dynamic viscosity p by
R
Measuring a statistically viable sample allows A to
be calculated, and thus an estimate of u* to be made
using Eq. (1). A water temperature appropriate for the
study area needs to be chosen. Then a modified
Prandtl's universal velocity distribution equation for a
parallel-sided conduit [Eq. (2)] or a circular conduit
ENCYCLOPEDIA OF CAVES
SCALLOPS
681
[Eq. (3)] can be used to calculate the mean flow veloc¬
ity at the time of creation of the scallops:
2 -n i,, (s)- , j +B L
2.5<fln(^) -l4+B L
( 2 )
(3)
where d is the hydraulic diameter of the conduit and
B l is Prandtl's bed roughness constant. Blumberg and
Curl (1974) showed through experimental flume work
on scallops that B L = 9.4.
An approximation to the groundwater flow velocity
(in meters per second) indicated by a scallop popula¬
tion can be calculated by dividing 3.5 by the mean
length of the scallops in centimeters.
CALCULATION OF DISCHARGE
AND OTHER PARAMETERS
Following the approach of Gale (1984), further
hydraulic parameters can be estimated from the values
obtained above and the following equations:
(m/m*) 2
Q = ua
Re =
udp^
P
r = (u*) 2 p f
P = UT
Wmax = 0.96 U*
where / is the Darcy—Weisbach friction factor, Q the
discharge, a the cross-sectional area of the conduit. Re
the conduit Reynolds number, F the Froude number, g
the gravitational acceleration, r the mean boundary
shear stress, P the power of flow per unit area of
boundary, and w max the maximum settling velocity of
material in suspension.
USES AND PITFALLS OF SCALLOP
DISCHARGE DATA
Curl's theory of scallop formation has been tested in
laboratory studies and in the field both in vadose and
phreatic conduits. Studies in network cave systems
have shown that paleodischarges calculated from
y 0.057
0.40
+ 0.11
■ 0.00
0.36
+ 0.08
-0.07
ujd -Q07
0.12
<-
+ 0.03
-0.03
0.31
+ 0.010
-0.07
/ 0.077*q Q2
+0.09
-0.07
FIGURE 3 Flow through the junctions (m 3 s 5 with estimates of
errors of flow. From Lauritzen (1982).
scallop data fit the requirements of the continuity
equation, whereby the same amount of incompressible
liquid has to pass through each cross-section of a tube
or tube system (Fig. 3).
Working in relict conduits where enough scallop
lengths can be measured, the paleovelocity and paleo-
discharge of the conduit can be calculated by assuming
a temperature of the water and measuring the conduit
dimensions. A number of other factors, however, have
to be considered. One problem is "Under what flow
conditions in the conduit does scallop formation
occur?" Field studies in phreatic conduits have shown
that the velocities indicated by the mean scallop length
(the scallop dominant discharge) correspond to the
upper 5% of the flow regime (Fig. 4). For vadose con¬
duits, however, a scallop dominant discharge is diffi¬
cult to define because of the unconstrained flow depth.
A complex relationship is seen to exist between the
velocity and depth of flow as discharge changes
(Charlton, 2003). Thresholds occur over discrete depth
ranges where there is little or no change in velocity;
these are observed during both rising and falling
stages. These thresholds may be related to changes in
hydraulic radius, and hence flow resistance at different
depths of flow.
Scalloped surfaces exhibit a number of small
depressions, which appear to be related to the intersec¬
tion of the rims of the scallops and therefore not
related to the flow velocity. The use of the Sauter
mean (A 32 ) rather than the arithmetic mean is used by
ENCYCLOPEDIA OF CAVES
682
s
% S 10 20 30
FIGURE 4 Flow duration curve for the period from September
1983 to April 1985 with the flow rate deduced from scallops for the
underground outlet for Lake Glomdal, Norway. The flow rate
deduced from scallops represents only 2% or less of the time of flow.
From Lauritzen et al. (1985).
some workers to suppress the importance of these
smaller features.
n
A32 = -
£A ?
i=l
Being hydraulic phenomena, the process of scallop
formation is independent of substrate; however, field
observation has shown that bedrock variation does
exert some control on scallop size. Homogeneous rock
with a uniform grain size and lack of small cavities
provides the best substrate for development. Scallops
are rarely developed in dolomite caves.
Studies of scallop distributions in an active phreatic
conduit have shown that distinct and adjacent scallop
distributions indicative of very different flow regimes
can coexist (Murphy et al., 2000). Contrast in size
between ceiling and lower wall scallop populations in
relict conduits have been explained by a partial draining
of the conduit with a corresponding change in discharge.
Such a situation could occur in the epiphreatic zone.
Workers in relict conduits have often assumed that the
presence of small scallops on the floor contrasting with
large scallops on the walls and ceiling is due to invasion
of the relict conduit by a vadose stream. This may not be
the case and the contrasting scallop sizes could be a
product of the phreatic phase of the conduit's history.
FIGURE 5 A lateral pothole on Buckeye Creek Cave, West
Virginia. Photograph courtesy of G. Springer.
Sims (2004) compared mean velocities required to
entrain sediment particles and to produce scallops.
Velocities which would entrain sand grains of I-mm
diameter would produce scallops with a wavelength
of approximately 10 cm. Scallop formation still occurs
at velocities two orders of magnitude below the mini¬
mum velocity required to entrain any particles.
The occurrence of scalloping superimposed on spe-
leothems means the scallops must postdate initial
draining of the conduit and initiation of speleothem
growth. Such occurrences allow for the possibility of
constraining the timing of paleohydraulic events by
dating the speleothems. This approach has been used
by Murphy et al. (2001) to date the occurrence of a sec¬
ond phreatic event in a conduit's history.
OTHER PHENOMENA
Pockets resemble hollow, bisected hemispheres or
hemiellipses (Springer and Wohl, 2002). They are
ENCYCLOPEDIA OF CAVES
SHALLOW SUBTERRANEAN HABITATS
683
readily distinguishable from scallops by their lack of
oversteepened upstream faces. The flow field in pock¬
ets can be forward, reversed, upward, or downward.
All four flow fields were recorded in Buckeye Creek
Cave in West Virginia, U.S.A. In contrast to scallops,
pockets do not migrate but are developed on defects
such as joints and stylolites, the morphology of the
pocket being dependent on the nature and orientation
of the defect. The presence of pockets within fields of
scallops suggests that some pockets may have origi¬
nated as preferentially enlarged scallops.
Lateral potholes may resemble hollow bisected
cylinders, hemispheres, hemiellipses, or upright tear¬
drops in vertical channel walls (Fig. 5). They are
eroded by sediment-laden vortical flow near the air-
water interface and are potential indicators of paleo-
flow depths and velocities. They are described from a
vadose canyon passage in Buckeye Creek Cave, West
Virginia, by Springer and Wohl (2002) where scallops
within lateral potholes record recirculating flows, with
flows descending at an angle of 10—30° from the
horizontal.
The preferential development of small-scale solu-
tional etching on one set of passage walls in gull rifts
in southwest England has been used by Self (1995) to
infer paleoground water flow.
See Also the Following Article
Karren, Cave
Bibliography
Blumberg, P. N., & Curl, R. L. (1974). Experimental and theoretical
studies of dissolution roughness. Journal of Fluid Mechanics, 75,
735-742.
Charlton, R. A. (2003). Towards defining a scallop dominant dis¬
charge for vadose conduits: Some preliminary results. Cave and
Karst Science, 30(1), 3—7.
Coleman, J. C. (1949). An indicator of water flow in caves.
Proceedings of the University of Bristol Speleological Society, 6(1),
57-67.
Curl, R. L. (1974). Deducing flow velocity in cave conduits from scal¬
lops. National Speleological Society Bulletin, 36, 1—5.
Gale, S. J. (1984). The hydraulics of conduit flow in carbonate aqui¬
fers. Journal of Hydrology, 70, 309—327.
Lauritzen, S.-E. (1982). The paleocurrents and morphology of
Pikhaggrottene, Svartisen, North Norway. Norsk Geographischen
Tidsskrift, 36, 183-209.
Lauritzen, S. -E., Abbot, J., Arnessen, R., Crossley, G., Grepperud, D.,
Ive, A., et al. (1985). Morphology and hydraulics of an active
phreatic conduit. Cave Science, 12, 139—146.
Murphy, P. J., Hall, A. M., & Cordingley, J. N. (2000). Anomalous
scallop distributions in Joint Hole, Chapel-le-Dale, North
Yorkshire, UK. Cave and Karst Science, 27(1), 29—32.
Murphy, P. J., Smallshire, R., & Midgley, C. (2001). The sediments of
Illusion Pot, Kingsdale, North Yorkshire, UK: Evidence for sub¬
glacial utilisation of a karst conduit in the Yorkshire Dales. Cave
and Karst Science, 28(1), 29—34.
Self, C. A. (1995). The relationship between the gull cave Sallys
Rift and the development of the River Avon east of Bath.
Proceedings of the University of Bristol Speleological Society, 20(2),
91-108.
Sims, M. (2004). Tortoises and hares: Dissolution, erosion and isos-
tasy in landscape evolution. Earth Surface Processes and Landforms,
29, 477-494.
Springer, G. S., & Wohl, E. E. (2002). Empirical and theoretical inves¬
tigations of sculpted forms in Buckeye Creek Cave, West
Virginia. Journal of Geology, 110, 469—481.
SHALLOW SUBTERRANEAN
HABITATS
T anja Pipan and David C. Culver 1
Karst Research Institute ZRC-SAZU, ^American University
INTRODUCTION
What are subterranean habitats? Speleobiologists {e.g.,
Botosaneanu, 1986; Juberthie, 2000) generally recognize
two primary categories of subterranean habitats: large
cavities (caves) and small interstitial cavities (gravel
and sand aquifers, and the soil). These habitats share
two important characteristics—the absence of light and
the presence of species both limited to and modified
for subterranean life. While species in both habitat
types are typically without eyes and pigment, large
cavity species have elongated appendages while ani¬
mals limited to interstitial habitats are often miniatur¬
ized with shortened appendages (Coineau, 2000). This
difference between large and small cavity species
makes it relatively easy to separate blind and depig-
mented species into the appropriate habitat category.
An example of these differences is shown for ingolfiel-
lid amphipods in Figure 1 .
However, there is one other category—shallow subter¬
ranean habitats (SSHs). SSHs are aphotic habitats such as
seeps, talus slopes, solution pockets, and tubes in epi-
karst. They are shallow (less than 10 m from the surface)
which means that they are much more intimately con¬
nected with the surface environment than deeper cavities.
They are typically more variable and with more organic
carbon than deeper subterranean habitats. The
habitable spaces within these habitats are considerably
larger than their inhabitants (Culver and Pipan, 2008).
They share with large cavities a habitable space large
enough that organisms are not in contact with solid sur¬
faces in all three dimensions. This contrasts with small
cavity habitats where cavity size is a major constraint in
the evolution of morphology. On the other hand, animals
limited to caves and SSHs have similar morphological
and ecological characteristics, such as appendage elonga¬
tion, specialization of extraoptic sensory organs, increased
Encyclopedia of Caves.
© 2012 Elsevier Inc. All rights reserved.
684
s
FIGURE 1 Diagrams of stygobiotic ingolfiellid amphipods.
(A) Trogloleleupia leleupi (12—20 mm) from a cave; (B) Ingolfiella sp.
(<1 mm) from interstitial habitats, shown at same scale as (A); (C)
Trogloleleupia opisthodorus (24—28 mm) from a cave; and (D) Ingolfiella
petkovskii (1 mm) from an interstitial habitat. Note the relatively short
appendages of I. petkovskii as well as its small size. From Coineau
(2000). Used with permission of Elsevier Ltd.
life span, resistance to starvation and reduced aggressive
behavior. This set of traits, especially the morphological
part, was termed troglomorphy by Christiansen, and this
term is widely used to describe the morphological conver¬
gence of obligate cave dwellers.
SSHs have a number of unique features relative to
other subterranean habitats, including (1) the areal
extent of an individual habitat is often small, usually
<0.1 km 2 , but many replicates exist; (2) they are rich in
organic matter relative to other subterranean habitats,
in part because of their close proximity to the surface;
(3) they have intimate connections to the surface,
resulting in greater environment variation than other
subterranean habitats; and (4) they have fauna which
includes troglomorphic species, some of which are
limited to SSHs.
Among terrestrial SSHs are the spaces between
rocks in areas of moderate to steep slope typically sta¬
bilized by moss, spaces in bedrock caused by weather¬
ing, similar spaces in volcanic terrains, clinker in lava,
air-filled epikarst spaces, and sometimes even leaf lit¬
ter. Juberthie (2000) used the term milieu soutermin
superficial or mesovoid shallow substratum (MSS) for this
type of habitat, especially the spaces between rocks,
and we use MSS as a collective term for these habitats
since it is a well-established term in the literature and
is self-descriptive.
Among aquatic SSHs are epikarst, the uppermost
layer of karst with poorly integrated solution cavities;
seepage springs, also called the hypotehninorheos; and
the underflow of streams and rivers, the hyporheos
and associated groundwater. For each habitat, we dis¬
cuss a representative example, give data on environ¬
mental variability, summarize available information on
organic carbon, and enumerate the fauna.
SEEPAGE SPRINGS
The hypotelminorheic habitat, or seepage spring
(Fig. 2), is (1) a persistent wet spot, a kind of perched
aquifer fed by subsurface water in a slight depression
in an area of low to moderate slope; (2) rich in organic
matter; (3) underlain by a clay layer typically 5—50 cm
beneath the surface; (4) with a drainage area typically
<10,000 m 2 ; and (5) with a characteristic dark color
derived from decaying leaves which are usually not
skeletonized (Culver et al., 2006). The habitat can occur
in a wide variety of geologic settings anywhere outside
of arid regions where there is a layer of impermeable
sediment, but it is probably less common in karst land¬
scapes because of the extensive occurrence of an
impermeable clay layer which prevents the downward
movement of water and the development of karst land¬
scapes. Most of the available habitat for the animals
comprises spaces between decomposing leaves and
sediment, and the animals literally live in their food.
Chemical and physical conditions vary considerably
between sites, but conductivity tends to be high, indi¬
cating that the water had been underground for some
time. Although oxygen concentrations varied consider¬
ably, the fauna does not seem to be especially sensitive
to this parameter. There was about 3.0 mg dissolved
organic carbon (DOC) L -1 in seepage springs in
FIGURE 2 Photograph of the authors at a seepage spring at
Scotts Run Park, near Washington, D.C., U.S.A. Photo courtesy of
W. K. Jones. Used with permission.
ENCYCLOPEDIA OF CAVES
SHALLOW SUBTERRANEAN HABITATS
685
Nanos, Slovenia, several times higher than most cave
water. Based on a 10-month monitoring period (March
2007 to January 2008) of a hypotelminorheic habitat in
Prince William Forest Park in Virginia, U.S.A., the
habitat was temporally variable (Fig. 3). From May
to September, hypotelminorheic temperatures were
depressed compared to the nearby surface stream, and
approximated surface water temperatures for the rest
of the year. In spite of the variability, the amplitude of
variation in hypotelminorheic temperatures is less
than that of surface waters. The maximum recorded
temperature in the hypotelminorheic was 22°C com¬
pared to 28°C in a nearby (<10 m) stream (Table 1).
This is a remarkable difference, given the superficial
nature of seepage springs. The differences may become
more important, given predictions of climatic variabil¬
ity and change.
FIGURE 3 Hourly temperature from April 7, 2007, to February 4,
2008, in a seepage spring and adjoining stream in Prince William
Forest Park, Virginia, U.S.A. Because of the scale, line thickness
indicates the extent of daily fluctuations.
TABLE 1 Statistical Properties of Temperature (in °C) Time Series
for a Seepage Spring in Prince William Forest Park, Virginia, U.S.A.
Data from Culver and Pipan (2008)
Seepage Spring
Surface Stream
Mean
12.80
15.10
Standard error
0.06
0.09
Standard deviation
4.88
7.51
Coefficient of variation
38.15
49.75
Range
20.03
27.47
Minimum
1.79
1.02
Maximum
21.82
28.49
Count
7274
7274
Based on a study of 50 seepage springs in the lower
Potomac River basin that drain hypotelminorheic habitats
within a radius of 45 km, a total of 15 macroinvertebrates
have been recorded, including 12 amphipod, 2 isopod,
and 1 gastropod species (Fig. 4, Table 2). Four of the
amphipods were probably accidentals—they were
uncommon and showed no evidence of reproduction. Of
the remaining 11 macroinvertebrate species, 7 were spe¬
cies living exclusively in subterranean habitats. Of these
seven, five are exclusively found in seeps: they are hypo¬
telminorheic specialists. Five of the seven stygobionts
were troglomorphic. Hypotelminorheic sites in Croatia
and Slovenia have a similar mixture of specialized and
nonspecialized amphipods.
TV
FIGURE 4 Photograph of the stygobiotic hypotelminorheic spe¬
cialist Stygobromus tenuis, from Scotts Run Park, Fairfax County,
Virginia. Photo courtesy of William K. Jones. Used with permission.
TABLE 2 List of Species Found in Seepage Springs
(Hypotelminorheic) in the Lower Potomac River Basin According
to Ecological Category, Whether They Are Limited to Seepage
Springs, and Whether They Show Troglomorphy
Group
Ecological
Category
Habitat
Specialist
Troglomorphic
Number of
Species
Amphipoda
Stygobiont
Yes
Yes
3
Amphipoda
Stygobiont
No
Yes
2
Amphipoda
Stygophile
No
No
3
Amphipoda
Accidental
No
No
4
Isopoda
Stygobiont
Yes
Weakly
1
Isopoda
Stygophile
No
No
1
Gastropoda
Stygobiont
Yes
Weakly
1
Source: Data from Culver and Pipan (2008).
ENCYCLOPEDIA OF CAVES
686
s
EPIKARST
Epikarst is a perched aquifer, the uppermost layer
of rock in karst regions, and the major point of contact
and transmission between surface and subterranean
water. Water is transmitted vertically either through
conduits or through small fissures to the water table.
Lateral transmission occurs through poorly integrated
cavities. Epikarst stores considerable volumes of water
which explains why, during a drought, cave streams
take much longer to dry up than surface streams in the
same area. Its vertical extent is usually 10 m or less. Its
principal characteristic is its heterogeneity, with many
solution pockets whose water chemistry is also quite
variable. Water dripping from epikarst (Fig. 5) typi¬
cally contains about 1.0 mg DOC L -1 , and while this
is relatively low, it is an important carbon source in
caves both in the establishment of the biofilm in cave
streams and in cave passages without active streams.
Epikarst is a nearly universal feature of karst areas,
except in arid zones and glaciated areas. Karst areas
cover approximately 15% of the earth's surface, and
epikarst is likely to occur over most of this area.
Although epikarst water occurs throughout karst
areas, it can usually only be sampled by filtering the
dripping and seeping water of caves (Pipan, 2005).
FIGURE 5 Photograph of water dripping from epikarst in Organ
Cave, West Virginia, U.S.A. Photo courtesy of Horton H. Hobbs III. Used
with permission.
Epikarst water shows considerable temperature buff¬
ering relative to surface water, and more temperature
buffering than hypotelminorheic water. An example is
the cave Zupanova jama in Slovenia (Fig. 6) where tem¬
perature varied by approximately 4°C over a 2-year
period. Drip rates of water, however, were highly vari¬
able, ranging from <1 mL min -1 to >100 mL min *;
with the seasonal pattern dependent on rainfall, but
displaying a lagged response. Even though epikarst
water is only a few meters below the surface, its resi¬
dence time may be weeks or even months.
In all of the places where epikarst fauna has been
sampled, stygobionts have been found, but the Slovenian
assemblages are by far the most diverse (Pipan, 2005). In
Zupanova jama in Slovenia, a total of 16 copepod species
were found in five drips and their associated pools, and
remarkably 14 of these are stygobionts (Table 3). Seven
of these species are epikarst specialists, including five
undescribed species, two of which are unique to
Zupanova jama. Other caves in Slovenia have similar
overall epikarst copepod species richness, although with
more stygophiles (species that can complete their life
cycles in either subterranean or epigean habitats) than
Zupanova jama.
HYPORHEIC
The hyporheic zone comprising water-filled spaces
between the grains of unconsolidated sediments beneath
Date
FIGURE 6 Discharge (upper panel) and temperature (lower
panel) variation of an epikarst drip in Zupanova jama, Slovenia,
based on monthly measurements taken over a 2-year period. Data
from Kogovsek (1990).
ENCYCLOPEDIA OF CAVES
SHALLOW SUBTERRANEAN HABITATS
687
and lateral to streams, the best studied of all interstitial
habitats is the surface—subsurface hydrological exchange
zone beneath and alongside the channels of rivers and
streams. The hyporheic zone of rivers is an ecotone
between surface and groundwater. The connection
between the hyporheic and permanent groundwater
(phreatic water) can be very direct or without any direct
connection at all. In the case of direct connections
between the hyporheic and permanent groundwater,
stygobiotic species are often found. Even though the
hyporheic appears to be highly uniform, it actually has a
series of upwellings and downwellings. Downwellings
typically have higher oxygen levels and more organic
matter. The exact position of these upwelling and down-
welling zones along the stream course depends on the
relative pressure of the subsurface and surface waters,
and other hydrological details. When there are
TABLE 3 List of Species of Copepods Found in Epikarst Drips
and Pools in Zupanova Jama, Slovenia, According to Ecological
Category, Whether They Are Limited to Seepage Springs, and
Whether They Show Troglomorphy
Copepod
Group
Ecological
Category
Habitat
Specialist
Troglomorphic
Number
of Species
Cyclopoida
Stygobiont
Yes
Yes
1
Cyclopoida
Stygobiont
No
Yes
1
Harpacticoida
Stygobiont
Yes
Yes
6
Harpacticoida
Stygobiont
No
Yes
6
Harpacticoida
Stygophile
No
No
2
Source: Data from Pipan (2005).
unconsolidated sediments along the stream bank, the
hyporheic can extend laterally tens of meters from the
stream bank.
Formed by a meander arm, hyporheic habitats in
the Lobau wetlands are part of the floodplain of the
Danube River near Vienna, Austria, and comprise the
Danube Flood Plain National Park. This UNESCO
Biosphere Reserve, with an area of 0.8 km 2 , has been
extensively sampled for decades. Bou-Rouch pumps,
specialized pumps designed for sampling in these
habitats, and minivideo cameras in shallow wells
revealed a complex habitat with areas of differing
porosity and permeability, variable oxygen levels, and
a rich fauna. A small 900 m 2 component of this flood-
plain, called "Lobau C" is a recent terrace of the back¬
water system "Eberschiittwasser—Mittelwasser" and is
a self-contained ecosystem with clear inputs and out¬
puts because of its position between two channels and
a dam. Loosely packed gravel, alternating with a thin
layer of finer sediments, extends several meters
beneath a thin soil cover. Temperature at a depth of
2 m varied between 3°C and 21°C over a two-and-a-
half-year period, while surface water temperatures
varied between 1°C and 26°C over the same period
(Fig. 7). DOC averaged 3.71 mg CL -1 (Danielopol et al.,
2000). Animals were found throughout the depth of
gravel, but were most common 0.5 m beneath the sur¬
face, and rare below 2 m. In Lobau C, at least 27 spe¬
cies have been found, 11 of them stygobionts (Table 4).
While these species are stygobiotic, they do not in gen¬
eral show the troglomoprhic syndrome of size
increases and appendage elongation, but rather are
miniaturized with shortened appendages (see Fig. 1).
Thus, they show more morphological kinship with
FIGURE 7 Variation in temperature and oxygen from 1991 through 1993 at site D10 in the Lobau wetlands: surface water temperature
(SW-T); groundwater temperature at 3.5 m (GW-T), and oxygen concentration at 3.5 m (GW-0 2 ). Data from Danielopol et al. (2000).
ENCYCLOPEDIA OF CAVES
688
s
TABLE 4 Number of Species and Stygobionts from the “Lobau C”
Area of Danube Flood Plain National Park, Austria
Group
Number of
Species
Number of
Stygobionts
Rotatoria
>1
1
Mollusca
>2
2
Copepoda: Cyclopoida
14
3
Copepoda:
Harpacticoida
7
2
Amphipoda
1
1
Isopoda
2
2
Source: Data from Danielopol and Pospisil (2001).
deep interstitial species rather than species from other
SSH habitats.
The hyporheic is an atypical habitat compared to
other SSHs because the animals are not troglomorphic,
but rather have the morphology of small body size
and small appendages characteristic of interstitial habi¬
tats, and because its linear reach can be extensive.
Whether it is useful to continue to include these as
SSHs remains to be seen.
MILIEU SOUTERRAIN SUPERFICIEL
(MSS)
Shallow subterranean terrestrial habitats include the
spaces between rocks in areas of moderate to steep
slope, spaces in bedrock caused by weathering, similar
spaces in volcanic terrains, air-filled epikarst spaces,
and even leaf litter. Initially, the importance of MSS
habitats was thought to be as dispersal corridors
between cave regions and did in fact account for some
disjunct distributions, but soon it was recognized that
the primary habitat for some troglobionts was the
MSS, rather than caves. MSS habitats can be thought of
as relatively large cavities embedded in a matrix of
small cavities, that is, the soil (Gers, 1998). An example
is Barranco de los Cochinos, an erosional habitat in
lava in a laurel ( Laums) forest at 940 m a.s.l. in the
Teno area, northwest Tenerife, Canary Islands. This
site comprises fragmentized basaltic rocks several mil¬
lion years in age, and covered by 40—60 cm of soil.
The surface is characterized by a dense tree covering,
low exposure to the sun, and rather high humidity.
Temperatures at a depth of 50 cm ranged from 9.2°C
to 16.4°C, compared to a range of 7.9—19.4°C on the
surface (Table 5). In fact, the annual pattern of
temperature fluctuation was quite similar between sur¬
face and MSS (Fig. 8), although there was no
detectable daily temperature cycle at the MSS site.
TABLE 5 Statistical Properties of Temperature (in °C) Time
Series for an MSS Site in Barranco de los Cochinos, Tenerife
Description
Surface
MSS
Mean
12.90
12.50
Standard error
0.025
0.021
Standard deviation
2.51
2.12
Coefficient of variation
19.45
17.00
Range
10.50
7.19
Minimum
7.93
9.24
Maximum
19.43
16.43
Count
10,055
10,055
Source: Data from Pipan et al. (2011).
FIGURE 8 Temperature profiles at hourly intervals for an MSS
site (solid line) and nearby surface site (dotted line) in a laurel forest
in Teno in northwest Tenerife, Canary Islands. Data from Pipan et al.
( 2011 ).
The Barranco de los Cochinos site has been excep¬
tionally well studied and a total of 73 invertebrates
species have been found. Of these, 41 are generalist
species, found both in surface and subterranean
habitats. Of the 32 species known from subterranean
habitats, 22 are soil specialists with shortened appen¬
dages and small size, and 10 are troglomorphic spe¬
cies, with elongated appendages and larger size. The
10 troglomorphic species are not usually found in soil,
and the 22 soil specialists may not necessarily be
permanent inhabitants of the MSS. The troglobionts
include spiders, cockroaches, beetles, millipeds, and
pseudoscorpions.
GENERALITIES
These SSHs share several features. First, all of these
habitats, with the exception of the hyporheic, support
ENCYCLOPEDIA OF CAVES
SHALLOW SUBTERRANEAN HABITATS
689
highly modified troglomorphic species, which have a
morphology that in many of its characters is conver¬
gent with that of the morphology of related deep
cave-dwelling species. Second, all of the superficial
subterranean environments have species that are
not only obligate subterranean species but also SSH
specialists. Third, SSHs are not generally resource-poor
habitats. This is especially true for seepage springs and
interstitial habitats. Resources are not as abundant in epi-
karst or MSS, but at least in the case of epikarst, there is
more DOC than in epikarst-fed streams (Simon et al.,
2007). Fourth, with the exception of light, environmental
conditions in SSHs are intermediate between surface con¬
ditions and deep subterranean conditions. Fifth, what
all these habitats share in common is that they are
aphotic. In the case of interstitial and epikarst habitats,
this is self-evident, hi the case of seepage springs, light
may be present at the exit of the seep, but the habitat
itself—the hypotelminorheic—is without light. For the
MSS long-term measurements indicate that no light
penetrates even 50 cm.
ORIGIN OF THE SSH FAUNA
It might seem curious that any species are ever
found in SSHs except for an occasional transient.
The dominant paradigm of colonization of deeper
subterranean habitats such as caves is that animals
were forced into these habitats by the vicissitudes of
climate change, such as glaciation, sea level change,
and aridity. This may be relevant to SSHs as well,
given that temperature extremes are never as high in
SSHs relative to surface habitats. If it is temperature
extremes that drive species underground, SSHs
may be a relatively favorable environment in this
respect.
In the case of aquatic SSHs (epikarst and hypotel¬
minorheic) both habitats offer an aquatic environment
that is less likely to dry out than some surface habitats.
Epikarst retains water long after surface water and
some cave streams have dried up and the clay that
underlies the hypotelminorheic also retains water dur¬
ing periods of drought. Both habitats may thus be a
refuge against climate change. Aside from darkness, it
seems likely that conditions in SSHs do not present a
formidable barrier to colonization, and may even pre¬
sent some advantages. For example, predators, at least
large ones, are absent in most SSHs because the size of
the habitat does not provide access. Organic carbon is
not scarce, or at least not as scarce as in caves, and
all of these habitats have a virtual rain of organic
matter—falling and decaying leaves in the hypotelmi¬
norheic, DOC in epikarst, organic debris in MSS, and
detrital accumulation in the hyporheic.
EVOLUTIONARY AND BIOGEOGRAPHIC
CONNECTIONS WITH OTHER
SUBTERRANEAN HABITATS
Depending on the proximity of SSHs to deeper sub¬
terranean habitats, especially caves, colonization of and
subsequent adaptation to SSHs may be an evolutionary
pathway to colonization of deeper, more extreme sub¬
terranean environments. If this is the case, then a phylo¬
genetic tree of a group with both SSH and deep cave
species should have a topology with SSH species more
basal than deep cave species. Unfortunately, no detailed
phylogenies are available for groups with both SSH and
subterranean species. Part of the problem may be classi¬
fication of the habitat. At first glance, the hypotelminor¬
heic might appear to be a surface habitat, while epikarst
species might simply appear to be cave species.
In many studies of subterranean fauna, the degree of
troglomorphy has been equated with the degree of adap¬
tation to subterranean environments and to the length of
time a lineage has been isolated in subterranean environ¬
ments. We suggest an alternative view—the degree of
troglomorphy reflects site-specific differences in the sub¬
terranean habitat the species occupy. Extreme appendage
elongation, reduction in metabolic rate, and increased lon¬
gevity, to cite only a few troglomorphic characters, may
occur in organisms in extreme, isolated subterranean
environments—deep caves and deep groundwater. More
typical troglobionts and stygobionts may be the inhabi¬
tants of SSHs. They are not necessarily phylogenetically
younger, they are just in a different habitat, one that is
dark but neither constant nor extremely food-poor.
See Also the Following Articles
Epikarst
Epikarst Communities
Bibliography
Botosaneanu, L. (Ed.), (1986). Stygofauna mundi. Leiden, The
Netherlands: E. J. Brill.
Coineau, N. (2000). Adaptations to interstitial groundwater life.
In H. Wilkens, D. C. Culver & W. F. Humphreys (Eds.), Subterranean
ecosystems (pp. 189—210). Amsterdam: Elsevier.
Culver, D. C., & Pipan, T. (2008). Superficial subterranean habitats—
gateway to the subterranean realm? Cave and Karst Science, 35, 5—12.
Culver, D. C., Pipan, T., & Gottstein, S. (2006). Hypotelminorheic—a
unqiue freshwater habitat. Subterranean Biology, 4, 1—8.
Danielopol, D. L., & Pospisil, P. (2001). Hidden biodiversity in the
groundwater of the Danube Flood Plain National Park, Austria.
Biodiversity and Conservation, 10, 1711—1721.
Danielopol, D. L., Pospisil, P., Dreher, J., Mosslacher, F., Torreiter, P.,
Geiger-Kaiser, M., et al. (2000). A groundwater ecosystem in the
Danube wetlands at Wien (Austria). In H. Wilkens, D. C. Culver
& W. F. Humphreys (Eds.), Subterranean ecosystems (pp. 481—511).
Amsterdam: Elsevier.
ENCYCLOPEDIA OF CAVES
690
s
Gers, C. (1998). Diversity of energy fluxes and interactions between
arthropod communities from soil to cave. Acta Oecologica, 19,
205-213.
Juberthie, C. (2000). The diversity of the karstic and pseudokarstic
hypogean habitats in the world. In H. Wilkens, D. C. Culver &
W. F. Humphreys (Eds.), Subterranean ecosystems (pp. 17—39).
Amsterdam: Elsevier.
Kogovsek, J. (1990). The properties of the precipitations seeping
through the Taborska Jama roof. Acta Carsologica, 19, 143-156.
Pipan, T. (2005). Epikarst — A promising habitat. Postojna—Ljubljana:
ZRC Publishing at Karst Research Institute ZRC-SAZU.
Pipan, T., Lopez, Oromi., P., Polak, S., & Culver, D. C. (2011).
Temperature variation and the presence of troglobionts in terres¬
trial shallow subterranean habitats. Journal of Natural History 45,
253-273.
Simon, K. S., Pipan., T., & Culver, D. C. (2007). A conceptual model
of the flow and distribution of organic carbon in caves. Journal of
Cave and Karst Studies, 69, 279—284.
SHOW CAVES
A rrigo A. Cigna
Union Internationale de Speleologie
A SHORT HISTORY OF SHOW CAVES
Caves have always attracted the attention of humans
since prehistory, but at that time the interest was mainly
quite practical, that is, to have a shelter, a sanctuary,
or a burial place. Later, until the Middle Ages, caves
were associated with the devil or hell in general and
people feared going into them. For this reason, bandits
could use caves as hiding places without problems
from undesired visitors.
Some historical show caves were known already in
ancient times. About 2000 year ago Plinius, a Roman
writer, described the "Dog's Cave" near Naples, Italy,
being visited by people because of the peculiar release
of carbon dioxide close to the floor which killed small
animals (hence its name) while standing people were
not affected. Other caves were visited not purely for
tourism but mainly for religious purposes: such
shrines may be found everywhere.
In Postojna Cave (Slovenia), on the walls of the
so-called Passage of the Ancient Names were found
old signatures left by occasional visitors, the most
ancient ones dating back to 1213,1323, and 1393 accord¬
ing to some authors of the nineteenth century. Around
1920, such signatures were scarcely visible on account
of the seepage; presently the oldest signature which can
be read easily dates 1412 and from the sixteenth century
onward they became rather abundant. This means that
from the sixteenth century the cave was visited more
frequently by many persons attracted by the under¬
ground world. Therefore, this period may be considered
the start of cave tourism.
In a more recent time the Cave of Antiparos in Cyclades,
Greece, became a great attraction in the seventeenth
century as documented by the many prints reproducing
the cave. The Kungur Cave, 100 km SE of Perm, near
Kungur (Urals), Russia, is an ice show cave already vis¬
ited in the eighteenth century. Probably it is the largest
show cave in gypsum. On August 13, 1772, the scientist
Joseph Banks landed on Staff a Island and in November
he wrote in the Scots Magazine : "... there is a cave in this
island which the natives call the Cave of Fingal." Since
that time this cave became one of the best known caves
of the world, inspiring poets and musicians. Its fame
was so great that it became the natural cave most repre¬
sented in paintings and engravings all over the world.
The Cango Cave (Oudtshoorn, South Africa) was
discovered around 1780 and the first recorded visit
was made in 1806. A few years later, a farmer bought
the land around the cave with the exclusion of the
entrance. The Governor included into the deeds the
condition that the farmer was obliged to leave, per¬
fectly free and undisturbed, the entrance of the cave,
to be considered as public property, with a road in his
land to reach the cave. This document has a historical
importance because it is probably the first attempt in
the world to legislate for cave protection.
In the United States, the Mammoth Cave, Kentucky,
was defined "the stellar attraction of the Mammoth
Cave National Park" by Gurnee and Gurnee (1990).
Already known in prehistory, in the late eighteenth
century the cave was mined for saltpeter to make gun¬
powder. But only at the beginning of the following
century. Mammoth Cave became a tourist attraction.
If a show cave is defined as a cave where a fee is
paid in order to have access and to visit it, then the old¬
est one is the Vilenica Cave in Slovenia. The cave is
close to the village of Sezana, just a few kilometers from
the Italian border. At the beginning of the seventeenth
century, the Count of Petac began to invite the people
of Trieste and some noble friends to visit the cave. On
certain holidays, a hundred meters from the entrance,
an area for the orchestra and a dance floor were set up
and the entire dripstone passage was illuminated with
torches and candles. Probably already in 1633, the
Count Benvenut Petac charged admission to visit the
cave. Part of the money was donated to the local church
of Lokev where masses were dedicated to "greater
safety" of the people in the cave.
THE ENVIRONMENTAL PROTECTION
OF SHOW CAVES
A cave is an environment with little contact with
the outside. For this reason its equilibrium may be eas¬
ily changed when additional energy is introduced
Encyclopedia of Caves.
© 2012 Elsevier Inc. All rights reserved.
SHOW CAVES
691
(Cigna, 1993). Obviously, such changes may occur
more frequently when the whole energy budget of the
cave is small, but in case of show caves the energy
budget is often not very small, because of their size
which is generally large. A river or a subterranean lake
plays an important role in keeping the natural equilib¬
rium because they may absorb any further input of
energy more easily than rock.
In a show cave both the visitors and the electric
lighting system release energy into the environment.
A person who is walking releases nearly as much
energy as a 200-W bulb at a temperature of about
37°C. Therefore, the total energy released by hundreds
or thousands of visitors in a day is not negligible as an
absolute amount. The heat released by the electric
lighting system has the same order of magnitude.
There are different ways to keep the additional
energy input into the cave as low as possible. A
limit of the number of visitors is given by the so-called
visitors' capacity, which is defined as the maximum
number of visitors acceptable in a time unit under
defined conditions which does not imply a permanent
modification of a relevant parameter. Otherwise,
instead of reducing the number of persons, the time
they spend in the cave may be reduced. This result
may be easily achieved when people enter the cave
through one entrance and exit along another passage,
instead of returning along the same pathway where
they got in.
Using high-efficiency lamps can reduce the contri¬
bution by the electric lighting system. A further reduc¬
tion can be obtained if the lamps are switched on only
when visitors are in the vicinity.
Another perturbation of the cave environment is
the lint (hair, dry-flaking skin, dust from shoes, and
lint from clothing) left by visitors. In caves visited
by a large number of people, the accumulation of lint
becomes a real problem to be solved by an accurate
removal. In fact, such lint would cause deterioration of
formations and reduce their pristine white beauty to a
blackened mess.
Lint released into a cave might be reduced by means
of air curtains at the entrance. Such a solution would
wash people entering the cave and, at the same time,
isolate the cave environment from outside since an air
curtain acts as an invisible door and prevents airflow
through it.
The protection of the environment of a show
cave is fundamental both from the point of view of
avoiding any damage to a nonrenewable patrimony
and the conservation of the source of income for
the cave management. Therefore, such a common
interest may have an important role in the imple¬
mentation of any action aiming to safeguard the cave
environment.
Visitors also release carbon dioxide as a result of
their breathing. Until few years ago, such carbon diox¬
ide was considered a threat to the cave formations
since it could have increased the water acidity and,
consequently, the corrosion instead of the deposition
of new formations. Further accurate studies (Bourges
et al., 1998) have shown that in many instances the car¬
bon dioxide produced by natural processes (oxidation
of organic matter in the soil above a cave) may intro¬
duce through the water percolating into the cave
amounts very much larger than the carbon dioxide
released by visitors.
When the water, with a relatively high concentra¬
tion of carbon dioxide, reaches the cave environment
it releases immediately part of such carbon dioxide,
which is not in equilibrium with the carbon dioxide
in air. Therefore, the chemical reaction moves toward
the deposition of calcium carbonate and the forma¬
tions continue to grow. In general, rather small caves
with a high visitor flux and without any input of nat¬
ural carbon dioxide might have formations corroded
because the chemical reactions would be reversed
when the carbon dioxide of air dissolves into water,
particularly when water vapor condenses on the cave
walls.
Another form of environmental pollution may occur
through a joint contribution by visitors and light.
Persons release into the cave spores or seeds of plants
and they may grow in the vicinity of lamps if the light
flux is high enough. The result is the so-called lampen-
flora, that is, green plants (generally algae, fern, moss)
developing on cave walls or formations close to a
light source. Such plants cover the surfaces with a
greenish layer, which can become included into the
calcite deposition and no longer removable. In fact, the
lampenflora may be washed away by bleach or hydro¬
gen peroxide if it is not covered by any calcite. Special
care must be taken to avoid any nuisance to the cave
fauna.
The growth of lampenflora can be avoided by em¬
ploying light sources with a very low emission of light
useful for the chlorophyllian process and low light
flux at the rock surface.
THE DEVELOPMENT OF A SHOW CAVE
A correct development of a show cave must take
into account both the protection of the environment
and the safety of the visitors. As it has been already
pointed out in the previous paragraph, the physical
and chemical equilibria of the environment should be
modified outside the range of the natural variations.
At the same time, any undue source of harm to the
visitors must be avoided. This means that the
ENCYCLOPEDIA OF CAVES
692
s
pathways must be strong enough to withstand the
very high humidity and, sometimes, also floods. In the
past, wooden structures were often used, but they had
to be replaced frequently; currently, some "green" peo¬
ple would still use wood because this material is natu¬
ral. Nevertheless, the rather short life of a wooden
structure in the cave environment implies an addi¬
tional cost which is not justified by any advantage. On
the contrary, the rotten wood supplies large amounts
of food modifying the equilibrium of the cave life.
In particular at present, the criterion to use only
structures which can be easily decommissioned is sub¬
stantially wrong because once it is no longer conve¬
nient to manage a show cave, no one will spend money
to take out structures inside the cave. Only when show
cave managers will be obliged to deposit a given
amount of money to assure the future decommission¬
ing of any structure, it is possible to use structures to
be easily disassembled.
In the meanwhile, it is preferable to use materials
which are compatible with the cave environment and
will not release pollutants in the long run. A material
with these characteristics and not expensive is con¬
crete. It may be conveniently used for pathways in
general.
The handrails in stainless steel are also a convenient
solution, particularly when they are also used as pipes
to provide water in different parts of the cave to wash
out the pathways. In fact, a higher cost of stainless
steel is justified by a lack of any maintenance after
many years of operation. Sometimes plastic may be
used under the condition that it does not contain any
contaminant (e.g., heavy metals or organic compounds
which may be released).
When an artificial entrance is needed in order to
give an easy access to the cave or to establish a circuit
by avoiding the return of visitors on the same path¬
way, it is absolutely necessary to install a system of
doors to stop any additional airflow in the cave. Up
until now, doors operated mechanically or manually
are normally used, but it would be most preferable
to install air curtains. This solution (suggested many
years ago by Russell Gurnee and Jeanne Gurnee) is
less expensive, quite safe, and has the great advantage
of avoiding any sense of claustrophobia to visitors.
In addition, it also decreases the release of lint by
people as reported previously.
The surveillance of the main parameters (tempera¬
ture, humidity, carbon dioxide, radon, etc.) can be
achieved by a monitoring network, which should
always be installed in any show cave. Presently, it is
possible to install networks at a very reasonable cost,
which are reliable and require little care, as, for
instance, data loggers, which can be discharged every
month, and the data transferred into a computer for
any further evaluation. Automatic networks directly
connected to a computer are operated more easily but,
of course, their cost is higher. In any case, it must be
stressed that any kind of monitoring network always
requires some attention to avoid malfunctioning and
calibration, possibly once a year.
Such monitoring networks have also another impor¬
tant advantage because they contribute much interest¬
ing data, which have greatly enlarged the knowledge of
the behavior of cave environment. A rather widespread
feeling among speleologists, and people in general, that
a cave is lost to science when it is developed as a tourist
attraction is not at all supported by the important
scientific results obtained just within many show caves.
Sometimes the borderline between use and abuse may
be difficult to define; nevertheless, a careful develop¬
ment continuously monitored may be the most efficient
way to protect a cave.
It is evident that the economy of a region around a
show-cave-to-be can be radically modified by the cave
development. Therefore, strenuous opposition to any
tourist visitation appears to be rather unfair also toward
the local people, particularly when a suitable compro¬
mise between strict conservation and a sound develop¬
ment can be found. But in any case, as it was previously
reported, cave development cannot be accepted if it is
not supported by appropriate preliminary research.
An evaluation of the number of show cave visitors
all around the world (Cigna and Burri, 2000), based
on data obtained for about 20% of all show caves
indicated a number of more than 150 million visitors
globally per year. By assuming a budget per person as
reported in Table 1, the total amount of money spent
to visit the show caves is around US$ 2.3 billion. The
number of the local people directly involved in the
show cave business (management and local services)
can be estimated to be several hundred per cave, that
is, some hundreds of thousands of individuals in the
world.
By taking into account that there are several hundred
other people working indirectly to each person directly
connected with a show cave (Forti and Cigna, 1989), a
TABLE f Rough Estimation of the Annual Direct and Local
Budget of a Show Cave per Visitor (2008 US$)
Direct income
6.5
Other local income:
Souvenirs & snacks
2.0
Meals
6.5
Transportation
2.5
Travel agency
2.5
Total
20.0
ENCYCLOPEDIA OF CAVES
SHOW CAVES
693
gross global figure of about 100 million people receive
salaries from the show cave business. Therefore, it can
be roughly assumed that behind each tourist in a show
cave there is about one employee directly or indirectly
connected.
In addition to show caves, the existence of karst
parks, which include a cave within their boundaries,
must also be considered. As reported by Halliday
(1981), the number of visitors of three top karst
national parks in the United States (Mammoth Cave,
Carlsbad Caverns, and Wind Cave) amounted to about
2500,000 tourists each year. Therefore, karst parks give
a further increase to the number of people involved in
the whole "karst" business.
There are many other human activities which
involve a larger number of people; nevertheless, the
figure reported above is not negligible and gives an
indication of the role that show caves play in the
global economy.
GUIDELINES FOR THE DEVELOPMENT
OF SHOW CAVES
Some guidelines aiming to supply a recommendation to
be endorsed for the development of show caves were
drafted over the past few years and received strong
recommendations from the UIS Department of Protection
and Management at both the 14th International Congress
of Speleology held in Kalamos, Greece, in August 2005
and the 15th International Congress of Speleology held
in Kerrville, TX, in July 2009. These guidelines are here
reported.
1. Development of a Wild Cave into a Show Cave
The development of a show cave can he seen as a positive
financial benefit to not only itself but also the area sur¬
rounding the cave. The pursuit of these antici-pated benefits
can sometimes cause pressure to be applied to hasten the
development of the cave. Before a proposal to develop a wild
cave into a show cave becomes a physical project, it is neces¬
sary to carry out a careful and detailed study to evaluate the
benefits and risks, by taking into account all perti-nent fac¬
tors such as the access, the synergy, and possible conflict
with other tourism-related activities in the surrounding
area, the availability of funds, and many other related fac¬
tors. The conversion should only take place if the results of
the studies are positive. A wild cave that is developed into a
show cave, and is subsequently abandoned, will inevitably
become unprotected and be subject to vandalism in a very
short time. A well-managed show cave assures the protection
of the cave itself, is a source of income for the local economy,
and may also contribute to a number of scientific researches.
1-1 A careful study of the suitability of the cave for
development, taking into account all factors influenc¬
ing it, must be carried out, and must be carefully eval¬
uated, before physical development work commences.
2. Access and Pathways within the Cave
In many caves it has been found to be desirable to provide an
easier access into the cave for visitors through a tunnel, or a
new entrance, excavated into the cave. But such an artificial
entrance coidd change the air circulation in the cave, causing
a disruption of the ecosystem. To avoid this, an air lock
should be installed in any new entrance into a cave. On the
other hand, it must be mentioned that in some very excep¬
tional cases a change in the air circulation coidd revitalize
the growth of formations. A decision not to install an air
lock must be taken only after a special study.
2-1 Any new access into a cave must be fitted with an
efficient air lock system, such as a double set of
doors, to avoid creating changes in the air circulation
within the cave.
Caves are natural databases, wherein an incredible amount
of information about the characteristics of the environment,
and the climate of the cave, are stored. Therefore, any inter¬
vention in the cave must be carried out with great care to
avoid the destruction of these natural databases.
2-2 Any development work carried out inside the
cave should avoid disturbing the structure, the
deposits, and the formations of the cave as much as
possible.
When a wild cave is developed into a show cave, pathways
and other features must be installed. This invariably
requires materials to be brought into the cave. These materi¬
als should have the least possible impact on both the aes¬
thetics of the cave and its underground environment.
Concrete is generally the closest substance to the rock that
the cave is formed in, but once concrete is cast it is
extremely expensive and difficult to modify or decommis¬
sion. Stainless steel has the distinct advantage that it lasts
for a long time and requires little to no maintenance but it
is expensive and requires special techniques to assemble and
install. Some recently developed plastic materials have the
advantage of a very long life, are easy to install, and are rel¬
atively easy to modify.
2-3 Only materials that are compatible with the cave,
and have the least impact on the cave, should be
used in a cave. Cement, concrete, stainless steel, and
plastics are examples of such materials.
The environment of a cave is usually isolated from the out¬
side and therefore the introduction of energy from the outside
will change the equilibrium balance of the cave. Such changes
ENCYCLOPEDIA OF CAVES
694
s
can be caused by the release of heat from the lighting system
and the visitors and also by the decay of organic material
brought into the cave, which introduces other substances into
the food chain of the cave ecosystem. In ice caves, the envi¬
ronmental characteristics are compatible with wood, which is
frequently used for the construction of pathways, as it is not
slippery.
2- 4 Organic material, such as wood, should never be
used in a cave unless it is an ice cave where, if neces¬
sary, it can be used for pathways.
3. Lighting
The energy balance of a cave shoidd not be modified beyond
its natural variations. Electric lighting releases both light
and heat inside the cave. Therefore, high-efficiency lamps are
preferred. Discharge lamps are efficient, as most of the
energy is transformed into light, but only cold cathode
lamps can be frequently sivitched on and off without incon¬
venience. Lightemitting diode (LED) lighting is also very
promising. As far as possible, the electric network of a cave
should be divided into zones to enable only the parts that
visitors are in to be lit. Where possible a noninterruptible
power supply should be provided to avoid problems for the
visitors in the event of a failure of an external power supply.
3- 1 Electric lighting should be provided in safe, well-
balanced networks. The power supply should prefer¬
ably be noninterruptible.
It is essential to ensure the safety of the visitors in the cave,
particularly in the event of a failure of the main power sup¬
ply. Emergency lighting should always be available whether
it is a complete noninterruptible power supply or an emer¬
gency lighting system with an independent power supply.
Local code requirements may be applicable and these may
permit battery lamps or a network of LEDs or similar
devices.
3-2 Adequate emergency lighting should be available
in the event of a power outage.
Lampenflora is a fairly common consequence of the introduc¬
tion of an artificial light supply into a cave. Many kinds of
algae, and other superior plants, may develop as a result of
the introduction of artificial light. An important method to
avoid the growth of green plant life is to use lamps that do
not release a light spectrum that can be absorbed by
chlorophylll.
3-3 Lighting should have an emission spectrum with
the lowest contribution to the absorption spectrum of
chlorophyll (around 440 nm and around 650 nm).
Another way to prevent the groivth of lampenflora is
the reduction of the energy reaching any surface where the
plants may live. The safe distance between the lamp and the
cave surface depends on the intensity of the lamp. As a
rough indication, a distance of one meter should be safe.
Special care should also be taken to avoid heating the forma¬
tions and any rock paintings that may exist.
3-4 Lighting sources should be installed at a distance
from any component of the cave to prevent the
growth of lampenflora and damage to the formations
and any rock paintings.
The lighting system shoidd be installed in such a way that
lights only in the portions of the cave occupied by visitors
are switched on, leaving the lighting in the portions of the
cave that are not occupied switched off. This is important
from the aspects of reducing the heating of the cave environ¬
ment and preventing the growth of lampenflora, as well as
decreasing the amount of energy required and its financial
cost.
3- 5 Lighting should be installed in a manner to
enable only the portions of the cave that are occupied
by visitors to be illuminated.
4. Frequency of Visits and Number of Visitors
The energy balance of a cave environment can be modified
by the release of heat by visitors. A human being, moving in
a cave, releases about 150 W — approximately the same as
a good incandescent lamp. Consequently, there is also a
limit on the number of visitors that can be brought into a
cave zvithout causing an irreversible effect on the climate of
the cave.
4- 1 A cave visitor capacity, per a defined time period,
should be determined and this capacity should not be
exceeded. Visitor capacity is defined as the number of
visitors to a given cave over a given time period, which
does not permanently change the environmental para¬
meters beyond their natural fluctuation range. A con¬
tinuous tour, utilizing an entrance and another exit,
can reduce the time that visitors spend in a cave, com¬
pared to the use of a single entrance/exit.
In addition to the normal tours for visitors, many show
caves have special activities, sometimes called adventure
tours, where visitors are provided with speleological equip¬
ment for use in wild sections of the cave. If such a practice
is not properly planned, it may cause serious damage to the
cave.
4-2 When visits to wild parts of a cave are arranged,
they must be carefully planned. In addition to pro¬
viding the participants with the necessary speleologi¬
cal safety equipment, there must be experienced
guides to show the wild caves. The pathway that
visitors travel along, must be clearly defined, for
ENCYCLOPEDIA OF CAVES
SHOW CAVES
695
example, with red and white tape, and the visitors
should not be allowed to walk beyond this pathway.
Special care must be taken to avoid any damage to
the cave environment, and the parts beyond the path¬
way must be maintained in a clean condition.
5. Preservation of the Surface Ecosystem When
Developing Buildings, Parking, Removal of
Surface Vegetation, and Waste Recovery
It is important that the siting of the aboveground facilities,
such as the buildings, parking, and waste recovery, be ivell
planned. There is a natural tendency to try to place these
development features as close as possible to the cave
entrance. Sometimes these features are built over the cave
itself, or relevant parts of it. The hydrogeology above the
cave must not be modified by any intervention such as
the watertight surface of a parking area. Any change in the
rainwater seepage into a cave can have a negative influence
on the cave and the growth of its formations. Care shoidd be
exercised also when making any change to the land above
the cave, including the removal of the vegetation and distur¬
bance of the soils above the bedrock.
5-1 Any siting of buildings, parking areas, and any
other intervention directly above the cave must be
avoided in order to keep the natural seepage of rain¬
water from the surface in its original condition.
6. Monitoring
After the environmental impact evaluation of the develop¬
ment, including any other study of the cave environment, it
is necessary to monitor the relevant parameters to ensure
that there is no deviation outside acceptable limits. Show
caves should maintain a monitoring network of the cave
environment to ensure that it remains ivithin acceptable
limits.
6-1 Monitoring of the cave climate should be under¬
taken. The air temperature, carbon dioxide, humidity,
radon (if its concentration is close to or above the
level prescribed by the law), and water temperature
(if applicable) should be monitored. Airflow in and
out of the cave could also be monitored.
When selecting scientists to undertake studies in a cave, it is
very important that only scientists zvho have good experi¬
ence with cave environments be engaged for cave-related
matters. Many otherwise competent scientists may not be
fully aware of cave environments. If incorrect advice is given
to the cave management, then this could result in endanger-
ment of the cave environment. Cave science is a highly spe¬
cialized field.
6- 2 Specialized cave scientists should be consulted
when there is a situation that warrants research in a
cave.
7. Cave Managers
The managers of a show cave must never forget that the
cave itself is "the golden goose" and that it must be pre¬
served with great care. It is necessary that persons involved
in the management of a show cave receive a suitable educa¬
tion, not only in the economic management of a show cave,
but also about the environmental issues concerning the pro¬
tection of the environment at large.
7- 1 Cave managers should be competent in both the
management of the economics of the show cave and
its environmental protection.
8. Training of the Guides
The guides in a show cave have a very important role, as
they are the " connection” between the cave and the visitor.
Unfortunately, in many instances the guides have not been
trained properly and, notwithstanding that they are doing
their best, the overall result will not be very good. It is very
important that the guides receive proper instructions about
the environmental aspects of the cave as ivell as dealing with
the public. It is important that guides are skilled in tactfidly
avoiding entering into discussions that can have a detrimen¬
tal effect on the overall tour. The guides are the guardians of
the cave and they must be ready to stop any misbehavior by
the visitors that could endanger the cave environment.
8-1 Cave guides should be trained to correctly inform
the visitors about the cave and its environment.
INFORMATION ON SHOW CAVES
IN THE WORLD
There are many books published in different coun¬
tries providing guides to the local caves. On one hand
they report a rather large amount of information but,
on the other hand, they are fully reliable only for
a short time after their publication. Information on
visitor details and even the existence of the show caves
can change frequently.
Recently, a rather useful way to obtain up-to-date
information became available. "Showcaves of the
World" is a website (Duckeck, 2000) that can be found
at www.showcaves.com/. This site changes and grows
continually, so the latest version may always be seen
on the web Figures 1—7.
ENCYCLOPEDIA OF CAVES
696
FIGURE 1 Kartchner Caverns (AZ, U.S.A.) are probably the show
caves in the world, managed according to the most up-to-date crite¬
ria. An electric train from the office building to the cave is available
for visitors, avoiding both acoustic and environmental pollution.
Photograph taken by the author.
FIGURE 3 Kartchner Caverns (Arizona). A monitoring network
should always be envisaged in a show cave in order to avoid relevant
changes of the cave environment. In the case of Kartchner Caverns,
such monitoring is particularly important since the relative humidity
within the cave is close to 100%, while outside the climate is very dry.
During the construction of a pathway, many pipes are placed to be
covered by concrete. Such pipes are used for circuits connecting any
sensor within the cave with a computer in the main building for pres¬
ent and future monitoring networks. Photograph taken by the author.
FIGURE 2 Kartchner Caverns (AZ, U.S.A.). Sometime in a cave
there are joints which may imply possible rock fall. In order to assure
the safest condition in a show cave, it is necessary to detect well in
advance any minor movement. A detector, monitoring possible dis¬
placement of a limestone layer, can be seen in this photo. Photograph
taken by the author.
FIGURE 4 Cango Caves (Oudtshoorn, South Africa) was discov¬
ered around 1780 and the first recorded visit was made in 1806; now
it is the most important show cave of Africa. Here a flowstone
deposited over a clay filling, which has been washed away succes¬
sively, can be seen. Iron handrails along the pathway in the forefront:
Nowadays such handrails are generally substituted by stainless steel
or plastic handrails, to avoid iron contamination of the cave environ¬
ment. Photograph taken by the author.
ENCYCLOPEDIA OF CAVES
SHOW CAVES
697
FIGURE 5 Cango Caves (Oudtshoorn, South Africa). In one of the
innermost parts of the cave, Cango 3, which is not open to tourism,
the floor is covered by calcite crystals; a plastic ribbon fixes the
boundaries of the trail, to avoid any destruction of the crystals due to
scientists moving around. Photograph taken by the author.
FIGURE 6 Hwanseon Cave (Daei, South Korea) is a very impor¬
tant show cave of Korea. Tourists reach the vicinity of the cave by
car or bus and then they have to climb a slope of a couple of kilo¬
meters to reach the cave entrance, which is here depicted. Photograph
taken by the author.
FIGURE 7 Frasassi Caves (Ancona, Italy) developed mainly in
the shallow phreatic zone, where rising hydrogen sulfide—rich
groundwater mixes with oxygenated seepage water. Presently, these
are the most important show caves in Italy. Details may be found
in www.frasassi.com. Shown here is Niagara Falls. Photo courtesy
of Ente Consorzio Frasassi.
See Also the Following Articles
Lampenflora
Recreational Caving
Bibliography
Bourges, F., D'Hults, D., & Mangin, A. (1998). Etude de VAven
d’Orgnac. Rapport final, Laboratoire Souterrain de Moulis C.N.R.
S. Geologie Environnement Conseil, pp. 1—84.
Cigna, A. A. (1993). Environmental management of tourist caves.
Environmental Geology, 21, 173—180.
Cigna, A. A., & Burri, E. (2000). Development, management and
economy of show caves. International Journal of Speleology, 29B
(1-4), 1-27.
Duckeck, J. (2000). Showcaves of the world, <www.showcaves.com/>.
Forti, P. & Cigna, A.A. (1989) Cave tourism in Italy: an overview.
Proceedings of the International Symposium on the 170th
Anniversary of Postojnska Jama. Postojna, Nov. 10-12, 1988,
Centre for Scientific Research SAZU & Postojnska Jama Tourist
and Hotel Organization, p. 46—53.
Gurnee, R., & Gurnee, J. (1990). Gurnee guide to American Caves.
Closter, NJ: R. H. Gurnee Inc.
Halliday, W.R. (1981) Karstic national parks: International economic
and cultural significance. Proceedings of the International
Symposium on Utilization of Karst Areas, Trieste, March 29—30,
1980, Commissione Grotte Boegan, CAI Trieste, pp. 135—144.
ENCYCLOPEDIA OF CAVES
698
s
SIEBENHENGSTE CAVE SYSTEM,
SWITZERLAND
Pierre-Yves Jeannin and Philipp Hauselmann
Hohlenforschergemeinschaft Region Hohgant (HRH) and
Swiss Institute for Speleology and Karst Studies (S1SKA),
Switzerland
GEOGRAPHICAL AND GEOLOGICAL
SETTING
With a length of more than 156 km and a depth range of
1340 m, the Reseau Siebenhengste-Hohgant Cave System
is one of the most important on Earth. Close to this net¬
work, several other large caves have been found, but
they remain unconnected to the main system so far.
The total length of explored conduits within the area is
about 320 km, and this ensemble is referred to as the
Siebenhengste Cave System in this article.
The system is located in the frontal Alpine range
(Helvetic Nappe), directly facing the Swiss Plateau
(Fig. 1). From the edge of Lake Thun, where the main
spring of the system (Batterich) is located, it extends
more than 20 km northeastward to the Schrattenfluh,
crossing the deep Emme Valley (Bitterli, 1988).
The mountain range enclosing the caves is the
southeast-dipping limb of an anticline. Its northwest¬
ern limit is formed by high cliffs. Elevations of the
summits located along the cliffs range between 1950
and 2190 m a.s.l. Patches of denuded karren fields
occur above 1700 m a.s.l., where limestone is exposed.
Below this limit, sandstone mainly crops out; therefore,
forest, meadow, and swamp cover most of the area.
The climate is humid and temperate, dominated by a
western wind. The average annual temperature is
about 2°C at 1800 m a.s.l., and annual precipitation
ranges between 1500 and 2000 mm. Southeastward,
the limestones/sandstone series disappears below the
thick flysch deposits of the Pennine Nappes (Fig. 2).
Karst features are developed predominantly within
the Schrattenkalk formation (Barremian to Aptian,
Cretaceous, Urgonian facies), which is usually
150—200 m thick. Detailed geological studies identified
6 different formations within the Schrattenkalk, which
is underlain by the Drusberg Marls (lower Barremian).
These marls are 30—50 m thick. Most of the under¬
ground rivers follow the dip on the top of this impervi¬
ous layer, often along the main faults. The Hohgant
sandstones (Eocene), which are locally up to 200 m
thick, overlie the Schrattenkalk (Fig. 2). In most of the
region, faults enable the surface waters to flow through
the sandstones down into the Urgonian limestones.
FIGURE 1 Location of the Siebenhengste region with the
water catchments of the St. Beatus Spring (bottom left) and of
the Batterich/Gelber Brunnen springs (Siebenhengste-Hohgant-
Schrattenfluh). Adapted from Jeannin et al, 2000.
FIGURE 2 Geological cross-section of the Siebenhengste range;
see Figure 3 for profile location.
An important longitudinal normal fault, stretch¬
ing from Lake Thun to Schrattenfluh (the Hohgant-
Sundlauenen Fault), disrupts the continuity of the
southeastward-dipping monocline of the Siebenhengste
range (Figs. 2 and 3). The offset on the fault is
200—1000 m, depending on the location. Several parallel
normal faults are also present. The normal faults in the
Siebenhengste region mainly developed during the
Lower Cretaceous to the Eocene. Another set of faults is
extensively developed in the area: the dextral strike-slip
Encyclopedia of Caves.
© 2012 Elsevier Inc. All rights reserved.
SIEBENHENGSTE CAVE SYSTEM, SWITZERLAND
699
faults and, to a lesser degree, the related sinistral strike-
slip faults. These faults are related to alpine tectonics
and are of Miocene age. They appear to have been
partly reactivated very recently.
West of the Siebenhengste, the frontal folds of the
Helvetic Nappe are observed in the Sigriswilergrat
Range, west of Justistal (Figs. 1 and 2), allowing place¬
ment of the original crest of the anticline between 1 and
2 km in front of (northwest of) the Siebenhengste cliffs.
HYDROLOGY
Most of the region is drained by underground karst
systems emerging at two main locations: the St. Beatus
spring and the Batterich/Gelberbrunnen springs.
The St. Beatus spring with a discharge ranging
between 10 L s -1 and 2—3 m 3 s -1 (average 72 L s _1 ) has
a catchment area of about 10.5 km 2 that lies in
the southeast part of the region. The Batterich/
Gelberbrunnen spring catchment extends at least 21 km
to the northeast, reaching the Schrattenfluh massif, as
proven by a tracing experiment (Fig. 1). As the main
spring (Batterich) lies below Lake Thun, the discharge
of the system is very difficult to measure. It probably
exceeds 20 m 3 s -1 during floods. The catchment area is
around 32 km 2 and is largely covered by sandstone
where surface flow may occur locally. Recharge rates
are therefore difficult to assess.
OVERVIEW OF THE CAVE SYSTEMS AND
HISTORY OF THEIR EXPLORATION
The different caves known at present sum to a total
length of more than 320 km of passage, lying between
480 and 2000 m a.s.l. The system has a distinct laby¬
rinth character, whose complex geometry can only be
explained by the development of several superim¬
posed cave systems that correspond to different and
more or less independent times and conditions.
Seven large caves enclose most of the known pas¬
sages in the area (Figs. 3 and 4). Many passages of the
caves are surveyed and mapped in detail, as shown in
Figure 5.
St. Beatus Cave
St. Beatus Cave is a spring cave that can be followed
from the spring upstream for approximately 2 km as a
crow flies (Hauselmann et al., 2004). The explored
length of the cave is 12 km with a height of +353 m
with respect to the entrance. Most of the cave is of
phreatic origin (elliptic passage), although canyons are
found in many places (keyhole cross-section). Most of
the passages lie on the top of the Drusberg marls.
A detailed study of the morphology of this cave and
its sediments was conducted (Hauselmann 2002),
allowing the identification of several speleogenetic
phases (see later discussion). Some smaller caves are
known in the vicinity of St. Beatus Cave.
The entrance of the cave has been known since pre¬
historic times and seems to have been inhabited by the
hermit called St. Beatus in the Early Middle Ages. The
first caving explorations were made in the nineteenth
century, and already in 1904 the entrance part of the
cave had become a show cave. Maybe one-third of
today's known passages were explored before World
War II. The evolution of the caving techniques (espe¬
cially cave diving) allowed the exploration of the
other two-thirds of the cave over the past 50 years.
A complete and systematic resurvey of the cave was
conducted in the 1990s, allowing the exploration of
about 1 km of new passages. Cave exploration is
almost complete by today's standards.
Barenschacht (“Bear’s Shaft”)
Barenschacht is the second largest cave in the region
with a length of 71 km and a depth of —979 m (as of
ENCYCLOPEDIA OF CAVES
700 s
FIGURE 4 Three-dimensional view of the cave system. From Jeannin et al., 2000. Used with permission.
October 2011). The name of the cave comes from three
bear skeletons found at the bottom of the first shaft,
some hundreds of meters from the entrance. The
entrance part presents a series of shafts leading to the
bottom of the limestone (Schrattenkalk) at a depth of
150 m (Bitterli et al., 1991). From this point down to
—550 m, the passage is steep, with small shafts and a
cross-section on the order of 5 m 2 . Most of this section
has developed along a thin calcareous bed within the
Eocene sandstone series (Hohgant series). Due to large
faults, deep shafts make it possible to reach an incredi¬
ble labyrinth of large elliptic conduits at a depth of
—900 m. The bottom part of the cave floods, as it is
located at the level of Lake Thun (Batterich spring).
Fligher levels remain dry and are richly decorated by
dripstones and flowstones, gypsum crystals, and some
aragonite formations.
With the exception of 2 km in the entrance part,
Barenschacht is almost exclusively of phreatic origin.
Direct infiltration into the cave is prevented by a thick
cover of marls and flyschs. In some places, the depth
of the conduits below the land surface reaches as
much as 1000 m. Therefore, the large cross-sections
found in the labyrinth (between 12 and 25 m 2 ) clearly
indicate that Barenschacht is the downstream collector
of the Batterich /Gelberbrunnen catchment area, that
is, the downstream part of most of the caves of the
Siebenhengste region.
Barenschacht was first discovered in 1965 and
explored in 1973 and 1974 down to a sump at —565 m.
Explorations were conducted by large teams recalling
the conquest of the Himalayan summits in the 1950s.
The sump was dived and considered too tight. In 1986,
an audacious caver dug underwater and was able to
squeeze through a very narrow flooded conduit for
about 40 m. A steep and large passage ended in
the darkness of very large shafts. This was the key to the
labyrinth at —900 m and the beginning of incredible
explorations by a very restricted group of cavers, able to
cope with the dive of both a nasty siphon and difficult
vertical caving. Nearly 40 km were explored at that time
(Funcken, 1994). In 1995, a tunnel was dug on top of the
sump, making it possible also for nondivers to reach the
deep part of the cave (Funcken et al., 2001). Today's
exploration is driven by the Coordination Group of
Barenschacht, which is part of the body of HRH.
The Reseau Siebenhengste-Hohgant
The Reseau Siebenhengste-Hohgant is located below
the Siebenhengste karren field. In the international
scene of speleology, it is the best-known system of the
region, because its exploration was quite spectacular.
This is also the reason why its name has been used for
the entire cave region. This cave network was con¬
nected to FI and Faustloch in the 1980s, giving today a
system of 156 km of connected passages with a depth
of 1340 m. The Siebenhengste labyrinth itself is nearly
108 km long.
Schematically, most of the entrances to the Reseau
start with a series of shafts and short meanders down
to a depth of 150—250 m (Hof, 1984). The Reseau ("net¬
work" in French) itself is reached at that depth, where a
labyrinth of elliptic or keyhole passages can be followed
for kilometers, upstream, downstream, or horizontally.
This labyrinth is clearly visible on the map of the caves
(Fig. 3). Similar to antennas, underground streams have
been explored downstream along large and straight
canyons that follow some of the main strike-slip faults
ENCYCLOPEDIA OF CAVES
SIEBENHENGSTE CAVE SYSTEM, SWITZERLAND
701
FIGURE 5 Example of a detailed map of the cave system. Adapted from Bitterli et ah, 1991.
ENCYCLOPEDIA OF CAVES
702
s
of the region. These all reach a perpendicular major
conduit coming from FI. Due to geological complica¬
tions (crossing a normal fault), the major conduit splits
into three-dimensional anastomoses in the so-called
Zone Profonde.
Exploration of the Siebenhengste massif started in
1966, but it was in 1972 and 1973 that the existence of
such a large cave system was revealed. Exploration was
then quick, with a length of 20 km being reached in
1976, 32 km (to a depth of —800 m) in 1978, and 45 km
in 1982. It was at that time that the FI (Innerbergli area)
was connected, giving a length of 60 km to the Reseau.
This quick exploration was not systematic at all, leaving
many unexplored passages. A more systematic explora¬
tion of these side passages was begun in 1982 and is still
ongoing today. As a consequence, the overall aspect of
the Reseau Siebenhengste-Flohgant did not change
much during this period, but the density of the conduits
has more than doubled. The exploration of this incredi¬
ble cave system is still far from finished, and new kilo¬
meters of passage are found every year. Exploration
and survey are achieved in a very systematic way in the
Reseau as well as at the surface. Fhmdreds of small
caves have been explored in the Siebenhengste karren
field, some of them reaching a length of 1 or 2 km; how¬
ever, only a restricted subset of them could be con¬
nected to the Reseau.
The Faustloch
The Faustloch is a 14-km-long cave (reaching —930 m),
which was connected to the Reseau Siebenhengste-
Flohgant in 1987. The entrance part, down to —200 m,
developed in the sandstone overlying the Urgonian lime¬
stone and is followed by two huge shafts (80 m and 60
m) crossing the limestone down to its bottom. From here
onward, an underground stream can be followed along a
canyon. At —450 m, the stream crosses the Drusberg
marls and goes deeper into a series of small and wet
shafts and meanders, developed in a siliceous limestone
(Flauterivian). At the top of the descent across the
Drusberg marls, a major elliptic conduit is found (cross-
section of 10—20 m 2 ). This conduit is the continuation of
the FI-Siebenhengste major collector (fossil phreatic con¬
duit). It can be followed downstream over several kilo¬
meters toward the Barenschacht and divides into several
branches. Two of them end on deep siphons (one of
them has been dived down to — 38 m for a length of 300
m), which he slightly higher than those of the
Barenschacht. The other branches are clogged by thick
sediments. Despite strong digging efforts, the continua¬
tion has not yet been opened.
The Faustloch was discovered in 1970, the entrance
being as large as a fist ("faust" in German). It took
several years to descend the shaft series down to
—350 m—first, because of large quantities of water cas¬
cading down the shafts, and, second, because it was
very difficult to set anchors in the sandstone. The first
third of the cave was explored between 1971 and
1978. In 1987, a large flood event occurred in the
Siebenhengste region, and the fossil passage coming
from the Reseau Siebenhengste-Flohgant was reacti¬
vated by a stream of several cubic meters per second.
This opened the passage upstream of Faustloch, allow¬
ing the Reseau to be connected. It also opened the
downstream end of the cave (Funcken and Moens,
2000). After a long digging effort, almost 10 km of pas¬
sage could be found toward the Barenschacht. In 1997,
a new flood clogged this passage again, and the most
remote part of the cave was not accessible until 2009,
when the access was dug open again.
Due to strong cold water cascades, the entrance
shaft series of Faustloch are a dangerous section of the
cave, where three cavers have lost their lives (one in
1976 and two in 1998).
FI (Innerbergli Area)
FI cave has a typical dendritic pattern. Its length is
nearly 32 km and its depth is 650 m. It was connected
in 1982 with the Reseau Siebenhengste-Hohgant. The
entrance shaft series is comprised of small pits and
meanders, with some tight passages. At a depth of
150 m, the Drusberg marls are found, and the passage
gets larger, thanks to the presence of some tributaries
converging at that point and to the collapse of the con¬
duit wall where it contacts the Drusberg marl. A large
active canyon can be followed for about 3 km before
disappearing into a large collapse room. From this
room onward, a phreatic fossil conduit continues for
another 3 km. It is joined by four main tributaries but
ends (clogged) very close to the Reseau. The connec¬
tion could be managed only through tight and wet
passages. The cave entrance was found in the summer
of 1981 and the main conduit grew over 15 km in
the 18 months prior to the connection to the Reseau
at Christmas 1982. A systematic exploration of the
tributaries has been undertaken over the past 30 years,
but the complete exploration of FI will require many
more years.
K2
This cave is parallel and quite similar to FI. Its
length is 14 km and its depth is —750 m. The entrance
shaft series, found in 1980, is followed by a severe
squeeze and a long and narrow meander down to a
room at —300 m. The passage out of this room is very
ENCYCLOPEDIA OF CAVES
SIEBENHENGSTE CAVE SYSTEM, SWITZERLAND
703
tight, and many cavers have had to turn back at this
point. After a short maze of small phreatic conduits, a
large canyon passage is reached and can be followed
for several kilometers. This section is very dark and
slippery. At a depth of —600 m, the main passage
splits into a complex three-dimensional labyrinth of
elliptic conduits (fossil phreatic passages) reaching a
siphon at its lowermost point. In 1991, a dive of the
siphon led to the deeper point of the cave (—750 m),
where any continuation does not seem possible. At
that time, this point was probably one of the remotest
parts of Switzerland. In 1992, a lower entrance was
found, giving access to the lower part of the cave and
reactivating exploration in this region. Many tributar¬
ies were explored, but no major continuation could be
found. The intermediate section of K2 is still being
explored with the hope of being able to connect it to
FI or to the Haglatsch cave (7.5 km), which develops
about 80 m straight above K2.
A2 (Hohlaub Area)
This 11-km-long cave (—690 m) can be considered as
the southwestern continuation of the Siebenhengste
labyrinth, but no connection has been achieved. The
entrance was found in 1973 and explored down to a
depth of 30 m. There, a very tight meander was found.
At least four groups have tried to squeeze through the
meander, where a strong draft indicated a continuation.
Only in 1986 was the 35-m-long passage traversed. A
mining action was organized in order to be able to
explore the galleries found beyond. A series of shafts
and meanders was explored down to a depth of 220 m.
A labyrinth of horizontal elliptic passages (fossil phre¬
atic conduits), corresponding to the Siebenhengste laby¬
rinth, crosscuts the shaft zone. Downstream, an active
canyon stretches toward the east to a collapse room,
where the water disappears. A fossil phreatic conduit
goes on toward Faustloch, but turns abruptly to the
right some hundreds of meters before reaching it. This
region, called Keller (Cellar), develops along a fault and
gives access to the lowermost point of the cave
(—690 m). It is still nearly 100 m above the nearby
Faustloch. Most of the cave was explored between 1986
and 1994; however, there are still a lot of passages to
discover (Gerber, 1994).
GENESIS AND AGE OF THE
CAVE SYSTEMS
Theories regarding the genesis of the Siebenhengste
cave system began to be developed immediately on
exploration of the cave, as a result of the exceptional
density of conduits. Various contradicting hypotheses
were discussed. In the middle of the 1980s, Flof (1984)
laid the groundwork for the current ideas. Jeannin
et al. (2000) and Hauselmann et al. (2002) refined this
model. The ideas are based on the observation of the
conduit morphology that allows recognition of paleo-
water tables.
Two major observations help to identify the position
of paleowater tables:
1. The change of a vadose conduit morphology
(canyon or meander) into a phreatic morphology
(elliptic tube) gives the exact height of the
paleowater table. However, it has been shown that
this paleowater table is not horizontal (Hauselmann
et al, 2003).
2. In looping passages, the position of the paleowater
table can be inferred to be higher than the top of the
loops, if they do not present any indication of free-
surface flow morphology (entrenchment).
By applying this concept, several phases (or levels)
of conduits have been recognized. The inferred paleo¬
water tables are considered to be linked to the position
of paleosprings ( i.e ., of paleovalley bottoms; regional
base level). Phases are labeled with elevations inferred
for the paleosprings. Due to the Alpine uplift, real ele¬
vations may have been different when the systems
were active.
The oldest speleogenetic phases are found in the
uppermost caves. Five phases could be identified at
1950, 1800, 1720, 1585, and 1505 m a.s.l., respectively.
The conduit morphology (scallops) indicates that water
flowed to the northwest and that the springs of those
hydrological systems were located in the Eriz Valley
(Fig. 6). Based on additional data, we hypothesize
that the Aare Valley (today's main valley) did not exist
when this part of the cave system was active and that
the inferred paleosprings were located along the
(mostly disappeared) Eriz paleovalley.
During the next series of speleogenetic phases,
water flowed to the southeast to springs located in the
Aare Valley (Fig. 7), meaning that a significant geo-
morphic event happened between phases 1505 and
1440, capable of turning the flow direction 180° within
the karst system. Evidence for springs at 1440, 1150,
1050, 890, 805, 760, 700, 660, and 558 m a.s.l. has been
found so far.
Cave sediments were investigated, and the follow¬
ing ages of the cave systems could be obtained by
applying the U-Th and the cosmogenic burial age
methods (Hauselmann et al, 2007, 2008):
• Phase 1950: must be older than 4350 ky
• Phase 1800: oldest found deposition of quartz sand
at 4350 (±600) ky
ENCYCLOPEDIA OF CAVES
704
• Phase 1800: first deposition of glacial pebbles at
1870 (±210) ky
• Phase 1585: active system (not necessarily phreatic)
between 1570 and 760 (±420) ky
• Phase 1505: active phreatic system between 740 and
600 (±170) ky
• Phase 1050: last deposition of quartz sand at 440
(±160) ky
• Phase 760: active phreatic system between more
than 350 ky and 207 (±27) ky
• Phase 700: active phreatic system between 207 (±27)
and 146 (±11) ky
• Phase 660: active phreatic system between 146 (±11)
and 28 (±11) ky
• Phase 558: active phreatic system since 28 (±11) ky
until today
These datings are novel for the Alps and made it
possible not only to date valley deepening phases with
relatively high precision, but they also contribute to
document the drain switch from Eriz to the Aare
Valley (due to its glacial incision), as well as incision
velocity of the valleys (0.12 mm a -1 in the Eriz,
1.2 mm a -1 in the Aare Valley) and erosion rates at the
Siebenhengste plateau surface (roughly 0.12 mm a -1
throughout, therefore proof of enhancing relief).
Further studies in the area proved that the oldest
glacial pebbles found in the Siebenhengste (1.87 Ma BP)
were transported from southern Valais to the north,
before the present-day valley pattern of the Alps
was established. All these data together make the
Siebenhengste a prime site for the geomorphological
evolution of this area of the Alps.
CAVE MINERALS
Many sections of the caves in the Siebenhengste
region are nicely decorated. White-to-yellowish flow-
stones are present in many passages, producing a
strong contrast with the dark limestone walls. Some of
the flowstones can be followed over hundreds of
meters, up to almost 1 km.
ENCYCLOPEDIA OF CAVES
SIEBENHENGSTE CAVE SYSTEM, SWITZERLAND
705
FIGURE 7 Younger phases of the cave system genesis.
ENCYCLOPEDIA OF CAVES
706
s
PRESENT AND PAST FAUNA
DISCOVERED IN THE CAVES
FIGURE 8 Sketch of a gypsum mantle covering and stalactite.
Illustration courtesy of A. Hof, 1984.
Stalactites and stalagmites are common features, but
they are never very large and are usually quite dis¬
persed. Some particular speleothems are calcite stalac¬
tites mantled in a gypsum crust that remain at a few
centimeters from the calcite core (Fig. 8). Larger gypsum
crystals are sometimes present at the tip of the stalactite.
Soda straws are found in many places, some of
them reaching more than 3 m in length. Helictites,
draperies, cave pearls, rimstone dams, cave shields,
moonmilk, and some spars are found in all parts of the
cave system, but mainly in the lower levels. Although
frequent, speleothems are generally not set in a very
dense pattern in the Siebenhengste caves, but really
decorate the conduit walls, roof, and floor.
Beside these very classical calcite formations, arago¬
nite is observed as frostwork or is interbedded in sta¬
lagmites or stalactites. Rarely, soft fibrous aragonite
aggregates looking similar to asbestos have also been
found. Gypsum is very frequently found in the caves
of the Siebenhengste region. Gypsum stars and needles
(up to 15 cm long) are the most common formations.
Gypsum crusts, snow, and mantles (Fig. 8) are also
quite common. Gypsum hair up to 30 cm is quite rare,
but can be seen in some cave parts. In Faustloch, a floor
is covered by a 20-cm-thick carpet of mirabilite crystals,
probably coated by some thenardite white crust.
In fact, very few data have been collected on miner¬
als and speleothems of the Siebenhengste region. More
systematic investigations would probably reveal many
other types of minerals and formations.
Many investigations of cave biology have been
undertaken in the Siebenhengste region, but have
never been brought to the attention of a wider public.
Investigations in St. Beatus Cave (Hauselmann et al.,
2004) revealed the existence of 26 mosses, 4 ferns, and
3 algae as well as 4 mushroom species; also, 11 bat
species have been observed, sometimes quite far from
the cave entrances. Many invertebrates have been col¬
lected and determined. The current numbers follow:
Collembola (11), Niphargus (3), Dytiscides (1), Asellus (2),
Diptera (4), Turbellaria (2), Ostracoda (1), Gasteropoda
(8), Copepoda (1), Diplopoda (1), Opiliones (1),
Arachnia (2), Acaria (2), Trichoptera (4), Coleoptera (4),
Thysanura (1), Lepidoptera (3), Oligochaeta (1), and
rainworms (2), giving a total of 54 species (24 of which
are described from St. Beatus Cave). Bythiospeum alpinum
(Mollusca: Gasteropoda: Prosobranchia: Hydrobiidea),
found in Barenschacht, is a new animal species for
science. The same is valid for Onychiurus dunarius, a
collembol first described from St. Beatus Cave.
Attention has also been paid to remains of past
fauna. Bone fragments of a cave bear ( Ursus spelaeus)
have been found in one cave, but these are the only
remains of Pleistocene fauna.
A large spectrum of Holocene animals (46 species
total) have been found, usually at the bottom of
entrance shafts, which acted as natural traps. The
Brown Bear ( Ursus Arctos), who was extinct in
Switzerland by around 1850, was present in many
caves. Bear nests have been found in two caves.
Moreover, a moose ( Alces alces) found in the
Innerbergli area dated back to 424 to 633 A.D. Many
species of small mammals and small bovines common
to the region were found in the caves as well and also
provide interesting information. For instance, the pres¬
ence of domestic animals and of species typical for a
forest environment is surprising in caves that open in
a denuded limestone pavement. In fact, these animals
are relicts of warmer periods, when forest still covered
much of the Siebenhengste region almost to the top of
the mountains.
The same type of observation has been made from
bat bones found in the caves. The presence of species
living today at least 500 m below their finding location
indicates relicts of a warmer climate (Morel, 1989).
Today, bats are still present but scarce. The very large
number of skeletons found in the caves seems to indi¬
cate that the bat population was much higher in the
past. Most of these bones probably date from a period
extending between 4000 B.C. and 1500 A.D. as con¬
firmed by 14C datings of bat bones from a nearby karst
ENCYCLOPEDIA OF CAVES
SINKING STREAMS AND LOSING STREAMS
707
area. Although still poorly investigated, bones found
in caves provide a rich record of the past fauna and
climate of Central Europe.
resultats. Stalactite, Journal de la Socie'te' Suisse de Spele'ologie, 39(2),
59-72 (in French).
Acknowledgments
SINKING STREAMS AND
LOSING STREAMS
The English was reviewed by Monique Hobbs. We also thank all the
cavers who explored this huge system and provided very valuable
work of exploration and cave mapping. Marco Filipponi and Duccio
Malinverni kindly revised Figures 6 and 7.
Bibliography
Bitterli, T. (1988). Das Karstsystem Sieben Hengste-Hohgant-
Schrattenfluh: Versuch einer Synthese. Stalactite, Journal de la
Socie'te' Suisse de Spele'ologie, 38(1—2), 10—22.
Bitterli, T., Funcken, L., & Jeannin, P.-Y. (1991). Le Barenschacht: Un
vieux reve de 950 metres de profondeur. Stalactite, Journal de la
Socie'te' Suisse de Spele'ologie, 41(2), 71—92 (in French).
Funcken, L. (1994). Barenschacht: Plus de 36 km post-siphon. Une
exploration hors du commun. Stalactite, Journal de la Socie'te' Suisse
de Spele'ologie, 44(2), 55—82 (in French).
Funcken, L., & Moens, M. (2000). Synthese des explorations au
Faustloch depuis 1987. Stalactite, Journal de la Socie'te' Suisse de
Spele'ologie, 50(1), 3—22 (in French).
Funcken, L., Moens, M., & Gillet, R. (2001). Barenschacht:
L'interminable exploration. Stalactite, Journal de la Socie'te' Suisse de
Spele'ologie, 51(1), 9—22 (in French).
Gerber, M. (1994). Le A2-Loubenegg (region des Sieben Hengste).
Stalactite, Journal de la Socie'te' Suisse de Spele'ologie, 44(1), 3—8
(in French).
Hauselmann, P. (2002). Cave genesis and its relationship to surface pro¬
cesses: Investigations in the Siebenhengste region (BE, Switzerland).
Switzerland: University of Fribourg.
Hauselmann, Ph., Jeannin, P.-Y., Monbaron, M. & Lauritzen, S. E.
(2002). Reconstruction of alpine Cenozoic paleorelief through the
analysis of caves at Siebenhengste (BE, Switzerland). Geodinamica
Acta, 15, 261-276.
Hauselmann, P., Bitterli, T., & Hochli, B. (2004). Die St. Beatus-Hohlen:
Entstehung, Geschichte, Erforsclmng. Allschwil, Switzerland: Speleo
Projects.
Hauselmann, P., Granger, D. E., Lauritzen, S.-E., & Jeannin, P.-Y.
(2007). Abrupt glacial valley incision at 0.8 Ma dated from cave
deposits in Switzerland. Geology, 35(2), 143—146.
Hauselmann, P., Jeannin, P.-Y., & Monbaron, M. (2003). Role of epi-
phreatic flow and soutirages in speleogenesis: The Barenschacht
example (BE, Switzerland). Zeitschrift fur Geomorphologie N.F.
47(2), 171-190.
Hauselmann, P., Lauritzen, S.-E., Jeannin, P.-Y., & Monbaron, M.
(2008). Glacier advances during the last 400 ka as evidenced in
St. Beatus Cave (BE, Switzerland). Quaternary International, 189,
173-189.
Hof, A. (1984). Sieben Hengste-Hohgant Hohle: Le Reseau. Le Trou,
Bidletin du Groupe Spele'o Lausanne, 34, 18—39 (in French).
Jeannin, P.-Y., Bitterli, T., & Hauselmann, P. (2000). Genesis of a
large cave system: The case study of the north of Lake Thun sys¬
tem (Canton Bern, Switzerland). In A. B. Klimchouk, D. C. Ford,
A. N. Palmer & W. Dreybrodt (Eds.), Speleogenesis, evolution
of karst aquifers (pp. 338—347). Huntsville, AL: National
Speleological Society.
Morel, P. (1989). Ossements de chauves-souris et climatologie:
Collecte systematique de squelettes de chiropteres dans des
systemes karstiques des Prealpes et Alpes Suisses—Premiers
Joseph A. Ray
Crawford Hydrology Laboratory, Western
Kentucky University
The purpose of this chapter is to discuss the character¬
istics and significance of influent streams, creeks, and
rivers that drain underground and to evaluate their
integral function in karst hydrology. Losing streams
are sometimes observed recharging sand and gravel
aquifers and volcanic material such as basalt. This
chapter, however, will restrict discussion to the more
common karst phenomena developed within soluble
rocks such as limestone. Many of the terms associated
with influent streams are often interchanged depending
on the practice of the user.
INFLUENT RIVERS AND STREAMS
The unique nature of losing and sinking influent
streams is their development and evolution of conduit-
flow routes and caves through soluble rocks. Large
volumes of concentrated recharge from losing and sink¬
ing rivers are central to the evolution of most of the
world's largest and most significant caves and springs in
karst (Fig. 1). Caves formed by influent streams may con¬
tain a major stream or active floodwater route or these
may be long abandoned high above active base level.
Abandoned flood routes often retain fluvial sediment
and features left by the former stream.
Where major rivers cross soluble rocks they gener¬
ally carve a low-gradient, continuous surface valley
through the region and usually function as the master
drainage or base level for local tributaries and springs;
however, certain rivers and most smaller streams
become influent and lose water where they traverse
soluble rocks. For example, influent water from the
upper Danube River resurfaces at the Aach Spring,
Germany's largest resurgence. Because of early con¬
flicts over water rights, this subsurface river was the
site of the first quantitative water trace to prove the
spring's source. The pioneering 2-day study, con¬
ducted by Knop in 1877, used hourly salt content mea¬
surements of Aach Spring to calculate a concentration
curve totaling nearly 10,000 kg of salt, about half of the
quantity deliberately injected into the Danube River
sinks (Kass, 1998).
The Takaka River in New Zealand begins losing base
flow into sediment-covered marble, 20 km from its
Encyclopedia of Caves.
© 2012 Elsevier Inc. All rights reserved.
708
s
FIGURE 1 Steep-gradient stream sinking into a large-capacity cave opening in southeastern Kentucky. The flood capacity of this cave may
exceed 20 m 3 s _1 . Photo courtesy of James C. Currens. Used with permission.
resurgence at Waikoropupu Spring. This large spring
maintains a low-flow discharge of >5 m 3 s -1 (Ford and
Williams, 1989). The Reka River in Slovenia sinks into
Skocjanska Cave and resurges 30 km south near Trieste,
Italy, forming a group of 16 springs with a mean dis¬
charge of >26 m 3 s -1 (Jennings, 1985). The Trebinjcica
River in the Dinaric karst is one of the largest rivers in
the world to sink into limestones and contributes to the
Ombla River Spring near Dubrovnik. One of the largest
springs in the world, the Ombla has a discharge ranging
from 4 to 140 m 3 s -1 .
River rises, supplied primarily by nearby losing riv¬
ers, are a common type of large spring in Florida. The
typical losing river reach is less than 5 km in length.
The Santa Fe River, a major tributary of the Suwannee
River in north-central Florida, may be the largest los¬
ing stream in the United States. The losing reach has
developed where the Santa Fe River incises a regional
escarpment or linear bluff. The influent conduit capac¬
ity ranges from about 5 to 16 m 3 s -1 at O'Leno State
Park, 4.8 km northeast of its resurgence at Santa Fe
Rise. Subsurface flow is locally visible in a karst win¬
dow about 3 km from the sink area, and when river
stage exceeds 14 m, the dry river channel is reoccupied
by flood waters (Martin and Dean, 1999).
Big Spring, Missouri's largest spring with a dis¬
charge ranging from 6 to 36 m 3 s _1 , receives water
from several losing tributaries of the Eleven Point
River. This regional-scale recharge was demonstrated
by groundwater tracer tests conducted by Aley over
distances of 60 km. Elk River is possibly the largest los¬
ing stream in West Virginia, with 8 km of dry river
channel, under-drained by caves and conduits feeding
Elk River springs (Jones, 1997).
The Edwards Aquifer in south-central Texas is one
of the largest, most important karst aquifer systems in
the United States. It contains a 5- to 25-km-wide, 350-
km-long recharge zone that is traversed by numerous
losing rivers and creeks. The Blanco, Medina, Frio/Dry
Frio, and Nueces/West Nueces rivers are four of the
larger losing-stream systems that contribute to the
karst aquifer with a combined average recharge of
ENCYCLOPEDIA OF CAVES
SINKING STREAMS AND LOSING STREAMS
709
FIGURE 2 Stream patterns and urban centers
in an 8250-km 2 region surrounding Mammoth Cave
National Park in Kentucky. A consistent stream den¬
sity occurs over most of the region except for the
streamless karst plain, oriented southwest to northeast.
Numerous sinking streams drain beneath this sinkhole
plain and contribute to large springs on the Green and
Barren rivers. These sinking stream watersheds are
identified with a gray tone.
20 m 3 s -1 . Large springs, including Comal, San
Marcos, and Leona springs, and high-volume artesian
wells, yielding groundwater flow under hydrostatic
pressure, are supplemented by this extensive network
of losing streams.
With more than half of the state covered by soluble
rocks, Kentucky contains many losing and sinking
streams. Figure 2 shows stream patterns in the region
surrounding Mammoth Cave National Park. The light-
shaded diagonal band showing an absence of streams
corresponds with the sinkhole plain. This gently rolling
landscape was an early settlement-transportation corri¬
dor because of the streamless, prairie terrain. The main
portion of the sinkhole plain, between the Green and
Barren rivers, is flanked to the southeast by dozens of
sinking streams. These influent streams comprise the
headwaters of the three largest karst drainage basins in
Kentucky, including groundwater in the National Park.
Blue Spring Creek, with a watershed of 37 km 2 , is the
largest of these sinking streams. Its basin comprises
nearly 10% of the 390-km 2 watershed of Gorin Mill
Spring, Kentucky's largest spring. Figure 3 shows the
sequence of multiple stream swallets that are activated
over a 5-km reach of Blue Spring Creek as discharge
increases from low flow to moderate and high flow.
Losing streams and subterranean meander cutoff
caves and springs are common in southwestern
Kentucky. These features frequently occur along the
West Fork of Red River and Sinking Fork of Little River
(Mylroie and Mylroie, 1991). The most significant losing
stream is probably the Sinking Creek/Boiling Spring
system in north-central Kentucky. Although 19 km of
Sinking Creek is shown on topographic maps upstream
of Boiling Springs, this 20- to 30-m-wide stream is actu¬
ally a dry channel for most of the year.
FUNCTION OF INFLUENT STREAMS
In order to comprehend the hydrology of influent
streams, one must also understand the structure of normal
streams. An idealized stream basin may be viewed as
having three basic components: the headwaters, or alloge¬
neic source; the main stem, or zone of transfer; and the
mouth, or discharge point (Schumm, 1977). Allogeneic
waters are defined as stream flow derived from nonlocal
sources. In karst drainage, influent allogeneic streams may
be derived from adjacent insoluble rock terranes or solu¬
ble rocks that include perching units such as shale, chert,
or sandstone. This allogeneic flow comprises the head¬
waters of most losing and sinking streams. Karst water¬
sheds that contain a significant amount of stream or
fluvial recharge are often termed fluviokarst (White, 1988).
Likewise, the concept of base level is important in
understanding influent streams. Base level refers to sea
level or the lowest level of master stream erosion.
Influent streams develop when they cross soluble rocks
along their transfer route to base-level rivers or seas. As
potential groundwater discharge locations (down-valley
or along master rivers) deepen over time, tributary
headwaters tend to keep pace by developing advanta¬
geous subsurface flow routes. Enlarging conduits typi¬
cally follow preferential pathways along bedrock
fractures and bedding planes and function as efficient
flow routes to local base level. Because karst conduit
development can occur some depth below base level,
losing reaches may occur along local base-level streams.
As incipient conduit-flow routes are established, con¬
vergent recharge from losing and sinking streams tends
to rapidly accelerate conduit enlargement. This preferen¬
tial and aggressive conduit erosion controls the location
of main flow routes and caves within the developing
ENCYCLOPEDIA OF CAVES
710
FIGURE 3 The sinking-stream complex of Blue
Spring Creek in central Kentucky. Flow variations are
compared during low-, moderate-, and high-flow con¬
ditions on three images of the same topographic map
segment. Active swallets or stream sinks are identified
with black triangles tangent to the flowing stream. The
influent capacity of successive swallets is exceeded as
stream discharge increases. Stream flooding and inter¬
mittent lakes are identified by a gray tone in the high-
flow map. More than 40 additional sink points are
located up to 2 km beyond the map border.
LOW FLOW MODERA TE FLOW
HIGH FLOW
1000 ” ' 0 METERS 1000 2000
karst drainage basin. Similar in function to local head¬
water stream networks, diffuse percolation from the
shallow weathered zone (epikarst) continually feeds the
under-draining trunk-flow routes created by losing and
sinking streams. This integrating process often creates
an efficient subsurface network, generally dendritic or
trellised in pattern, that functions as a unified karst
drainage basin or watershed.
Jakucs (1977) pointed out that much classic theory
on the development of karst ignored the erosional and
hydrologic influence of through-flowing and losing
streams derived from adjacent nonkarst terrane. He
described several varieties of erosion within karst ter¬
rane by allogeneic or "alien" streams and illustrated
the evolution of a chain of swallow-holes formed as
these streams incise soluble rocks.
In many cases, the largest karst caverns are created
in the zones where allogeneic rivers pass through the
subsurface. Because gravel, cobbles, boulders, and other
debris can be swept underground by sinking streams,
features resembling normal surface landforms such as
floodplains, alluvial bars, and stream meanders can be
formed within flood-route caves. This transported
sedimentary load contributes to mechanical abrasion
(corrasion) along with the solutional enlargement
(corrosion) of passages by aggressive flood waters.
Exceptional flood-route caves and mazes are often
formed where influent tributaries converge, where
floodwater routes diverge around obstacles, and within
effluent distributaries (Palmer, 1975). Accordingly, sink¬
hole collapse and karst-window development tend to
focus along the trunk route and near the boundary
zones of subsurface rivers.
Losing and sinking streams can concentrate large
amounts of floodwater runoff and contaminants into a
karst aquifer at specific points. Obviously, this type of
rapid recharge makes karst groundwater extremely
vulnerable to pollution. Education of the agricultural
industry, urban developers, and the public about the
dynamics of karst drainage is vital if we are to protect
and restore these important groundwater resources.
EVOLUTIONARY SEQUENCE OF LOSING
AND SINKING STREAMS
The subsurface flow route of influent streams may
range from minor cutoffs to regional trunk conduits.
ENCYCLOPEDIA OF CAVES
SINKING STREAMS AND LOSING STREAMS
711
(A) Losing Stream, Perennial Flow
(B) Losing Stream, Dry Reaches
FIGURE 4 Three basic types of influent streams draining to a base-
level or master river. These generalized illustrations utilize the terms
Q r for allogeneic discharge or headwater source and Q c for conduit-
flow capacity in the subsurface transfer zone. Arrows show flow direc¬
tion from under-draining conduits to springs in the discharge zone.
Case (A) shows a losing stream with perennial flow, where base-flow
allogeneic discharge is greater than conduit-flow capacity. Case (B)
shows a losing stream where maximum allogeneic discharge is greater
than conduit-flow capacity, but base-flow allogeneic discharge is less
than conduit-flow capacity; therefore, a dry reach occurs during base
flow. Case (C) shows a terminal sinking stream where conduit-flow
capacity is greater than maximum allogeneic discharge. Adapted from
White (1999).
As an evolving surface drainage network initially
incises soluble rocks, subterranean meander cutoffs
and short-distance detours of flow into the subsurface
may begin to develop. Where the exposure of soluble
rocks is limited, the depth and extent of conduit flow
are accordingly minimized. Such watercourses usually
regain lost conduit flow at downstream springs.
This intravalley losing and gaining karst flow is
probably the most common type of stream behavior in
soluble rocks. In addition to flow in highly soluble
rocks, losing and gaining hydrology can occur in thin
or interbedded limes tone/shale units that lack the
potential for deep conduit flow. Likewise, less soluble
rocks such as dolomite locally develop this type of
interrupted stream flow. Often, normal stream morphol¬
ogy remains largely unaffected by the stream loss, and
many topographic maps do not reveal the subsurface
diversions.
The idealized diagrams in Figure 4 illustrate a rea¬
sonable evolutionary sequence in the development of
losing and sinking streams. Three types of influent
streams tributary to a base-level river are shown and
are compared by means of a generalized water bal¬
ance. Evapotranspiration and artificial withdrawals are
ignored in these simplified relationships.
Case (A) illustrates karst flow in a losing stream
where influent conduit capacity is exceeded during all
conditions. The base flow from the allogeneic water¬
shed, Q a (base), is greater than the conduit-flow capac¬
ity (Q c ); consequently, surface flow is only diminished
within the losing-stream reach. Due to subtle flow varia¬
tions, some losing reaches can be difficult to identify.
This less common (and less recognized) situation may
occur during incipient karst development, because of
persistent conduit clogging, or along broad rivers with
large base flows. Streams with abundant loads of coarse
sediment often lose a portion of surface flow entirely
through its alluvium. This type of intergranular flow
may be locally indistinguishable from karst diversions.
Case (B) shows a losing stream that develops a dry
reach during base flow. Maximum allogeneic discharge
during flood, Q a (max), is greater than the conduit-flow
capacity (Q c ), causing a continuous overflow channel
to be fluvially eroded and maintained. However, the
conduit-flow capacity (Q c ) is greater than the allogeneic
base flow, Q a (base), which causes a dry reach below
the losing point. This is probably the most common
type of influent stream. Conduit flow may return to the
surface at a down-valley spring within the same water¬
shed (intrabasin flow) or it may divert beneath a
surface divide into a neighboring watershed (interbasin
flow). Because of their frequency and significance
in karst hydrology, groundwater systems or spring
basins dominated by losing streams may be classified
as type I or overflow allogeneic karst basins.
Case (C) illustrates a sinking stream that is pirated or
diverted entirely underground. Conduit-flow capacity
Q c is greater than the maximum allogeneic discharge,
Q a (max). Because of the exceptional conduit-flow
capacity, a surface escape route or overflow channel is
not maintained. An abandoned karst valley or sinkhole-
dominated plain often separates the stream sink and its
resurgence. Groundwater systems dominated by sink¬
ing streams in blind valleys may be classified as type II
or underflow allogeneic karst basins.
These three cases are simplified. Actual influent
streams are usually more complex, as they may exhibit
ENCYCLOPEDIA OF CAVES
712
s
overlap or transitional situations based on weather
trends or unique conditions. For example, case (A)
may resemble case (B) during severe drought, whereas
case (B) may resemble case (A) in wet years. Also,
during extreme flood events some sinking streams,
case (C), may yield intermittent lakes and overland
flow through degraded gaps or hanging valleys that
are progressively being abandoned.
Some karst areas contain relatively thick surficial
deposits of soil or sediment. Streams that lose water
through appreciable bedrock cover tend to maintain
surface overflow routes. Influent soil pipes or channels
formed in sediment frequently collapse, partially clog¬
ging flow routes. Cobbles and gravel concentrated
from the cover material tend to collect within these
unstable conduits and may help to maintain a limited
capacity of leakage into the under-drain system.
Depending on season or flow condition, losing and
sinking stream flow may advance to, or retreat from, a
series of sink points along a considerable stream reach.
The 32-km-long dry channel of Lost River in southern
Indiana is an exceptional example. Such streams have
a tendency of adjustment between the influent capacity
of under-draining conduits and loss of base flow
within a stream reach. Because winter base flow
exceeds summer low flow, the influent front tends to
migrate between the two conditions. Figure 3 shows a
channel reach of 3 km over which flow migrates
between low and moderate conditions, which gener¬
ally corresponds to the two base-flow conditions. Also,
excessive runoff from developed or agricultural areas
increases the magnitude of peak runoff and may alter
the expected base flow discharge. Knowledge of these
potential flow dynamics is important during hydro¬
geologic mapping and karst inventories for dye tracing
studies.
In summary, many rivers drain through the karst
regions of the world and are altered by interaction with
underlying conduit flow developed within soluble rocks.
Countless allogeneic tributary streams lose water or sink
entirely and contribute to the primary groundwater cir¬
culation routes through karst watersheds and drainage
basins. These main conduit-flow routes function as trans¬
fer paths from headwater influent streams to discharge
points at springs. Many of the largest and most impor¬
tant caves in the world were developed by influent
recharge. The active trunk-flow routes serve as master
groundwater collectors of diffuse recharge from infiltra¬
tion and runoff into sinkholes. Due to limited filtration of
influent stream recharge and rapid groundwater veloci¬
ties, most karst aquifers are extremely vulnerable to pol¬
lution from surface activities. Knowledge of the recharge
sources and variability inherent to karst drainage is vital
for protecting and restoring threatened cave resources,
water supplies, and many perennial spring-fed streams.
See Also the Following Articles
Hydrogeology of Karst Aquifers
Water Tracing in Karst Aquifers
Bibliography
Ford, D. C., & Williams, P. W. (1989). Karst geomorphology and hydrol¬
ogy. London: Unwin Hyman.
Jakucs, L. (1977). Morphogenetics of karst regions. New York: John
Wiley & Sons.
Jennings, J. N. (1985). Karst geomorphology. Oxford: Basil Blackwell.
Jones, W. K. (1997). Karst hydrology atlas of West Virginia. Charles
Town, WV: Karst Waters Institute.
Kass, W. (1998). Tracing technique in geohydrology. Rotterdam: A.A.
Balkema.
Martin, J. B., & Dean, R. W. (1999). Temperature as a natural tracer
of short residence times for groundwater in karst aquifers.
In A. N. Palmer, M. V. Palmer & I. D. Sasowsky (Eds.), Karst
modeling (pp. 236—242). Charles Town, WV: Karst Waters
Institute (Special Publication 5).
Mylroie, J. E., & Mylroie, J. R. (1991). Meander cutoff caves and self
piracy: The consequences of meander incision into soluble rocks.
National Speleological Society Bulletin, 52, 33—44.
Palmer, A. N. (1975). The origin of maze caves. National Speleological
Society Bulletin, 37, 56—76.
Schumm, S. A. (1977). The fluvial system. New York: John Wiley &
Sons.
White, W. B. (1988). Geomorphology and hydrology of karst terrains.
New York: Oxford University Press.
White, W. B. (1999). Conceptual models for karstic aquifers.
In A. N. Palmer, M. V. Palmer & I. D. Sasowsky (Eds.), Karst
modeling (pp. 11 — 16). Charles Town, WV: Karst Waters Institute
(Special Publication 5).
SISTEMA HUAUTLA, MEXICO
C. William Steele and James H. Smith, Jr. '
Boy Scouts of America, ^Environmental Protection Agency
Sistema Huautla, the second deepest cave in the Western
Hemisphere, is located in the southern Mexican state of
Oaxaca. It is among the 12 deepest caves in the world,
and one of the longest of the world's deep caves.
The exploration of the Huautla caves is still ongoing,
nearly 47 years since cavers first discovered them.
The stories connected to the exploration are well docu¬
mented in magazines, books, caving publications, pho¬
tography, maps, and film (Steele, 2009; Stone et al., 2002).
Sistema Huautla was the first cave outside of Europe to
be explored deeper than 1000 m.
CAVE DESCRIPTION
The integrated Sistema Huautla is comprised of
20 entrances and 62,099 m of passage to a depth of
1475 m. This most complex vertical system has more
independent deep routes than any other cave in the
Encyclopedia of Caves.
© 2012 Elsevier Inc. All rights reserved.
SISTEMA HUAUTLA, MEXICO
713
world: three routes over 1000 m deep, one over 900 m,
two over 800 m, one over 700 m, and two over 600 m.
The connection of Nita Nanta with the rest of the sys¬
tem in 1987 marked the first and, still one of the very
few times two caves each over 1000 m deep have been
joined together. The world's deepest down and up tra¬
verse, a spectacular crossover trip, can be made: 1225 m
down via Nita Nanta's highest entrance, through spec¬
tacular stream galleries of San Agustin, and 1107 m
back up Li Nita, all without retracing a single footstep.
This down and up traverse would be 11 km in traverse
length, although the two entrances are less than a kilo¬
meter apart on the surface. Such a trip is likely never to
be done due to the logistics of rigging well over 100
pitches and diving two widely separated sumps at the
bottom.
The most remote recess of Sistema Huautla is
beyond a series of water-filled tunnels at —842 m in
San Agustin. Sump dives of 450 and 170 m in length
reach air-filled tunnels that measure 20 m wide and
20 m tall that ultimately lead to Perseverance Hall, the
second largest chamber in Sistema Huautla. From
Perseverance Hall, measuring at 120 m wide and
120 m long and 30 m tall, the passage dips steeply to
galleries that lead to the deepest level and final sump
of Sistema Huautla, 1475 m below the highest
entrance. The final sump and farthest exploration is
located 2.5 km from the 842-m sump or 6.6 km from
the San Agustin entrance.
Hidden in the depths of the Sierra Mazateca, 640 m
below the entrance of San Agustin, is the jewel of
Sistema Huautla, Anthodite Hall. Anthodite Hall, at
300 m long and 200 m wide and 70 m high, is one of
the largest chambers in the region. The ceiling and
walls are covered with milky white to clear spe-
leothems, the majority of which are large anthodites.
These radiating pin-cushions are half a meter in length.
The caves of Huautla were first discovered in 1965
by cavers from Austin, TX. The caves were initially
explored through the rest of the 1960s by Canadian
cavers and cavers from assorted American states such
as Texas, Georgia, Indiana, and Tennessee. In 1967 the
distinction of having the deepest cave in the Western
Hemisphere came to the Huautla area caves with the
FIGURE 1 Sistema Huautla is located in the north¬
east corner of the state of Oaxaca, Mexico. Map by
James H. Smith.
Sistema Huautla
Municipios de Huautla de Jimenez, San Miguel
Huautepec, Mazatlan Villa de Flores
Oaxaca, Mexico
Plan View
FIGURE 2 Sistema Huautla is a complex network
of cave streams with 20 known entrances. The
pattern of stream flow is determined by the structural
geology of the karst groundwater basin. Map by Stone
(2003).
kilometers
ENCYCLOPEDIA OF CAVES
714
s
exploration of Sotano del Rio Iglesia to a depth of 531 m.
This was surpassed the following year when Sotano de
San Agustin was explored to a depth of 612 m.
Following a hiatus from 1970 to 1976, expeditions to
Huautla began again in 1976 and carried on steadily
into the twenty-first century. These were organized
FIGURE 3 Sistema Huautla is a cave composed of
many separate streams that descend vertically to a
base level and then converge to form one large stream
before discharging out a spring entrance. Map by Stone
(2003).
kilometers
FIGURE 4 Exploration beyond the 840 Sump to Sump 9 in Sotano de San Agustin is the greatest single feat of U.S. deep cave exploration
ever accomplished. Map by Barbara Anne Am Ende.
ENCYCLOPEDIA OF CAVES
SISTEMA HUAUTLA, MEXICO
715
Huautla Cave System
San Agustin Entrance
to the San Agustin Sump
Side View
-300 meters
below
the San
Agustin
entrance
meters -700
below
the Nita
Nanta
entrance
-800
- -500
Legend
Cave passage
Water in cave passage
- -600
I
(side view)
Lakes and sumps in
cave passage (top view)
Waterfall
- -700
qP
Breakdown (boulders)
Flowstone
- -800
>
Floor slope (points uphill)
Vertical floor drop-off
L -900
(hash marks on down side)
Ceiling drop-down
(hash marks on lower side)
Stream
77 m
Space
Drop
Alternate route
—►to Camp Three
(not used in '94)
620 m depot -•
Grand '
Cascade
-840 (San Agustin)
-1300
-1325 m (Nita Nanta)
drafted by Barbara Anne am Ende
FIGURE 5 Anthodite Hall, the jewel box of Sistema Huautla, contains the most spectacular anthodite speleothems found in the cave
system. Anthodite Hall is the largest chamber in Sistema Huautla. Map by Barbara Anne Am Ende.
FIGURE 6 Sistema Huautla is a vertical drainage
system composed of separate streams that descend
through the mountain down steeply descending
passages and a multitude of vertical shafts. More than
62 km of cave passages are represented with 1475 m of
vertical relief. Map by Bill Stone.
0
2
3
kilometers
Copyright 2003 Bill Stone
ENCYCLOPEDIA OF CAVES
716
s
and led by Richard Schrieber, Bill Stone, Jim Smith,
Mark Minton, and Bill Steele. In 2007, Sotano del Rio
Inglesia was connected to Sistema Huautla. These
expeditions are well documented in the newsletters of
the Association for Mexican Cave Studies.
GEOLOGY
Sistema Huautla is located in the Sierra Mazateca of
the Sierra Madre Oriental del Sur. The Sierra
Mazateca, a northwest—southeast trending mountain
range which attains altitudes of up to 2200 m, contains
the easternmost mountains to the Gulf of Mexico. The
Sierra Mazateca is flanked to the east by the foothills
and coastal plain of the Veracruz Basin and to the west
FIGURE 7 Anthodites are speleothems that are characterized by
radiating clusters of needles or quills and generally are composed of
aragonite, one of the crystalline forms of calcium carbonate (CaCCE).
Anthodites may grow from the ceiling, walls, and floor. They may
form single clusters, huge clusters, or chandeliers. This anthodite
was found in Anthodite Hall, Sistema Huautla. Photograph taken by
James H. Smith.
by the intermountain basin of the Canada de
Oaxaquena (a.k.a. Tehuacan Valley). The Canada de
Oaxaquena is drained to the east by the Rio Tomellin
and Rio Salado, which converge to form the Rio Santo
Domingo. The Rio Santo Domingo cuts across the
Sierra Madre Oriental del Sur and converges with the
Rio Papaloapan in the Veracruz Basin.
The Sierra Mazateca range receives orographic precipi¬
tation that decreases in amount from the east to the west.
Annual rainfall averages range from 6 m along the east¬
ern front of the Sierra Mazateca at Tenango to 400 mm at
Teotitlan del Camino in the arid Canada de Oaxaquena.
The Sistema Huautla area, which ranges to an altitude of
2100 m, receives approximately 2.5 m of annual rainfall.
The abundant rainfall is one of the factors essential to the
formation of deep vertical drainage systems.
FIGURE 8 Laura Smith is descending a waterfall on a rope
during the exploration of Cueva Agua Carlota, Huautla. In order to
follow the stream, cave explorers must use rope techniques perfected
specifically for cave exploration. Photograph taken by James H. Smith.
ENCYCLOPEDIA OF CAVES
SISTEMA HUAUTLA, MEXICO
717
HYDROGEOLOGY
The Sistema Huautla karst groundwater basin is an
elongate basin, 4 km wide and 14.5 km long, oriented
north—south, and is defined by its structural geology.
The basin is bordered on the west by allochthonous
sandstones and shale of Jurassic Age and to the north
and east by cretaceous limestones. The drainage of the
basin is characterized as subterranean with more than
100 km of active and paleo conduits that have been
explored and surveyed. The drainage outlet or spring
for the Sistema Huautla Cave is to the south in the Rio
Santo Domingo Canyon at an elevation of 300 m above
sea level. Subterranean groundwater flow is determined
by regional strike and plunge of dipping limestone
strata. The vertical extent of the subterranean drainage
was verified by dye tracing. From the highest mapped
FIGURE 9 At —740 m deep in lower La Grieta, Sistema Huautla,
explorers search for a potential connection to Sotano de Agua
Carrizo, an 848-m-deep cave, the bottom of which is located only a
few meters from La Grieta's stream passage. This photograph was
taken near the Triple Connection between Sotano San Agustin, La
Grieta, and Nita Nanta (the highest entrance). Photograph taken by
James H. Smith.
entrance of Sistema Huautla, Nita Nanta, to the Sistema
Huautla Resurgence, the vertical extent of 1760 m is
second only to Sistema Cheve, which dye traced to
2525 m (Smith, 2020).
BIOLOGY
The cave life in the Huautla System is highly diverse
and rich in cave-adapted animals. Of the 48 species of
invertebrates known from the caves of Huautla, 10 are
highly adapted troglobitic forms. Many groups of tropi¬
cal arachnid have been found inhabiting the caves. The
most striking animal is the scorpion Alacran tartarus
(Francke, 1982). Found deep in the system, it is usually
found near or in the water. One of the rarest troglobitic
tarantula spiders, Schizopelma grieta is also found deep in
the caves. Another, as yet unnamed, spider of the genus
Pholocophora has also been found in the caves. An ambly-
pigid with very reduced eyes, Paraphrynus grubbsi, is
known to be in four of the caves. An undescribed schizo-
mid of the genus Schizomus has also been found. Two
cave-adapted millipedes, Cleidogona baroqua and Mexi-
cambala fishi, have been found. Among the insect groups
the collembola Psendosinella bonita and the beetle Platynus
urqui have been described, while a large nicoletiid thysa-
nuran, common on the mudbanks of the streams,
remains undescribed. Closer examination of the caves
will no doubt reveal more species of cave-adapted life. A
new species of the spider genera, Ctenus, Maymena,
Coryssocnemis, Metagonia, and Modismus await descri¬
ption, as do opilionids of the genera Hoplobunus and
Karos, millipedes of the genus Sphaeriodesmus, and cam-
podeid diplurans.
ARCHAEOLOGY
Few archaeological remains have been discovered in
the passages integrated into the Sistema Huautla; how¬
ever, a site in an area cave, named Blade Cave, suggests
use by ancient Mazatecs. Evidence from the cave
includes 60 ceramic vessels that date from the Early
Urban stage (1—300 A.D.) through the Early City-State
stage (750—1250 A.D.). Many of the vessels appear to
have held perishable offerings. Frequently, wealth items
such as beads and pendants of jade, stone, shell, and
coral were included in these offerings or added to the
vessels later. Evidence of blood sacrifice is present in
the form of remains of animals, in particular dog man¬
dibles. Four bifacially chipped blades, one 18 cm in
length, were likely used as knives for such sacrifices.
Auto-sacrificial bloodletting was possibly occurring as
smaller prismatic obsidian blades were found. Human
skeletal material is abundant in the Blade Cave
ENCYCLOPEDIA OF CAVES
718
s
assemblage. It is not certain what these individuals rep¬
resent. The investigations have amassed an abundance
of detailed contextual information on ancient Mazatec
cave ritual. Ethnographic research leads to the belief
that the ceremonies conducted in Blade Cave were
primarily petitions for rain (Smith, 1987).
Bibliography
Francke, O. F. (1982). Studies of the scorpion subfamilies
Superstitioninae and Typhlochactinae, with descriptions of a new
genus (Scorpiones, Chactoidea). Association for Mexican Cave Studies
Bulletin, 8, 51—61.
Smith, J. H. (2002). Hydrogeology of the sistema huautla karst groundwater
basin, bulletin 9. Austin, TX: Association for Mexican Cave Studies.
Steele, C. W. (2009). Huautla: Thirty years in one of the world's deepest
caves. Dayton, OH: Cave Books.
Steele, J. F. (1987). Blade Cave: An archaeological preservation study in
the Mazatec Region, Oaxaca, Mexico. San Antonio: University of
Texas.
Stone, W. C., Am Ende, B., & Paulsen, M. (2002). Beyond the deep: The
deadly descent into the world's most treacherous cave. New York:
Warner Books.
SOIL PIPING AND SINKHOLE
FAILURES
Barry Beck
PE. LaMoreaux & Associates, Inc.
If you visit Mammoth Cave National Park in Kentucky,
you may hike down into huge forested sinkholes like
Cedar Sink. These low areas are completely enclosed
by steep rock walls, and they are many hectares in size
and more than 100 meters deep, possibly with a stream
flowing out of a cave on one side and into a cave on
the other. If you live in Central Florida you may have
seen a house or a road that collapsed into a sinkhole—
one that is smaller than those mentioned in Kentucky,
but significantly more devastating. If you are a farmer
in some areas of Slovenia, the only land with sufficient
soil to grow crops is at the bottom of the numerous
bowl-shaped basins that dot the landscape, thousands
of feet across and tens of feet deep; these are locally
termed dolinas, the original term for sinkholes. If you
visit West Texas, you may go to see the Devil's
Sinkhole, a gaping collapse in the ground surface more
than 30 meters in diameter that drops vertically into an
even larger cave room extending down to 100 meters
below the ground surface; whatever cave lies below is
completely plugged with the debris from the upward-
stoping roof collapse. All of these different features
have been termed sinkholes, and they are part of the
same overall process, even though each one is a little
bit different.
Sinkholes are generally thought of as "natural
enclosed depressions found in karst landscapes." In
European terminology they are generally called dolines,
an Anglicized form of the Serbo-Croatian word dolina.
Although they vary over a wide range of sizes and
shapes, they are generally subcircular in plan, and
they vary in cross-section from saucer shaped through
bowl shaped to cylinder shaped. Depending on the
type of sinkhole under discussion, they may vary
from 0.3 m in diameter and 1 m deep for small cover
collapses, to a kilometer or more in diameter and hun¬
dreds of meters deep for the polygenetic sinkholes of
mature karst landscapes. In the following description
of the origin and types of sinkholes, it is assumed that
the reader is generally familiar with karst processes
and terminology. Moreover, this discussion generally
refers to traditional karst developed on limestones or
dolostones (dolomites). Some useful sources are given
in the bibliography.
Although five different types of sinkholes are gener¬
ally identified by geologists (Beck, 1988), these are the
result of only two different processes: the transport of
surficial material downward along solutionally enlarged
channels, or collapse of the rock roof over large bedrock
cavities (Fig. 1). The natural solution process dissolves
the limestone most rapidly at its surface. If the lime¬
stone is bare, or almost bare (that is, exposed at the
ground surface), water flows over the top of the lime¬
stone toward the points of easiest infiltration, usually
at joint intersections, and then downward. As water
converges toward these points of infiltration, vertical
solution pathways develop, and the converging flow
dissolves and lowers the limestone surface, forming
a bowl-shaped depression. This is the classic solution
sinkhole (Fig. 1A).
Because the solution process is imperceptibly slow,
solution sinkholes are not generally hazards to the
engineering of human-made structures. Their role is to
act as drainage inlets into a conduit drainage network.
They are environmental hazards in that they are a
direct input to the deeper karstic groundwater.
Contaminants introduced into an open sinkhole (of
any type) will flow rapidly into the karst drainage
network with little or no degradation or filtering of the
contaminants.
If the limestone is mantled (covered) by an accumu¬
lation of residual soil (insoluble residue from the disso¬
lution of the limestone), the terrain is termed a subsoil
karst. If the limestone is mantled by sediments of an
outside origin (for example, marine sands or glacial
sediments) the terrain is termed a mantled karst. In terms
of the processes that will form sinkholes, the two set¬
tings are generally the same. The overlying unconsoli¬
dated sediments may simply be termed cover. Water
infiltrates through the cover to the limestone surface.
Encyclopedia of Caves.
© 2012 Elsevier Inc. All rights reserved.
SOIL PIPING AND SINKHOLE FAILURES
719
-PdlUk
A. Solution Sinkhole B. Cave-Collapse Sinkhole
C. Formation of a Cover-Collapse
Unconsolidated Sediment Sinkhole (L to R).
1. Solution sinkhole (A) is par¬
tially infilled and plugged, or
completely covered as shown. If
it is completely covered, it is
called a buried sinkhole.
2. Water infiltrating through the soil and
down the central shaft initiates erosion of
soil forming a void in the soil above the
shaft. The roof of the soil void gradually
breaks down and the void grows upward.
3. Upward erosion eventually reaches the
ground surface, which collapses leaving a
steep-walled pit in the soil. The limestone
shaft may be visible or may be covered with
collapsed sediment. Surface erosion will
modify the shaft to a funnel shape over time.
D. Formation of a Cover-Subsidence
Sinkhole (L to R).
1. Solution sinkhole (A) is partially in-filled
or completely covered as shown.
2. Slow movement of sediment down
the central shaft of the sinkhole gradu¬
ally undermines the land surface.
3. Overlying unconsolidated sediments
slowly settle and deform in response to the
undermining. The rate of settlement may
be inches (cm’s) per year.
E. Generalized Conceptual Cross-Section of Mature
Topography in a Mantled Karst
Note: For the sake of simplicity the limestone surface is shown to be
smooth and regular. However, in detail it is well-known to be highly irregu- prepared by pol
lar, on a scale of feet to a few tens of feet (a few tenths to several meters).
-P.E. LaMoreaux & Associates-
FIGURE 1 Sinkhole formation. Prepared by P.E. LaMoreaux & Associates, Inc. Used with permission.
then downslope to the vertical drains and downward.
In the process, the cover sediment is eroded down
the vertical conduits. This process, internal erosion of
the cover sediment, forms what have been termed subsi¬
dence sinkholes. Jennings (1971) describes them as fol¬
lows: "Where superficial deposits or thick residual soils
overlie karst rocks, dolines (sinkholes) can develop
through spasmodic subsidence and more continuous
piping of these materials into widened joints and solu¬
tion pipes in the bedrock beneath. They vary very
much in size and shape. A quick movement of subsi¬
dence may temporarily produce a cylindrical hole
which rapidly weathers into a gentler, conical or bowl¬
shaped depression" (p. 126). The process is diagram-
matically shown in Figures 1C and IE. These have also
been called alluvial dolines. However, that term is some¬
what misleading and has not been generally accepted.
If the covering sediment is somewhat cohesive, an
arched void may form in the sediment immediately
above the limestone drain, as the sediment is eroded
ENCYCLOPEDIA OF CAVES
720
s
FIGURE 2 A cover-collapse sinkhole that developed rapidly
(overnight) near Orlando, Florida. It is approximately 10 meters in
diameter and 5 meters deep. Note that only surficial sand (cover) is
exposed in the hole. The limestone is approximately 30 meters below
the surface in this area. Photo by the author.
down the drainage conduit. As upward erosion of the
cover sediment continues, the void may grow larger
over time while the eroded sediment is removed down
the conduit or shaft. Or, if collapse of the soil is rela¬
tively rapid, the void may simply migrate vertically
upward, as the crumbling sediment from the arched
roof accumulates on the floor (Fig. 1C, parts 1—3).
More cohesive strata within the cover, such as a clay
layer, may impede the upward erosion and cause the
cavity to widen, rather than grow upward for a period
of time. Eventually the upward growth of the void
may leave only a thin roof of soil that is not strong
enough to support its own weight, resulting in ground
collapse. At that moment a hole suddenly appears in
the ground surface; this is termed a cover-collapse sink¬
hole (Fig. 2). If the limestone is not too far below
ground, and if the previously collapsed sediment was
removed down the drain as the collapse grew upward,
then the limestone and drain may be visible in the
hole. More often, all that is visible in the hole is col¬
lapsed sediment. A cover-collapse sinkhole is a sub-
type of a subsidence sinkhole.
The process of localized internal erosion of the soil is
often called soil piping by geologists, in this case caused
by subsurface karstic erosion. Soil piping was described
by White (1988): "Concentrated runoff ... can form
channelized paths through the soil to a solution open¬
ing in the underlying bedrock. The increased hydraulic
gradient and increased flow velocities speed the
removal of soil from the subsurface. Because the pave¬
ment or the building itself, or even the root-bound soil
zone, acts as a supporting structure, the loss of soil may
not be noticed until a substantial cavity has formed. At
some point structural support is lost and the building
collapses into the hole" (p. 357). This downward
FIGURE 3 The remnants of a home engulfed as a large cover-
collapse sinkhole developed over several hours. Keystone Heights,
Florida (east of Gainesville). The occupants were able to evacuate
safely, although their Christmas tree and presents were lost. The sur¬
ficial sediment is clayey sand, which spans the upward eroding
subsurface void allowing it to grow to a significant size before it
collapsed. The limestone at this location is approximately 30 meters
below the land surface. Photo by the author.
erosion of unconsolidated sediment has been termed
ravelling by engineers; it is the same process.
When the earth suddenly collapses beneath a
building, a roadway, or a railroad, this is an obvious
disruption to human-made infrastructure (Fig. 3). Such
unexpected damages due to cover-collapse sinkholes
are common in karst areas, although the actual rate of
incidence is not high. Where detailed records have
been compiled and averaged, a corrected incidence of
1.7 sinkholes per square mile per year is the highest
reported. Rates of less than one sinkhole per square
mile per year are commonly reported in high incidence
areas. Inasmuch as most sinkholes are less than
60 meters in diameter, the actual risk of a sinkhole
damaging an individual 180-square-meter home in
ENCYCLOPEDIA OF CAVES
SOIL PIPING AND SINKHOLE FAILURES
721
Pinellas County, Florida, has been calculated as on the
order of 1 chance in 5000.
The incidence of cover-collapse sinkholes is often
increased by the ponding or damming of surface-water
flow due to human construction and drainage. Storm
water retention basins or waste treatment lagoons have
frequently emptied rapidly into sinkholes formed in
FIGURE 4 Damage to a home in East Tennessee caused by a
cover-subsidence sinkhole. This damage developed slowly over sev¬
eral years. However, the area is underlain by limestone and has
other characteristics of karst topography. Drilling showed a depres¬
sion in the limestone surface in the area of the damage, and a slight
depression in the ground surface could also be seen. Photo by Arthur
J. Pettit, P.E. LaMoreaux & Associates, Inc. Used with permission.
the bottom. This is obviously a source of groundwater
contamination. Large reservoirs also cause sinkholes
to collapse in their bed, often causing unacceptable
leakage and costing hundreds of thousands of dollars
to repair. In some cases, episodes of intense ground-
water withdrawal due to well pumping also trigger
the collapse of cover-collapse sinkholes. This is particu¬
larly true in Florida where high-intensity irrigation for
freeze protection of crops often causes tens of sink¬
holes to form overnight, some extremely damaging.
The surface expression of a sinkhole may also
develop imperceptibly slowly. In karst terrain a gener¬
ally circular area of the ground may slowly subside at
the rate of inches to feet per year due to subsurface
karstic erosion, not local compaction. These features
have been termed cover-subsidence sinkholes. The down¬
ward erosion may take place slowly, grain by grain, if
the sediment is loose and granular, or it may take place
by plastic flow into cavernous voids. It is also possible
that the surface sediment is gradually "sagging" down¬
ward as the area below it is undermined by karstic
erosion removing the structural support. The time-
frame involved is years. Cover-subsidence sinkholes
have been documented due to their damage to human-
made rigid structures (Figs. 4 and 5). Whereas an inch
of settlement spread out over a year is unnoticeable,
after 10 years a 25-cm-deep depression has formed.
Ground Surface
Top of Clay
otice the depression in the top of
ie clay beneath the white area and
ie subsidence in the shallow sand
yers at the same location,
arking sinkhole activity.
FIGURE 5 Ground-penetrating radar graph showing a cross-section of shallow sediment layers that have settled due to a cover-subsidence
sinkhole in Tampa, Florida. The surficial sediment is loose sand and a clay layer occurs approximately 2 to 3 meters below ground surface.
Note that the clay stratum has subsided as it has been undermined by deeper karstic erosion. Note the small, but obvious, area of settlement
in the surface sands directly over the clay subsidence. The limestone is approximately 6 meters below ground surface and cannot be seen on
this graph. This cover-subsidence sinkhole developed slowly over several years, partially beneath a home, damaging it to such an extent that
it was uninhabitable. Radar graph by P.E. LaMoreaux & Associates. Used with permission.
ENCYCLOPEDIA OF CAVES
722
s
If this is beneath a foundation, significant damage may
occur. We do not yet have a complete understanding
of the mechanisms by which cover subsidence occurs.
Both cover-collapse and cover-subsidence sinkholes are
subtypes of subsidence sinkholes.
Sinkholes are generally thought of as bowl-shaped,
funnel-shaped, or cylindrical depressions in the
ground surface. However, during and after the devel¬
opment of these basins, depending on their origin, nat¬
ural processes will begin to fill them. The sides will
erode outward and the eroded sediment will accumu¬
late in the basin, creating a broader, shallower depres¬
sion. This low area will also collect runoff and
sediment from the surrounding area. It may be wet
and swampy, possibly even forming a small lake,
where organic growth will fill the basin over time.
These processes, and others, may fill the sinkhole
completely, eventually leaving no sign of its previous
existence. Such a feature is now termed a buried sink¬
hole (Fig. 1C, part 1).
It is also possible for external sedimentation pro¬
cesses, which operate more rapidly than the rate at
which karst develops, to completely cover a karst ter¬
rain and fill all the surface depressions. In the northern
United States and Europe, many karst areas have been
blanketed by a thick mantle of glacial debris. In some
areas of Florida, the limestone is covered by a blanket
of marine or coastal sands, obscuring the underlying
karst until the moment when the ground surface col¬
lapses due to ongoing karstic erosion. Figure 1C,
part 1, shows a buried solution sinkhole. It is also pos¬
sible to have cover-collapse or cover-subsidence sink¬
holes buried. As Waltham (1989) points out, "... these
phenomena develop either by pre-burial subaerial ero¬
sion or subsequent subsoil erosion ..." (p. 35). That is,
the karstic erosion process will continue to operate
while the terrain is covered, creating modifications to
the unconsolidated strata overlying the limestone.
All four of the aforementioned types of sinkholes—
solution, cover-collapse, cover-subsidence, and bur¬
ied—are the result of the downward movement of
material along solutionally enlarged channels through
the limestone. However, the dissolution of the lime¬
stone by laterally flowing groundwater at depth may
also produce conduit drainage systems—caves. These
caves continue to grow as the groundwater flowing
through them continually dissolves the limestone,
over time making the passages larger and larger.
Simultaneously, the upper surface of the limestone is
also being lowered by solutional attack. These pro¬
cesses may eventually result in a situation where a
large cave is close to the limestone surface with only a
thin rock roof over the void. If this rock roof suddenly
collapses into the cave, a bedrock-collapse or cave-collapse
sinkhole is formed (Fig. IB).
Bedrock-collapse sinkholes are generally very steep
walled, or even overhung. They are often deeper than
they are wide and relatively cylindrical. Because the
walls are in solid rock, they will maintain this profile
for a long time, in terms of a human life span. Thus,
examples of bedrock-collapse sinkholes are not diffi¬
cult to find. Karst geologists generally agree that bed-
rock-collapse sinkholes are rarely observed in the act
of collapsing because the rate of surface down-wasting
is extremely slow on a human time scale. That is, bed-
rock-collapse sinkholes are rare, except on a geologic
time scale. The Florida Sinkhole Research Institute col¬
lected data on more than 1700 sinkholes that devel¬
oped in Florida, and not one of these was definitely a
bedrock-collapse sinkhole.
The Mitchell Plain of southern Indiana is an example
of the long-term product of karstic erosion. It is pock¬
marked with sinkholes. The largest and deepest of
them are generally formed over major, active cave pas¬
sages due to roof collapse. One of the rare examples of
the formation of a cave-collapse or bedrock-collapse
sinkhole within historic time occurred at the Colglazier
entrance to Blue Spring Cave, on the Mitchell Plain
southwest of Bedford, Indiana. What had formerly been
a shallow marshy depression grew into a cave-collapse
sinkhole nearly 2800 cubic meters in volume during a
heavy rainstorm in 1941. Although other cave-collapse
sinkholes occur across the Mitchell Plain over large
cave passages, the incidence of collapse in the human
timeframe is extremely rare, as mentioned above.
Europeans favor a different terminology. They favor
using the Anglicized Serbo-Croatian term doline, and they
include all collapse (a rapid process) dolines in one sub¬
group, within which they differentiate cave or rock col¬
lapse from overburden collapse. Sinkholes due to the
slow karstic erosion of cover (herein cover subsidence) are
called suffosion dolines. The suffosion process is defined as
the gradual winnowing and downwashing of fines by a
combination of physical and chemical processes.
According to European usage, suffosion dolines are small
depressions usually only a few meters in diameter and
depth. This specific example does not match the features
herein called cover-subsidence sinkholes, although it could
describe one possible example. The term suffosion is little
used in the United States, and the process of winnowing
and downwashing of fines leaving coarser materials in
place does not accurately describe the processes occurring
in the development of cover-subsidence sinkholes.
Therefore, we will eschew the European terminology in
favor of the American nomenclature, which appears to
more correctly characterize the erosion process.
Subtypes of sinkholes (dolines) are routinely defined
as if they occurred in isolation, but, in fact, in the evo¬
lution of a complex karst landscape these features are
stages in time rather than subtypes. Most sinkholes in
ENCYCLOPEDIA OF CAVES
SOLUTION CAVES IN REGIONS OF HIGH RELIEF
723
a mature, mantled karst landscape are polygenetic, and
most karst landscapes will eventually be mantled, at
least by residual soil, except for those few cases where
the area is underlain by hundreds of meters of very
pure limestone.
Solution and internal drainage are the basic para¬
digm underlying karst landscape development. As the
landscape evolves and cover sediments are deposited
(either as residuum from the insoluble residue in the
limestone or transported into the area from outside
sources), other erosional processes will also play a
role, integrated with the underlying paradigm. As the
cover sediments are eroded downward through the
internal drainage network, the solution process con¬
tinues accompanied by episodes of cover collapse,
infilling of karstic depressions, and possible cover
subsidence under the appropriate conditions. Where
enlarging subsurface karst features approach the top of
solid rock, cavern roof collapse can also impact the
landscape. The Mitchell Plain of southern Indiana is a
good example of such a complex karst landscape.
According to this concept, a sinkhole is part of a
long-term process and it may have different forms
and surface expressions at different times. When the
sinkhole is completely filled, there is no surface indi¬
cation of its presence, but the epikarstic drainage fea¬
ture still exists and continues to function below
ground. This has been termed a buried sinkhole, which
is a valid and appropriate concept despite the fact
that no depression exists during this stage of sinkhole
development. In recognition of the complex evolu¬
tionary process that characterizes the development of
karstic drainage and subsurface erosion, and in view
of the varying surface expressions that this process
may have as it evolves and the landscape matures,
and considering that under some conditions there
may be no ground surface expression of the underly¬
ing karst drainage, the traditional definition of a sink¬
hole as "a depression" appears inappropriate. The
following is a functional, process-oriented definition
of a sinkhole, or doline:
Sinkholes (dolines) are the surface (and near¬
surface) expressions of the internal drainage and ero¬
sion process in karst terrain, usually characterized by
depressions in the land surface. Formative processes
include bedrock solution, downward transport of over¬
burden sediment, and/or bedrock collapse. Most large,
mature sinkholes have a complex origin involving all
three processes. At some stages in sinkhole develop¬
ment, a surface depression may not be present.
See Also the Following Articles
Closed Depressions in Karst Areas
Epikarst
Bibliography
Beck, B. F. (1988). Environmental and engineering effects of sink¬
holes—The processes behind the problems. Environmental Geology
and Water Science, 12(2), 71—78.
Beck, B. F. (1991). On calculating the risk of sinkhole collapse.
In E. H. Kastning & K. M. Kastning (Eds.), Appalachian karst
(pp. 231—236). Huntsville, AL: National Speleological Society.
Ford, D., & Williams, P. W. (1989). Karst geomorphology and hydrolog]/.
London: Chapman & Hall.
Jennings, J. N. (1971). Karst. Cambridge, MA: The MIT Press.
Parise, M., & Gunn, J. (2007). Natural and anthropogenic hazards in karst
areas. Geological Society of London.
Sowers, G. F. (1996). Building on sinkholes. New York: ASCE Press.
Sweeting, M. M. (1972). Karst landforms. London: Macmillan.
Waltham, A. C. (1989). Ground subsidence. Glasgow: Blackie & Sons.
Waltham, T., Bell, F., & Culshaw, M. (2005). Sinkholes and subsidence.
Chichester, U.K.: Springer.
White, W. B. (1988). Geomorphology and hydrology of karst terrains.
New York: Oxford University Press.
Williams, P. W. (2003). Dolines. In J. Gunn (Ed.), Encyclopedia of caves
and karst science (pp. 304—310). New York: Fitzroy Dearborn.
SOLUTION CAVES IN REGIONS
OF HIGH RELIEF
Philipp Hauselmann
Swiss Institute for Speleology and Karst Studies
La Chaux-de-Ponds, Switzerland
INTRODUCTION
The topic of solution caves in lrigh-relief areas,
although they are caves constrained by two conditions,
is remarkably broad and variable. It is therefore diffi¬
cult to choose good examples of the many existing
caves. We restrict ourselves voluntarily to epigenic
caves, formed by rainwater percolating from above,
because, although recent research in hypogenic caves
revealed that they are much more numerous than pre¬
viously thought (Klimchouk, 2007), virtually all large
cave systems in mountainous ranges are epigenic and
may show only some minor hypogenic chambers or
passages.
It is apparent that there is no definite geomorpho-
logical trait that characterizes all these caves; on the
contrary, they seem even more diverse than the coun¬
tryside that surrounds them. Therefore, it is not possi¬
ble to define what caves in mountains have to look
like. However, caves in high-relief areas share some
common traits that are not directly linked to, but do
influence, morphology. They are summarized below.
It is broadly known that caves are natural archives
that register events in the Earth's history. It is less
known that this conservation of events is not only
reflected in sediments (deposition), but also in the
Encyclopedia of Caves.
© 2012 Elsevier Inc. All rights reserved.
724
s
cave's morphology (erosion). Both are equally impor¬
tant if one wants to know what happened in time. The
difficulty is often to link sedimentological and mor¬
phological observations. An article on the importance
of studying caves therefore gives a broad recipe.
Examples of widely known caves then illustrate the
diversity of settings and morphology. In many cases it
is apparent that the cave's morphology is not sufficient
to understand the morphogenesis of the landscape
without taking into account other geological param¬
eters. However, other examples show that insight into
cave-forming processes is the only tool to reconstruct
the evolution of the surface.
Literature research reveals that the geomorphology
of many alpinotype caves is only known in its broad
outlines; in many cases, the morphological description
amounts to one or two phrases in the text. This fact
makes it very difficult to review their geomorphologi-
cal traits and significance, but on the other hand, it
shows how much research is still needed to under¬
stand these caves. Consequently, the last paragraph,
although entitled "Conclusions," also gives ideas for
future research.
GENERAL CHARACTERISTICS OF CAVES
IN HIGH-RELIEF AREAS
Caves in high-relief areas differ from caves in the
lowlands in several aspects. Infiltrating water tries
always to flow down vertically until it reaches either
impervious strata or the base level. In high-relief areas,
large and long shafts and steep meandering canyons
are created this way. In a general perception, a cave
with many shafts and meandering canyons is com¬
monly regarded as alpinotype, although it is by no
means restricted to the Alps. On the contrary, in low¬
lands, caves are most usually more or less flat, with
small shafts between, and meandering canyons tend to
be subhorizontal.
Due to orogenetic processes, the limestone strata are
seldom subhorizontal; most of the time, the strata are
more or less inclined. Because many caves preferen¬
tially develop on bedding planes (Lowe, 1992;
Filipponi, 2009), eventually also at the intersection
of bedding planes with faults and fractures, cave
passages almost invariably are steeply inclined. This is
evident for vadose passages following stratal dip, but
it is commonly also the case for phreatic passages fol¬
lowing strike. The permeability of the bedding plane
being inhomogeneous, the water seeks its way along
the most permeable places (Ford and Williams, 2007)
and thus wiggles along the bedding plane, creating the
so-called looping passages (Fig. 1). "Looping" in this
case does not necessarily mean "making a closed
loop," although lateral passages that do loop often
exist.
The temperature of the infiltrating water is lower in
mountain chains. Colder water can contain more CO 2
(and thus may be more aggressive), whereas the lower
temperature should slow down the reaction kinetics.
Therefore, it would be expected that the water in high-
relief areas retains more aggressivity over a longer dis¬
tance, thus enhancing speleogenetic processes in depth.
However, the availability of CO 2 is not only due to
water temperature, but also strongly influenced by veg¬
etation at the surface (Bogli, 1978). This vegetation is
reduced or absent in high altitudes. Moreover,
investigations (Jeannin, 1990) have shown that the
water temperature relatively quickly adjusts to the cave
temperature. The subject has still to be investigated.
Glaciers at the surface in high mountains produce a
high load of sediment suspended in water, which then
is often called glacier milk. This enriched water has two
effects on caves: where it flows torrentially through
passages, it increases mechanical erosion and may dig
out large canyons with polished walls. Where the
water velocity gets very low, the sediment may settle
out and eventually obstruct the passages, thus creating
the silty fillings that are found in many caves in the
Alps and elsewhere (Bini et al., 1998). The absolute
effect of glacier milk on passage enlargement is still
unknown.
Climatic variation in mountain chains is usually much
larger than in lowlands. In wintertime, precipitation falls
as snow, and there is almost no water available. In
springtime, snowmelt releases huge amounts of water in
a comparably short time, the caves are often flooded,
and discharge is near its maximum. In summertime,
thunderstorms form, releasing huge precipitation. There
is an ongoing debate as to whether climate and its vari¬
ability has a notable influence on cave-forming processes
(and whether such influence can be seen in passage
morphology).
FIGURE 1 A looping passage formed along an inclined bedding
plane. Such phreatic and epiphreatic passages enlarge at those places
where permeability of the bedding plane is largest.
ENCYCLOPEDIA OF CAVES
SOLUTION CAVES IN REGIONS OF HIGH RELIEF
725
WHY IS IT IMPORTANT TO STUDY
CAVES IN HIGH-RELIEF AREAS?
In the lowlands, surface erosion is commonly slow
and, in some cases, even negligible. Therefore, geo-
morphological markers of the past are often recogniz¬
able and present. Commonly, these superficial forms
can be linked to cave levels and cave sediments.
Dating of one sedimentary deposit then gives
the age range of the other one. The interrelation
between surface and caves is often relatively clear in
the lowlands.
On the other hand, mountain chains generally have
large erosion rates. In midlatitudes, glaciation is another
way of erosion that often leaves no trace of prior events.
The overall result is that mountainous regions do very
rarely present traces of older evolutionary stages. In
order to reconstruct the geomorphic evolution of the
surface in high-relief areas, caves are therefore the only
means to reach this goal. It also becomes evident that
caves are a valuable tool for geomorphology and
Quaternary geology (Audra et al., 2006).
THE RELATIVE CHRONOLOGY
A considerable problem exists with caves being a
hollow form within the parent rock, so they cannot be
dated in themselves. On the other hand, the exact mor¬
phology of the void contains much information about
processes that took place at the surface. It is thus very
important to date the void. This can be done by apply¬
ing the concept of relative chronology.
Relative chronologies are based on the interrelation
between cave morphology (which expresses the
removal of rock) and the sediments contained within
the cave (which expresses depositional events). Within
a sedimentary pile, the laws of stratigraphy can be
applied, which in the principle of superposition indi¬
cate that a younger sediment overlies an older one.
Morphological indications, on the other hand, also
give chronological information. A keyhole passage in
an epigenic cave informs us that a phreatic phase was
followed by a vadose one. Successions of speleogenetic
phases are found in many cave systems. This in itself
is also chronological information.
The task now is to connect the sediments of several,
basically independent, sedimentary profiles and to link
them with the morphological succession of the cave
passages. Thus, the sedimentary profiles are not inde¬
pendent from each other, and a relative chronology of
erosional and depositional events over the whole cave
can be made.
Example of a Relative Chronology
Figure 2 shows a real situation encountered in St.
Beatus Cave, in Switzerland.
To the right side is a typical keyhole passage which
proves that a phreatic initiation of the ellipse on top
was followed by a canyon incision. In the middle part
of the figure, the meander gradually disappears and is
replaced by a more or less elliptic passage that con¬
tinues toward the left side of the figure. We see there¬
fore a transition of a vadose feature into a phreatic
one, and thus an old water level. In the profile to the
right, we observe flowstone deposition that was trun¬
cated by the river incising the meander. Therefore, the
flowstone predates the canyon, but postdates the initial
genesis of the elliptic passage to the right. The mean¬
der changes into an elliptic passage; thus the two
forms are contemporaneous. Consequently, the older
flowstone disappears in the area of this transition.
Within all the passages, silts were deposited. They are
younger than the meander incision, and younger than
the passage to the left, and proof of an inundation of
the whole cave. Stalagmites grow on the silts and are
partially still active. This example can be written as a
table. For practical reasons, that table is not rewritten
with each sedimentary succession found. Instead, the
single sedimentary sequence is coupled with morphol¬
ogy, and is written as a column in the table. The next
FIGURE 2 A schematic (but real) example to establish a relative chronology between morphology (erosional events) and sediments
(depositional events) within a cave. (For explanation see text.)
ENCYCLOPEDIA OF CAVES
726
s
TABLE 1 Chronological Table with Columnar Writing of
Figure 2
Sequence at Left
Sequence at Right
Phreatic genesis of top ellipse
Water level lowering
Phreatic genesis
Deposition of flowstone
Phreatic genesis
Erosion of flowstone
Phreatic genesis
Erosion of meander
Water level lowering
Silt deposition
Silt deposition
Stalagmite growth
Stalagmite growth
sedimentary sequence, again coupled with morphol¬
ogy, is written as another column. Thus, the above
example would then look like Table 1. If we continue
upstream and downstream of that profile, we find sev¬
eral other morphological indications and sedimentary
successions, each of them having a link with our initial
profile until we encounter the next paleowater level
and thus the next morphological change. There, the
links have to be established again. The table thus
slowly grows and gets more complete.
When the basic table is established, absolute dating
is used to assign ages to the respective formations.
This has a twofold benefit: on the one hand the dates
ascertain that the relative chronology is consistent,
because the absolute ages should be concordant with
the relative ones. On the other hand, the relative chro¬
nology can be used to assess whether a date is to be
trusted or whether it might have experienced changes,
for example, U leakage from speleothems.
EXAMPLES OF CAVES
Caves in high-relief areas can be divided into three
main groups: (1) large, multiphase labyrinths, often of
considerable age; (2) vadose shafts and meandering
canyons leading steeply down; and (3) long inclined
passages. Of course, there is almost always some mix
of these types. Nevertheless, these three groups cover
virtually all cave types in mountain ranges.
EXAMPLES OF CAVES:
LARGE LABYRINTHS
Considering that relief and thus hydrogeological
gradients are steep, one might expect mostly vertical
to steeply inclined caves in mountainous regions.
However, some of the densest labyrinths of the
world's known caves develop in the mountains. Most
of these labyrinths contain passages of different ages,
and typically represent floodwater mazes that are
often intersected by younger vadose passages.
Holloch Cave, Switzerland
Holloch Cave was, in the 1960s, considered to be the
longest cave in the world. At about the same time, Alfred
Bogli executed considerable scientific work within this
cave. A summary of his work around Holloch Cave is
published in a well-known book on karst hydrography
and physical speleology (Bogli, 1978). Holloch Cave is sit¬
uated in central Switzerland. The main entrance is at an
altitude of 733 m asl. The cave develops in Schrattenkalk
(Urgonien, Lower Cretaceous) of the Helvetic realm.
Holloch Cave represents the lower end of a vast
catchment area extending between 700 and 2300 m in
altitude. Whereas the entrance of Holloch is dry in nor¬
mal seasons, it may flood during exceptionally wet
times. This fact indicates that Holloch Cave has a huge
flooding zone that may reach more than 320 m in the
far parts of the cave. In fact, 2/3 of the currently
explored 195 km can be flooded during exceptional
events (Battig and Wildberger, 2007; Fig. 3).
As it follows from the floods, and as shown on
Figure 3, Holloch can be considered an anastomotic
floodwater maze following Palmer (1991). This is very
nicely seen also by its passage morphology: elliptical
tubes of phreatic or floodwater origin abound every¬
where (Fig. 4); entrenched meandering canyons occur
only very rarely.
Bogli (1970), on the base of distribution of the large
passages, concluded that Holloch was made by three
speleogenetic phases that were conditioned by valley
deepening in the spring area. However, in Bogli's
time, no absolute dating was possible. A recent paper
(Wildberger et al., 2010) tried to couple different
obtained age information (absolute and relative) with a
general geomorphological evolution of the surface, in
order to get ideas about the possible onset of karstifica-
tion in the region of Holloch. The results are yet incon¬
clusive; however, karstification of the area should have
begun at around 3 Ma BP.
Reseau Siebenhengste-Hohgant, Switzerland
Reseau means network—and this is a good descrip¬
tion of the cave system of Siebenhengste. Together
with many caves that are still not connected to the
main system, the area comprises 320 km of cave pas¬
sages, to which the Reseau itself contributes 156 km
with a depth of 1340 m. This makes the area of
Siebenhengste one of the large cave areas of the world.
ENCYCLOPEDIA OF CAVES
SOLUTION CAVES IN REGIONS OF HIGH RELIEF
727
FIGURE 3 Map and projected section of Holloch Cave (Switzerland) and neighboring Silberen System. Plan pattern as well as heights of
flood (projection) clearly show that Holloch Cave can be qualified as an epiphreatic floodwater maze. Adapted from Wildberger et al., 2010.
FIGURE 4 A passage in Holloch cave reveals its phreatic form
that is still enlarged in the epiphreatic zone. This passage gets regu¬
larly flooded, especially in snowmelt and in summer thunderstorms.
Photo courtesy of J. Godat.
The Siebenhengste area is located in western-central
Switzerland, north of Lake Thun. The caves occur
between 460 m asl (100 m below the level of the lake).
and the summit of Siebenhengste at 1950 m. The system
develops in Schrattenkalk (Urgonian, Lower Cretaceous).
A general view of the cave area is given in Figure 5.
The Siebenhengste caves are quite diverse in their
character due to different genesis and age. In the fol¬
lowing paragraphs, we describe the three main mor¬
phological parts.
The caves FI and K2 (Fig. 5) show a typical den¬
dritic pattern (Palmer, 1991). They are caves with very
minor phreatic development: most of the passages are
of vadose origin.
The labyrinth of the Siebenhengste itself represents
a dense maze of old, phreatic tubes interlaced with
many younger shafts and meandering canyons (Hof
et al., 1984). Morphological research revealed that there
were five speleogenetic phases responsible for the gen¬
esis of the different phreatic passages.
From the labyrinth, mostly vadose rivers flow
downward toward the zone profonde (deep part) of the
Siebenhengste, which again forms a 3D labyrinth with
vadose shafts and meandering canyons coming from
the surface. From there onward, an old phreatic tube
that has its origin in the downstream part of FI con¬
tinues toward Faustloch and farther to the southwest.
ENCYCLOPEDIA OF CAVES
728
FIGURE 5 Caves in the Siebenhengste area, Switzerland. The map shows a total of 320 km of passages, of which 156 km are
currently interconnected. The difference in plan pattern between dendritic (FI, K2) and phreatic multilevel mazes (Barenschacht) can clearly
be seen.
ENCYCLOPEDIA OF CAVES
SOLUTION CAVES IN REGIONS OF HIGH RELIEF
729
Barenschacht is, like Holloch, a floodwater maze
cave. While several speleogenetic phases can be recog¬
nized in Barenschacht (Hauselmann et ah, 2002), there
are almost no vadose meandering canyons and afflu¬
ents from the surface that is covered by impermeable
flysch. Therefore, all the maze is of (epi)phreatic
origin.
The Reseau Siebenhengste-Hohgant is one of the
few well-explored cave systems where a lot of scien¬
tific work has been done. Geomorphological investiga¬
tions of the passage morphology revealed the existence
of 14 speleogenetic phases (Hauselmann et ah, 2002).
All these phases can be linked with ancient valley
floors, and therefore give valuable information on the
evolution of the surface outside. The uppermost (and
therefore oldest) of the 5 levels had their spring in the
northeast of the Siebenhengste labyrinth in the Eriz
valley (Fig. 5). The following 9 speleogenetic phases,
on the contrary, had their spring in the Aare valley.
Dating of the several cave phases by cosmogenic iso¬
topes and U-Th series indicated that the uppermost 5
phases were active between more than 4.4 and 0.8 Myr
BP, and that the valley incision rate during that period
was around 0.12 mm yr -1 (Hauselmann et ah, 2007).
The subsequent incision of the Aare valley, between
0.8 Myr BP and today, happened 10 times faster with
an incision rate of roughly 1.2 mm yr -1 .
EXAMPLES OF CAVES: VERTICAL TO
SUBVERTICAL SHAFTS AND
MEANDERI NG CA NYONS
Although the labyrinths presented above all occur
within the mountains, the shaft-and-canyon type is
considered a typical alpine cave. By number, these caves
exceed all other types by far. For instance, all but three
of the 176 caves in the Innerbergli karren field
(Siebenhengste, Switzerland) contain mostly shafts and
meandering canyons.
Many of these shaft systems may be smallish. To
develop vertical or subvertical systems of more than
500 m in altitudinal difference, the limestone thickness
(or its folds) has to be sufficiently large to allow that.
This is not everywhere the case.
Krubera, Georgia
The currently most famous of these shaft systems is
Krubera (Voronja) cave, Georgia. By its depth of currently
—2191 m, it is the world's deepest cave (Klimchouk et ah,
2009). A succession of shafts and steep canyons leads
down to several sumps which were dived to show that
the passages continue behind. This is therefore a typical
alpine shaft system, albeit of exceptional dimensions
(Fig. 6).
Krubera Cave is situated in the Arabika massif
(Western Caucasus, Georgia). The entrance opens at
2256 m asl within a glacial trough. The rocks are com¬
posed of Upper Jurassic and Lower Cretaceous lime¬
stones that are folded and complexly faulted and that
dip continuously down to the Black Sea shore.
The cave map makes very quickly clear that the cave
is essentially vertical: a total depth of —2191 m and a
length of about 14 km lies within an area of only circa
700 X 400 m. Geomorphological observations point out
that most of the cave was formed in the vadose zone;
only minor passages seem to have their origin in an
ancient phreatic zone (Fig. 7). This in itself is not very
surprising for the given geologic setting. However, two
facts make the cave interesting from the geomorphologi¬
cal point of view: first, the very low gradient between
the water table occurring in the cave (110 m asl) and the
spring area distant around 12—13 km at the Black Sea
coast, and second, U-Th dating of stalagmites occurring
at a depth of 640 and 436 m asl that yielded an age of
between 200 and 300 ka. These age dates reveal that
these parts of the cave were already in the vadose zone
at that time. The current knowledge indicates that the
initiation of the upper phreatic tubes could have hap¬
pened during the Messinian salinity crisis, and the
vadose cave formed during the Plio-Pleistocene, more or
less concordant with uplift, which was very huge in the
central parts of the massif (estimated to be 2—2.5 km
during Plio-Pleistocene; Fig. 7).
Krubera Cave shows that not only phreatic tubes
connected to paleosprings can be used for reconstruct¬
ing the paleogeomorphology of a landscape, but in ideal
cases, also vadose caves are very valuable. Therefore,
scientific observations in all kinds of alpinotype caves
are to be promoted.
EXAMPLES OF CAVES: LONG
INCLINED PASSAGES
Caves that present long inclined passages are almost
invariably caves that develop along a geological bound¬
ary with impermeable rocks; they therefore always
contain an active river. Generally, these caves are char¬
acterized by a vertical shaft zone where the waters
transect the limestone in the most direct way before
reaching the impervious layers. Often, the limestone
thickness is rather small (150—250 m)—as a matter of
fact, many of the deepest caves of the world actually
belong to this cave style. The most famous of them is
Gouffre Berger (Vercors, France), which was, in 1956,
the first cave known to reach —1000 m (Caillault et ah,
1997).
ENCYCLOPEDIA OF CAVES
730
FIGURE 6 Projected section of the deepest cave of
the world, Krubera (Voronja), Caucasus, Georgia. The
cave is mainly a succession of steep meandering can¬
yons and shafts. Almost all the cave is of vadose origin,
with the exception of the subhorizontal branch at
around -300 m, which is interpreted as being a possible
Messinian cave.
FIGURE 7 Genetic interpretation of the genesis of Krubera Cave in connection with recent uplift of the Caucasus range. From Klimchouk
et al., 2009.
ENCYCLOPEDIA OF CAVES
SOLUTION CAVES IN REGIONS OF HIGH RELIEF
731
FIGURE 8 Projected cross-section of Jean-Bernard Cave (France). Several entrances (with numbers) connect to partly vadose, partly phre¬
atic channels. The phreatic cave had its spring in the area of the entrance V4. The rest of the cave is essentially vadose and formed by a canyon
which is sometimes very high (80—100 m). Adapted from Lips et at., 1993.
The genesis of this cave type is actually favored by
orogenetic processes that overthrust (and fold) lime¬
stone nappes onto other rocks and so create the possi¬
bility of large altitudinal differences within the same
(thin) limestone sequence.
The Jean-Bernard Cave (France)
The Gouffre Jean-Bernard was the world's deepest
cave in the beginning of the 1990s. It is situated in the
Pre-Alps in the east of France. Its several entrances are
between 1840 and 2320 m asl. The cave develops in
Urgonian limestones of 150- to 200-m thickness, which
are overthrust, folded, and faulted. The cave follows a
minor syncline within the limestone series.
For many reasons, the cave is again considered an
intermediate system between the simple shaft-canyon
type (as is Gouffre Berger) and a complex labyrinth
like the Siebenhengste: its upper part consists not only
of vadose shaft systems rapidly reaching the impervi¬
ous base level (Fig. 8), it also contains ancient phreatic
tubes that actually had their paleospring in the lower
entrances (Lips et al., 1993). The lower part of the cave
contains one large single meandering canyon, some¬
times up to 100 m high, where the river flows. This
river sinks between breakdown; the following passage
leading farther downward is of much smaller size.
Jean-Bernard Cave has the peculiarity that the low¬
ermost end of the cave (716 m asl) is lower than the
altitude of the resurgence (780 m asl). This caused a
vigorous debate as to whether this would be possible
or whether this is simply caused by topographic
errors. The current idea is that the mapping is correct
and that indeed the cave is deeper than its spring,
because the river diverts laterally from the cave and
resurfaces on the flank of the syncline, whereas the
small annex passage goes farther downward along the
synclinal axis and was drained by a now-abandoned
spring downstream from the present one.
There is at present no comprehensive dating of the
cave that permits determination of its age. A hypothe¬
sis is that the ancient phreatic passages occurring on
top could have formed before the Messinian salinity
crisis. The only proven fact is that when these passages
form, the valleys and the overall geomorphology of
the Alps in that region has to have been completely
different from their aspect of today.
Cassowary System (Papua New Guinea)
The Cassowary System is situated in the island of
New Britain (Papua New Guinea), at an altitude of
1445 m asl (Audra et ah, 2001). The system is, with
— 1178 m in altitudinal difference, the deepest cave of
the Southern Hemisphere; it can be followed from the
ponor entrances all the way to the spring at 267 m asl
(Fig. 9). The cave is formed within pure, only slightly
compacted, Miocene limestone, overlain by alteritic
residues of former volcanoclastic rocks. The area was
subject to vigorous uplift beginning in the Upper
Pliocene.
The cave character differs from the general notion
of shafts going down to an impervious layer and a
meandering passage following it to the spring: first,
there is no impervious layer present and then the cave
presents only very few shafts: it is mainly composed of
a long, inclined tube going from ponor to resurgence
that was later incised to form a canyon. No ancient
phreatic system is present; the cave was therefore
formed in a single phase, which suggests a recent
ENCYCLOPEDIA OF CAVES
732
s
Arcturus Muruk
FIGURE 9 Projected cross-section of Cassowary System (Papua New Guinea). The cave monotonously dips down toward its spring in
the Galowe Canyon. Dating as well as morphological consideration led to the hypothesis that the cave must be very young. Adapted from
Audra et al., 2001.
development. The plan view supports this hypothesis:
the cave goes practically straightforward from ponor
to spring with basically no lateral wiggles.
The Nakanai mountains, where the cave is situated,
present special conditions for speleogenesis and karsti-
fication: the presence of pure, but only slightly consoli¬
dated, limestone offers high potential for both active
corrosion as well as mechanical erosion. The rainfall in
the area (10 to 12 m yr _1 !) provides huge amounts of
waters, which furthermore are concentrated at the sur¬
face due to the argillaceous alteritic cover of the former
volcanoclastic rocks. The presence of a rainforest with
a highly productive growth increases the CO 2 content
of the water to a considerable extent, and the vigorous
recent uplift of the area is responsible for a high
hydraulic gradient.
Dating of some speleothems gave ages of less than
50 ka BP, and paleomagnetic analyses of clay sedi¬
ments showed normal polarity. The age of the cave
system is therefore less than 780 ka BP. Audra et al.
(2001) suppose an age of 100 to 200 ka BP. The cave
system is therefore an example of a recent, single¬
phase ponor-to-spring system. In that respect, it is a
good example of an extreme end-member of cave
formation.
CONCLUSIONS
From the above examples, it is clear that there is no
simple geomorphological trait that links all the caves
in high-relief areas together. While the basic cave¬
forming processes are always the same, every cave is
dependent on local conditions.
When reading speleological literature about alpino-
type caves, it appears that very many caves actually
are vertical shafts without fossil passages. The ques¬
tion is now whether this is speleological reality (there
are actually quite few fossil passages) or whether this
is exploration bias (many cavers are not geomorpholo¬
gists and simply want to go as deep as possible, so all
possible horizontal passages are neglected). My per¬
sonal experience rather tends toward the second expla¬
nation: very many shaft caves I know do in reality
intersect (or follow) fossil phreatic passages. As a first
conclusion, one can thus say that caves have to be
explored more thoroughly, and complete and well-
drawn maps (and longitudinal sections) that contain
all relevant information must be made.
When literature is searched specifically for scientifi¬
cally interesting information on caves in high-relief
areas, the reader is left with some few articles com¬
pared to the mass of such caves that are explored. In a
general sense, there is still much fieldwork to be done
almost everywhere. Rewards are most certain, because
caves now are known to be archives of the past,
recording not only valley deepening phases, but also
in registering climatic variations, both in speleothems
and in other sediments (and even in their passage mor¬
phology). Therefore, the second conclusion follows:
almost everywhere there is still a huge amount of
work to be done.
As stated above, caves everywhere, but especially in
high-relief areas, are very important for reconstructing
the evolution of the landscape outside the cave. They
are actually, in very many places, even the single best
possible means to get insight into valley deepening
processes, glaciations, and climatic variations. Such
research is not purely academic: caves may offer
insight into incision rates of valleys, which may directly
condition the potential suitability of, for instance, a
deposit of radioactive waste planned by the govern¬
ment. Now, much of this work cannot be done by "tra¬
ditional" geomorphologists, simply because they
cannot reach the cave. Either cavers become geomor¬
phologists, or the geomorphologists teach the cavers to
ENCYCLOPEDIA OF CAVES
SOLUTION CAVES IN REGIONS OF MODERATE RELIEF
733
bring as much information as possible back to the sur¬
face. Experience has shown that it may work! If "nor¬
mal" scientists regard cavers no longer as "dirt-loving
and strange," but as collaborators, general knowledge
will improve tremendously. The third conclusion: do
collaborate!
While caves in high-relief areas are often not very
easily accessible, they may offer some advantages to
caves in the lowlands. For instance, due to the gener¬
ally more diverse climate, looping caves present larger
loops, which make it easier to decipher speleogenetic
processes and to gain insight into basic principles of
speleogenesis. In following years, it is expected that
there will be important discoveries.
See Also the Following Article
Solution Caves in Regions of Moderate Relief
Bibliography
Audra, P., Bini, A., Gabrovsek, F., Hauselmann, Ph., Hoblea, F.,
Jeannin, P-.Y., Kunaver, ]., Monbaron, M., Sustersic, F., Tognini, P.,
Trimmel, H., & Wildberger, A. (2006). Cave genesis in the Alps
between the Miocene and today: A review. Zeitschrift fur
Geomorphologie N.F., 50(2), 153—176.
Audra, P., de Coninck, P., & Sounier, J.-P. (2001). Nakanai—20 ans
if exploration. Antibes: Association Hemisphere Sud.
Battig, G., & Wildberger, A. (2007). Ein Vergleich des Holloch-
Hochwassers vom August 2005 mit seinen Vorgangern. Stalactite,
57(1), 26—34 [in German],
Bini, A., Tognini, P., & Zuccoli, L. (1998). Rapport entre karst et gla¬
ciers durant les glaciations dans les vallees prealpines du Sud des
Alpes. Karstologia, 32(2), 7—28 [in French],
Bogli, A. (1970). Das Hollocli und sein Karst. Neuchatel: La Baconniere.
Bogli, A. (1978). Karsthydrographie und physische speldologie. Berlin:
Springer Verlag.
Caillault, S., Haffner, D., & Krattinger, T. (1997). Spele'o dans le
Vercors. Aix-en-Provence: Edisud.
Filipponi, M. (2009). Spatial analysis of karst conduit networks and deter¬
mination of parameters controlling the speleogenesis along preferential
lithostratigraphic horizons. Switzerland: Ecole Polytechnique
Federale de Lausanne.
Ford, D. C., Ik Williams, P. (2007). Karst hydrogeology and geomorphol¬
ogy. Chichester, U.K. Wiley.
Hauselmann, P., Granger, D. E., Jeannin, P.-Y., & Lauritzen, S.-E.
(2007). Abrupt glacial valley incision at 0.8 Ma dated from cave
deposits in Switzerland. Geology, 35(2), 143—146.
Hauselmann, P., Jeannin, P.-Y., Monbaron, M., & Lauritzen, S.-E.
(2002). Reconstruction of Alpine Cenozoic paleorelief through the
analysis of caves at Siebenhengste (BE, Switzerland). Geodinamica
Acta, 15, 261-276.
Hof, A., Rouiller, P., & Jeannin, P.-Y. (1984). Le Reseau.
Hohlenforschung im Gebiet Sieben Hengste-Hohgant, 0, 1—106.
Jeannin, P.-Y. (1990). Temperatures dans la zone vadose du karst.
Bidletin du Centre d'Hydroge'ologie de VUniversite de Neuchatel, 9,
89-102 [in French],
Klimchouk, A. B. (2007). Hypogene speleogenesis: Hydrogeological and
morphogenetic perspective. Socorro, NM: National Cave and Karst
Research Institute.
Klimchouk, A. B., Samokhin, G. V., & Kasian, Y. M. (2009). The deep¬
est cave in the world in the Arabika massif (Western Caucasus)
and its hydrogeological and paleogeographic significance.
In Proceedings of the 15th International Congress of Speleology
(pp. 898-905). Kerrville, TX.
Lips, B., Gresse, A., Delamette, M., & Maire, R. (1993). Le Gouffre
Jean-Bernard (-1602 m, Haute-Savoie, Fr.). Karstologia, 21(1),
1-14.
Lowe, D. J. (1992). The origin of limestone caverns: An inception horizon
hypothesis. : Manchester Polytechnic.
Palmer, A. N. (1991). Origin and morphology of limestone caves.
Geological Society of America Bulletin, 103, 1—21.
Wildberger, A., Geyh, M., Groner, U., Hauselmann, Ph., Heller, F., &
Ploetze, M. (2010). Dating speleothems from the Silberen Cave
system and surrounding areas: Speleogenesis in the Muota Valley
(Central Switzerland). Zeitschrift fur Geomorphologie, 54(2),
307-328.
SOLUTION CAVES IN REGIONS OF
MODERATE RELIEF
Arthur N. Palmer
State University of New York
Caves in regions of moderate relief are widely regarded
as the standard to which all others are compared.
Tectonic stability is their most significant characteristic.
The presence of moderate relief implies that uplift of the
land is slow and erosional processes are able to keep
pace. As a result, rivers easily erode to their local base
levels, and base-level control is reflected in cave pas¬
sages. The relation between caves and regional geomor-
phic history is stronger than in any other setting.
CHARACTERISTICS OF KARST REGIONS
OF MODERATE RELIEF
In this article, the standard cave-forming process is
considered to be dissolution of carbonate rocks by
groundwater rich in carbonic acid derived from the
soil. Some solution caves in moderate-relief karst are
the product of quite different mechanisms (e.g., by sul¬
furic acid or by gypsum dissolution) and are discussed
in other articles.
Relief by itself has little influence on cave origin. For
example, alpine caves form by the same mechanisms as
those in moderate relief, but they show different rela¬
tionships to the landscape and to the geologic structure.
Most regions of moderate relief have a well-developed
soil cover, except where it has been lost through glacial
scour, deforestation, or overgrazing. Soil is the source
for most of the carbonic acid that allows groundwater to
dissolve caves, and it enhances the deposition of carbon¬
ate speleothems where groundwater seeps into aerated
caves. It also affects the nature of groundwater recharge,
and therefore cave patterns.
Encyclopedia of Caves.
© 2012 Elsevier Inc. All rights reserved.
734
GEOGRAPHIC DISTRIBUTION
s
most caves are of alpine, thermal, sulfuric acid, and
volcanic origin, or have formed in gypsum.
Regions of moderate relief include plateaus of rela¬
tively undeformed rock, as well as the roots of
deformed mountains that uplifted hundreds of mil¬
lions of years ago and are tectonically nearly stable
today. These landscapes are well illustrated by the low
karst plateaus of the east-central United States and the
Appalachian Mountains (Fig. 1). These were the main
field areas for the classic American karst studies of the
late nineteenth century through the mid-twentieth cen¬
tury. Base-level and water-table control of cave devel¬
opment were topics of great weight among American
karst scientists, but were far less conspicuous in
European studies. Many of these differences can be
traced to the strong focus on alpine karst by most
European karst scientists. Nevertheless, many impor¬
tant karst areas throughout the world are in plateaus
and ancient mountains of moderate relief.
The most cavernous states in the United States,
ranked by number of known caves, are Tennessee,
Kentucky, West Virginia, Alabama, Virginia, and
Missouri. Each has more than 3000 significant caves,
and all are in the moderate-relief interior plateaus or
Appalachian Mountains. The extensive Florida karst
shares many of the same cave-forming processes,
although it has much lower relief and is influenced by
the unusually high porosity of its carbonate rock and
in local areas by mixing of freshwater and seawater. In
the mountainous and semi-arid western United States,
Generalized geologic cross-section through line on map
I I Mississippian
Wl Silurian-Devonian
■■ Cambrian-Ordovician
mainly insoluble
FIGURE 1 Map of karst regions of the 48 conterminous United
States, showing areas discussed in the text. A = margins of the
Illinois Basin (Indiana and Kentucky); B = Ozark Plateaus of
Missouri and Arkansas; C = Appalachian Plateaus; D = Ridge and
Valley Province (folded Appalachians); E, E = karst areas modified
by continental glaciation. The cross-section shows the distribution,
simplified structure, and ages of the three major cavernous rock
sequences.
ORIGIN OF SOLUTION CAVES
Most solution caves in carbonate rocks are formed
by the circulation of groundwater from upland sur¬
faces through a network of fractures, partings, and
intergranular pores, to springs at lower elevations.
Groundwater recharge takes place in several ways,
including diffuse seepage through soil, small trickles
delivered by sinkholes, and concentrated inflow as
sinking streams.
As water passes through the atmosphere and soil, it
absorbs carbon dioxide, which combines with the
water to form carbonic acid. This acid, though weak, is
abundant and is the most common agent of carbonate-
rock dissolution. Where the water first enters lime¬
stone or dolomite, it dissolves all available openings in
the rock at nearly uniform rates, producing a maze of
widened fissures and pores (the epikarst). In areas of
moderate relief, most of these openings are soil-filled.
Water drains from the base of the epikarst through a
relatively small number of major conduits, which
deliver solutionally aggressive water deep into the sol¬
uble rock. It is along these paths that caves form. As
the openings enlarge, soil subsides into them, forming
sinkholes at the surface that funnel increasing amounts
of water into the ground as their catchment areas
expand.
Water descends by gravity along the steepest avail¬
able openings until it reaches the zone where all open¬
ings are filled with water. The top of this zone is the
water table. Above the water table is the vadose zone, and
below it is the phreatic zone. (Most groundwater specia¬
lists prefer to call them the unsaturated and saturated
zones, but these terms blur the distinction between
hydrologic and chemical saturation.) Cave passages that
form in the vadose zone have continuously descending
profiles that typically consist of canyon-like passages
interrupted by vertical shafts, so that the resulting pat¬
tern is step-like. Canyons develop mainly along the part¬
ings between rock strata and grow by entrenchment of
their floors by flowing streams (Fig. 2). Shafts form
where the descending water follows steeply inclined or
vertical openings, mainly joints or faults.
At and below the water table, gravity is offset to
varying degrees by the downward increase in hydro¬
static pressure. Phreatic water follows the paths of
least resistance to the nearest available outlets, usu¬
ally river valleys. Passages of phreatic origin are
tubular or fissure-like conduits with low overall gra¬
dients (Fig. 3), and most have looping profiles that
fall and rise irregularly along their length. Because
ENCYCLOPEDIA OF CAVES
SOLUTION CAVES IN REGIONS OF MODERATE RELIEF
735
FIGURE 2 A typical canyon passage in McFail's Cave, New York,
formed by vadose flow.
fractures and partings tend to become narrower with
depth, most phreatic passages follow rather shallow
paths at, or not far below, the water table. This char¬
acteristic is especially common in regions of moderate
relief.
As surface rivers entrench their valleys, groundwater
seeks progressively lower outlets, and new cave pas¬
sages form along those paths. Much of this entrench¬
ment is caused by uplift of the Earth's crust, while
stream erosion tries to keep pace. Therefore, younger
passages are not necessarily lower in absolute elevation
above sea level. Changes in the rate of river entrench¬
ment leave their mark on both valleys and caves. Steep-
walled valleys are produced by rapid entrenchment.
Wide flat-floored valley bottoms (or their remnant ter¬
races) indicate slow entrenchment, static conditions, or
even rising river levels, and major cave passages tend
to form at those elevations. Sediment accumulates in
valleys and caves when river levels rise.
FIGURE 3 A tubular passage in Sullivan Cave, Indiana, formed
by phreatic flow (i.e., at or below the water table). Although the
main stream has since diverted to a lower route, periodic overflow
still utilizes the passage shown here.
The erosional-depositional history of stream valleys
is controlled by significant events such as climate
changes, sea-level fluctuations, rearrangement of river
patterns, changes in sediment supply, changes in uplift
rate, or depression of the continent by the weight of
glaciers. At the surface, much of the evidence for this
history is destroyed by later erosion. A great deal
more evidence is preserved in caves.
Cave origin is not as thoroughly understood as the
above paragraphs might imply. There is still debate on
many topics, such as: When does cave development
actually begin? Of all the original openings in soluble
rock, which ones evolve into caves? How deep can
caves form below the surface? How does the local set¬
ting affect cave patterns? What do caves tell us about
past geologic history and climates? Each question is
best examined in karst regions of moderate relief,
because they not only show the greatest geologic and
hydrologic variety, but also are tectonically stable
enough that cave features can be preserved for lengthy
periods of time.
FIELD EXAMPLES
Several well-documented karst areas are briefly
described here, with emphasis on the details that help
to answer the questions posed above. These areas are
located in the low plateaus of the east-central United
States and in the Appalachian Mountains and have
three major karst-bearing sequences in common: lime¬
stones and dolomites of Cambrian-Ordovician, Silurian-
Devonian, and Mississippian age. The same strata,
arranged in different settings, produce a great variety
of cave types.
ENCYCLOPEDIA OF CAVES
736
s
W Crawford Upland Mitchell Plateau E
sulfate-rich
approx 25 km
FIGURE 4 Cross-section through the karst area of southern
Indiana. 1 = mainly insoluble sandstones and shales, 2 = Mississippian
limestones, 3 = insoluble siltstones. A = Caves in exposed limestones
in the Mitchell Plateau. B = Caves beneath insoluble caprock along the
eastern edge of the Crawford Upland. C = Deep flow routes guided
by interbeds of gypsum, with outlets through the caprock along faults.
Dissected Plateaus of Southern Indiana and
Western Kentucky
Perhaps no other karst area in the world shows the
relation between river erosion and cave development
as clearly as these low-relief plateaus, where carbonate
rocks are exposed at the surface around the perimeter
of the Illinois structural basin (Fig. 1). In general, the
rocks dip gently at less than half a degree toward the
center of the basin, which is located in southern
Illinois (Fig. 4). The main cavernous units are carbon¬
ate rocks of Mississippian age with a thickness of
100—200 m. Most of them exhibit prominent bedding.
At considerable depth below the surface, gypsum is
interbedded with limestone and dolomite in the mid¬
dle part of the sequence. At depths of less than about
100 m, dissolution by groundwater has removed most
of the gypsum.
Where the carbonate rocks are exposed at the sur¬
face they form a vast sinkhole plain covering thou¬
sands of square kilometers (the Pennyroyal Plateau of
Kentucky and Mitchell Plateau of Indiana). Many river
caves are located in this region, including some of the
longest caves in the United States. Binkley's Cave, Blue
Spring Cave, and the Lost River Cave System (all in
Indiana) and Hidden River Cave (Kentucky) all have
mapped lengths of 30 to 40 km. The erosional relief of
the sinkhole plain is generally less than 50 m, and so
the caves rarely contain more than one or two distinct
levels of passage development. In river valleys, post¬
glacial sediment has accumulated to thicknesses as
much as 25 m, flooding and choking some cave pas¬
sages that he below present base level.
Farther in the down-dip direction, the carbonate
rocks are capped by insoluble sandstone and shale,
which form a dissected upland of irregular limestone
ridges capped by the more resistant insoluble rocks (the
Chester Upland of Kentucky and Crawford Upland of
Indiana). Multilevel caves are numerous, because the
protective caprock inhibits erosional destruction of
the underlying caves. The most extensive caves are the
Mammoth Cave System (628 km of surveyed length),
the Fisher Ridge Cave System (182 km), and the
Whigpistle Cave System (56 km), all of which are close
neighbors at the southeastern edge of the Illinois Basin
in Kentucky. (A cave system is a group of several inter¬
connected caves explored separately and later joined by
the discovery of interconnecting passages.) Mammoth is
the world's longest known cave, more than twice as
long as any other, and it is likely that future exploration
will connect it with one or both of the neighboring
caves.
Still farther down-dip, the insoluble caprock
becomes continuous. Confined groundwater moving
through the confined underlying carbonate rocks is
able to escape upward through the caprock only along
sparse high-angle faults. Springs fed by this water
have high sulfate contents, showing that the inter¬
bedded gypsum is hosting deep groundwater flow
through the carbonates. Hydrogen sulfide in the spring
water indicates that sulfate reduction is also present at
depth.
By scanning these three regions from their down-
dip to up-dip ends, it is possible to envision how the
karst landscape must have evolved with time.
Presumably the deep gypsum dissolution in the con¬
fined part of the karst aquifer has prepared the way
for later cave development at shallower depths. In
places, the exposed carbonate rocks are fragmented
and contorted and contain veins of calcite and quartz,
all of which are indicators of former gypsum bodies.
However, extensive geologic mapping of caves in the
capped ridges and sinkhole plain shows that the cave
patterns are closely adjusted to present-day shallow
groundwater flow and must have had very little (if
any) inheritance from earlier deep-seated dissolution.
The clues that indicate passage development by
shallow groundwater flow are fairly clear. In the
prominently bedded rocks of this region, most vadose
water preferentially follows the dip of the strata,
except where it jogs downward along joint-controlled
shafts. Where it reaches the water table, the water loses
its dip tendency, and most of the water follows paths
that have no dip trend. Instead, many phreatic pas¬
sages closely parallel the strike of the beds, which is
approximately the line of intersection between the dip¬
ping beds and the less steeply sloping water table.
Most passages consist of a vadose upstream section
and a phreatic downstream section. At the present or
former vadose-phreatic transition point, the passages
change from down-dip canyons to low-gradient tubes
with irregular profiles. In Indiana the lower parts of
the carbonate sequence are massive and prominently
jointed, so that narrow, fissure-like passages are
formed in which the vadose-phreatic transition is not
so clear.
ENCYCLOPEDIA OF CAVES
SOLUTION CAVES IN REGIONS OF MODERATE RELIEF
SE NW
FIGURE 5 Idealized profile through the Mammoth Cave
System, Kentucky, and its relation to the surrounding landscape.
A = Cavernous limestones (from bottom to top: St. Louis Limestone,
Ste. Genevieve Limestone, and Girkin Formation). B = Resistant cap-
rock, mainly of sandstones and shales. C = Impure, poorly cavernous
limestone. 1 = large Tertiary-age canyons, with thick sediment fill,
180—210 m above sea level; 2, 3 = major cave levels with vadose-
phreatic transitions, 168 and 152 m above sea level; 4 = Late
Pleistocene passages; 5 = sediment fill from postglacial rise in base
level; 6 = springs fed by active cave streams; 7 = vadose passages
(canyons and shafts).
Mapping of these transition points in any part of
the region shows a tight clustering at a few consistent
altitudes. At Mammoth Cave these are 200—210 m,
180—190 m, 168 m, and 152 m (Fig. 5). These same
levels of development can be traced through much of
the region, although their altitudes and spacing vary.
In each of the lowest two levels in Mammoth Cave, the
vadose-phreatic transitions in different passages fall
within a meter of each other. If deep-seated dissolution
had prepared the way for later cave development, this
tight correlation would be absent. No good example of
control of cave development by ancient, deep-seated
dissolution has yet been demonstrated in either
Kentucky or Indiana, although it is likely that a few
may yet be revealed.
The drainage history of the region has been greatly
clarified by cave evidence. The uppermost passages
are large canyons partly or completely filled with sand
and gravel deposited by cave streams. These passages
represent slow dissection by surface rivers, alternating
with periods when river valleys and cave passages
were partly filled with sediment. This alternating
sequence may reflect climate changes from humid
(entrenchment) to relatively dry (sediment accumula¬
tion), but it may also have been the result of sea-level
changes. Throughout this period the Pennyroyal
Plateau and Mitchell Plateau formed a low-relief sur¬
face near base level with little or no karst develop¬
ment. The largest cave passages in the Chester and
Crawford Uplands are located at roughly the same ele¬
vations, and many of them appear to have formed by
737
water draining through the upland ridges and fed by
runoff from the adjacent plains.
A widespread episode of sediment accumulation
followed, covering the plains and filling the caves to
depths of up to 30 m. Soon afterward, rivers began to
entrench rapidly, allowing the Pennyroyal Plateau and
Mitchell Plateau to develop many sinkholes and caves.
This rapid entrenchment was apparently triggered by
rerouting of rivers by glacial advances, which led to
the development of the present Ohio River. The large
canyon passages and sediment fill in the upper levels
of the caves date from the Late Tertiary Period, which
predates the first significant continental glaciation.
Rapid entrenchment and development of the sinkhole
plain took place during the Quaternary Period, when
glacial ice masses periodically covered much of high-
latitude North America. Glaciers fell short of the pres¬
ent karst areas, but their effect on the caves was
strongly felt in the form of changes in river entrench¬
ment and valley filling.
Dating of the passages in Mammoth Cave has
recently been accomplished with cosmogenic alumi¬
num and beryllium isotopes in quartz-rich sediment
(Granger et ah, 2001). The oldest cave sediments in the
large upper-level canyons are up to 4 million years
old, and the canyons themselves are presumably much
older. The period of widespread sedimentation took
place about 2.6 million years ago, just before the onset
of extensive North American glaciation. The 168-m
and 152-m passage levels date from roughly 1.5 and
1.0 million years ago and correlate with adjustments of
the Ohio River drainage pattern. Some of these
changes were caused by diversion of surface rivers to
more southerly routes. The contribution of caves to the
interpretation and timing of the regional geomorphic
and climatic history is clearly demonstrated.
The Ozark Plateaus
Karst of the Ozark Plateaus of Missouri and Arkansas
(Fig. 1) resembles that of the sinkhole plains of Indiana
and Kentucky, except that most of its caves are in
dolomite. This rock type is widely considered to be
inferior to limestone in cave-forming potential, but the
caves seem not to have noticed. Ordovician and
Mississippian carbonates are the main cavernous units.
In general they dip gently away from the St. Francois
Mountains of southeastern Missouri at less than half a
degree, but with local broad, gentle warps in the
regional structure. Chert is so abundant in some of the
carbonate rocks that residual chert fragments have
accumulated to depths as great as 40 m, subduing the
surface karst topography. The chert fragments also
retard flood runoff into caves, and some surface rivers
ENCYCLOPEDIA OF CAVES
738
s
FIGURE 6 Map of Crevice Cave, Missouri, showing an ideal
branching passage pattern guided by prominent bedding. The rela¬
tion between dip-oriented canyons and the strike-oriented main
tubular passage is shown. E = entrance. From map by Paul Hauck.
fed by karst springs have relatively low flood peaks as
a result.
Caves are numerous and widespread in the Ozark
Plateaus. The longest are sprawling branchworks com¬
posed of sinuous, low-gradient passages guided by
bedding. Vertical shafts are sparse, and the low topo¬
graphic relief prevents them from being very deep.
Crevice Cave is the largest cave in the region, with
46 km of mapped passages (Fig. 6). The Ozarks are
also well known for large karst springs. Some of these
are fed by deep phreatic tubes that extend more than
60 m below base level.
In the 1930s through 1950s the well-known geologist
] Harlan Bretz (1956) used qualitative field observa¬
tions in this region to support his view that caves orig¬
inate deep in the phreatic zone and are later invaded
by vadose streams only late in their history. This
hypothesis has gone out of favor in the past half cen¬
tury. In recent years, however, certain aspects of this
model are being seen in more favorable light. The geo-
morphic history of Ozark caves has yet to be studied
in detail using quantitative methods.
Appalachian Plateaus
This region stretches all the way from central New
York to northern Alabama (Fig. 1). Its karst is devel¬
oped mainly in Mississippian carbonates, with maxi¬
mum thicknesses of about 500 m. At the northeastern
end the cavernous rocks are much thinner and of
Silurian-Devonian age. The Appalachian Plateaus con¬
tain the largest caves of the east, including the Friar's
Hole Cave System, Organ Cave System, and Scott
Hollow Cave (West Virginia); Blue Spring Cave and
Cumberland Caverns (Tennessee); and Sloan's Valley
Cave (Kentucky), all of which have surveyed lengths
between 40 and 70 km.
In many ways the geology and caves of this region
resemble those of the Mammoth Cave area of Kentucky,
but the Appalachian Plateaus have greater relief and
more obvious control by faults and insoluble beds. The
strata have been warped into broad folds, typically with
dips of only a few degrees. In many places the carbon¬
ate rocks are overlain by a thick impermeable cap and
are exposed only in narrow bands along escarpments
and entrenched river valleys. Perhaps the most dra¬
matic example is the Obey River Gorge of northeastern
Tennessee, where a broad, nearly flat plateau of sand¬
stone and shale has been breached by abrupt V-shaped
canyons up to 300 m deep, in which carbonate rocks are
exposed only in the lower 25—30%. Many complex
caves have developed along the valley bottoms parallel
to the rivers. These cave patterns reflect the widening of
joints and partings by stress release from erosional
unloading along valleys (Sasowsky and White, 1994).
Many caves in northern West Virginia have a similar
setting. A conspicuous exception in the latter area is
Simmons-Mingo Cave, which crosses beneath a promi¬
nent surface divide along a linear fault zone for a dis¬
tance of 3.2 km.
In parts of southeastern West Virginia, the insoluble
caprock is of irregular thickness and forms only the
tops of hilly plateaus. It has been breached by small
surface streams high above base level, and caves can
extend from these points of focused recharge to remote
springs along the plateau edges. Caves do not follow
the valley bottoms as faithfully as they do in the exam¬
ples cited above. The Friar's Hole Cave System is the
best example.
Along the plateau edges, many caves are perched
on underlying insoluble beds. In southeastern West
Virginia, hundreds of kilometers of vadose passages
have formed at the contact between limestones of
the Greenbrier Group and underlying shale of the
Maccrady Formation (Fig. 7). Erosion by vadose flow
has entrenched canyons into the shale as much as 15 m
deep. Flow along the contact is mainly in the dip direc¬
tion, but is locally deflected from that direction by
intersecting fractures. As a result, vadose passages are
not oriented so relentlessly down the dip as in the
Mammoth Cave area. The vadose passages feed con¬
duits of phreatic origin that follow the strike of the
beds, often along synclinal troughs. Dye tracing shows
that some of these conduits feed springs at distances
up to 20 km.
ENCYCLOPEDIA OF CAVES
SOLUTION CAVES IN REGIONS OF MODERATE RELIEF
739
FIGURE 7 Cave development at the contact between limestones of
the Greenbrier Group (ceiling) and shales of the Maccrady Formation
(walls and floor) in Ludington Cave, southeastern West Virginia.
Faults exert a strong effect on some caves in the
Appalachian Plateaus. The deepest vertical shaft in
any cave of the eastern United States, in Ellison's
Cave, northwestern Georgia, is located along a large
strike-slip fault. The fault zone features slickensides,
pulverized rock, displacement of solutional rock sur¬
faces, and occasional mild earthquakes. Upper-level
stream passages in the cave are perched on shaly beds
hundreds of meters above the local valley floors, but
where they intersect the fault the water plunges down
shafts up to 180 m deep.
Some caves in the region consist of extensive net¬
works of intersecting fissures, in which all major frac¬
tures within a large area of soluble rock have enlarged
more or less simultaneously by dissolution. Network
caves of this type can form in various ways, but, in the
areas described here, the largest are located where
seepage takes place through a thin overlying cap of
quartz sandstone. Water infiltrates thorough the sand¬
stone as if through a sponge, and enlarges each frac¬
ture at or near the top of the underlying carbonate
rock at fairly uniform rates. This process can take place
in either phreatic or vadose conditions, and often both
in sequence. The best example of a sandstone-capped
network in the Appalachian Plateaus is Anvil Cave,
Alabama (Fig. 8). Twenty kilometers of tightly labyrin¬
thine fissure passages are constrained to the top of
the Monteagle Limestone beneath a thin caprock of
Hartselle Sandstone no more than 10 m thick. This and
other similar caves have undergone some enlargement
by backflooding from nearby rivers.
Ridge and Valley Province, Appalachian
Mountains
The intensely folded and faulted eastern region of
the Appalachian Mountains (Ridge and Valley Province)
A i
Cross Section
Hartselle Sandstone
100 m
Alt. (m) 190 - B
160
Flint Creek
Monteagle Limestone
FIGURE 8 Map of Anvil Cave, in northern Alabama, a joint-
controlled network cave beneath a thin permeable sandstone cap.
From map by William Varnedoe.
consists of numerous subparallel ridges and valleys
that follow the trends of the folds (Fig. 1). Resistant
quartz sandstones and conglomerates are the main
ridge-formers. Carbonate rocks up to 2 km thick are
exposed in linear bands over much of the region,
mostly in valleys and along the flanks of ridges. Caves
in this region contain the same general passage types
as those in the plateaus farther west, but the overall
cave patterns show some differences. Because the aver¬
age dips are much steeper than in the plateaus, vadose
water tends to follow short, steep paths interrupted by
many shafts and fissures along discordant fractures.
Major phreatic passages usually have elongate patterns
extending along the strike of the beds.
A good example is the group of caves at Trout
Rock, in northeastern West Virginia (Figs. 9 and 10).
These include several levels of strike-oriented passages
in steeply dipping limestone, with a network maze at
the crest of a small anticline. Nearby Rexrode Cave
and Smoke Hole Caverns have linear patterns tightly
constrained along the strike of beds that are almost
perfectly vertical.
The Burnsville Cove System of Virginia illustrates a
variant of the tendency for strike orientation. The
main passage drains approximately along the axis of a
syncline, with many joint-controlled tributaries, over¬
flow routes, and local networks superimposed on the
main trend. Gravel and large boulders carried through
the cave during peak flow suggest that part of the
cave's complexity is due to enlargement by periodic
floodwaters.
In the western part of the Ridge and Valley
Province most of the caves are in Silurian-Devonian
limestones only about 100—200 m thick. The thinness
of these rocks helps to constrain many of the caves to
ENCYCLOPEDIA OF CAVES
740
s
strike-oriented trends. Along the eastern margin of the
region, prolonged weathering of the thicker Cambrian-
Ordovician carbonate rocks has formed a broad, roll¬
ing lowland that extends through western Virginia
and southeastern Pennsylvania. Caves in this area are
not as large as might be expected from the thickness of
soluble rock. The largest accessible caves tend to
occupy small residual knobs (e.g., Luray Caverns,
Virginia). Elongation of caves along the strike of the
beds is less prominent than in the thinner rocks to the
west, because the massive nature of the carbonates and
the intense structural deformation allow groundwater
to flow in a variety of somewhat unpredictable direc¬
tions along complex fractures.
Caves in the folded Appalachians have been used to
validate some of the most widely accepted views on
cave origin. The idea that levels of cave development
correlate with periods of rather static base level in
nearby river valleys had been proposed by various
geologists in the first half of the twentieth century.
Many of these studies were not entirely convincing
because some "levels" were caused simply by perching
of water on insoluble beds, and correlations between
caves in mountain ranges were clouded in places by
irregular rates of tectonic uplift. The elevations of
major cave passages in the Appalachians correlate well
with river terraces, which are remnants of former
floodplains produced when rivers pause in their
downward entrenchment and widen their valleys. The
caves are not perched on insoluble beds, because most
of them are oriented along the strike of steeply dipping
beds or cut across geologic structures. Most caves in
the region could be attributed to dissolution at the top
of the phreatic zone, because those in dipping beds
concentrate in narrow horizontal bands that terminate
abruptly in both the up-dip and down-dip directions.
Although these observations do not hold true for every
cave in the region, this general view is still valid for
the majority.
Caves in the Potomac River basin cluster vertically in
a few groups whose elevations decrease in the down¬
stream direction at roughly the same rate as the present
river gradients (White, 2009). This suggests that the
caves, while forming, were adjusted to river patterns
not much different from those of today. The various
cave levels become more distinct in the downstream
direction, apparently because base-level control is better
defined in the mature lower parts of the river valley.
Several excellent examples of sandstone-capped
mazes are located in the Ridge and Valley Province
where a thin cap of Oriskany Sandstone overlies
jointed limestones of the Devonian Helderberg Group,
and where the dip is locally rather low. Some of these
mazes have also been enlarged partly by periodic
flooding from nearby rivers, so the zone of maximum
FIGURE 9 Composite map of caves at Trout Rock, West Virginia,
showing their relation to the geologic structure. Hamilton Cave is a
network maze along the crest of a local anticlinal flexure. Trout Cave
and New Trout Cave are located along the anticlinal flank. Structural
contours are drawn at intervals of 10 m. The local dip of the strata is
shown in degrees. E = entrance. Cave maps by David West. Geology by
A. and M. Palmer.
passage development is not necessarily at the very top
of the limestone. In fact, flooding has produced local
mazes superimposed on the branching pattern of
many Appalachian caves, even where a sandstone cap
is absent. Mazes are especially common along the
crests of anticlines where fractures have been enlarged
by tension (Fig. 9).
Glaciated Plateaus of Northeastern and North-
Central United States
Extensive caps of glacial ice covered the north-
central and northeastern states several times during
the past 2 million years. Where glacial deposits are
thick (as in northern Illinois, Indiana, and Ohio) preex¬
isting karst surfaces were completely buried, and there
is no substantial field evidence for caves in these areas.
However, where the glacial deposits are thin or absent,
preglacial caves have survived with only minor
disturbance.
Howe Caverns and related caves in New York,
located in gently dipping Silurian-Devonian carbo¬
nates, are examples of preglacial caves that have sur¬
vived at least one phase of continental glaciation. Ice
sheets were more than a kilometer thick over the
region, and yet the caves remain more or less intact.
Some of the overlying surface karst was partly or
ENCYCLOPEDIA OF CAVES
SOLUTION CAVES IN REGIONS OF MODERATE RELIEF
741
completely masked by glacial sediments, and patterns
of groundwater flow were altered slightly by blockage
of inputs and springs by these deposits.
During the waning phases of the latest glaciation,
about 14,000 years ago, the local north-flowing river
(Schoharie Creek) was blocked by ice to form a pro¬
glacial lake up to 150 m deep. Thinly laminated clays
were deposited on the surface during this event, and
they are also present in every major cave in the
Schoharie Valley (Fig. 11) but absent in caves in
FIGURE 10 Strike-oriented passage in steeply dipping Devonian
limestone. New Trout Cave, West Virginia.
surrounding drainage basins. In Howe Caverns and
nearby caves the clays occupy the lowest passage
levels, and postglacial entrenchment has extended at
most only 10—20 cm below their base. Radiometric dat¬
ing of speleothems in these caves includes ages of more
than 350,000 years, showing that the caves had reached
their present state of maturity well before the onset of
the latest (Wisconsinan) glaciation about 70,000 years
ago. Only a few minor cave passages can be demon¬
strated to have formed in the 14,000 years since the
last glacial retreat. Some are diversions from upper
levels, and others are floodwater passages formed
under the most favorable conditions of aggressive
water, short flow path, and steep hydraulic gradient.
A few caves in the Schoharie Valley contain narrow
passages up to 0.5 m 2 in cross-sectional area that were
formed by water diverted around passages that are
blocked by glacial deposits.
Sinking streams are most common in karst of mod¬
erate relief, where runoff from large areas of insoluble
material feeds swallow holes in adjacent cavernous
rock. Like most surface streams, they tend to exhibit
severe flooding. Floods have their greatest effect on
caves that are at least partly air-filled. Cave enlarge¬
ment is rapid during floods because the incoming
water tends to have a low concentration of dissolved
solids. Solutional growth is aided by abrasion by
stream-borne solid particles. Caves that have been
FIGURE 11 Laminated glacial lake clays
in Howe Caverns, New York, were deposited
in proglacial Lake Schoharie roughly 14,000
years ago. The bedrock floor beneath the clay
lies about 1 meter below the present level of
the main cave stream (in foreground). Width
of photo is approximately 50 cm.
ENCYCLOPEDIA OF CAVES
742
s
influenced greatly by floodwater exhibit complex pat¬
terns, with diversion mazes, dead-end fissures, and
solution pockets superimposed on the basic pattern of
stream passages. Such caves occur in all the regions
described above, but they are fairly sparse, accounting
for an average of about 10% of all caves. The percent¬
age is much higher in glaciated regions because of
diversion of surface water into caves by low-perme¬
ability glacial deposits, which also block many springs.
An example is Mystery Cave, in southeastern
Minnesota, located in a low-relief plateau of highly
fractured Ordovician dolomites and limestones. It
developed as a subterranean cutoff of several mean¬
ders in a surface river, across a reach of particularly
steep gradient. The entire cave consists of a fissure net¬
work with 21 km of mapped passages. The local dip is
0.3—0.8° to the west, and the groundwater flow is to the
east, against the dip, along fractures discordant to the
bedding. The maze pattern is a result of periodically
aggressive water following steep hydraulic gradients
through a network of fairly wide initial fissures. The
stresses of glacial loading and unloading may have
aided the mechanical widening of joints, although
much of the cave predates the latest glaciation. Uranium-
thorium radiometric dates for calcite speleothems
cluster at 100,000—150,000 and 10,000—15,000 years,
which are respectively pre-Wisconsinan and post-
Wisconsinan. Flowstone in a high-level passage dates
to more than 350,000 years. The close relation of the
cave to the present river pattern is intriguing, because
the cave probably began to form at least half a million
years ago.
Isotopic analysis of calcite speleothems in this mid¬
continental region has provided paleoclimatic evidence
that is both more precise and geographically more spe¬
cific than that of other methods such as coring of
marine sediments. The fact that these speleothems can
be dated fairly accurately to ages of at least 350,000
years makes them doubly useful for paleoclimatic
studies. Variations in oxygen isotopes are temperature-
related, and although there are difficulties in translat¬
ing them directly into former temperatures, they are
useful in detecting paleoclimatic trends. Variations in
carbon isotopes help to determine the type of vegeta¬
tion in the area at any time.
SUMMARY
The examples described here represent the typical
range of caves found in karst regions of moderate
relief. Even with such a brief overview it is possible to
answer the questions posed earlier:
• When does cave development begin? This depends on
the definition of terms as much as on geologic
setting. Slow flow far below river levels can dissolve
primitive routes in highly soluble materials like
gypsum (as in Indiana), but there is evidence for
well-developed deep-seated conduit flow through
carbonates as well (as in Missouri). However, in
most karst areas of moderate relief there is little
evidence that deep flow has determined the patterns
of presently traversable caves. The close association
between caves and the present cycle of landscape
development argues against a widespread influence
by older deep-seated processes. This conclusion is
supported by the correlation of vadose-phreatic
transition points between cave passages and with
surface features such as river terraces.
• Which of the original fractures and pores in a soluble
rock evolve into caves? Those that are widest, have
the steepest hydraulic gradients, and shortest paths
are favored. Mystery Cave, Minnesota, illustrates
rapid development in such a setting, even though
the cave cuts diagonally across the strata in a
direction opposite that of the stratal dip. The limited
depth of phreatic loops in most caves, and the
tendency for phreatic tubes to follow the strike of
the strata, show that shallow paths are the most
favorable, owing to a decrease in fissure width with
depth. Concentrations of caves in zones of stress
release in deep canyons also show the importance of
initial fracture width, as well as the close association
with river levels.
• How far below the surface can caves form? The flow
systems that extend deepest below river levels
appear to follow major faults (even in relatively
undisturbed strata such as those of Indiana
and Kentucky). Although the depth of cave
development is limited in the moderate-relief
regions described here, phreatic caves at least
100 m below the present base level are known from
geochemical studies, water tracing, diving in
water-filled caves, and measurement of high-
amplitude phreatic loops in presently air-filled
caves.
• What is the effect of local geologic setting on cave
development? The nature of groundwater recharge
controls the overall cave pattern. Most caves fed
by typical karst recharge are branchworks ( e.g .,
Fig. 6). Thick overlying insoluble rocks restrict
recharge to a few points, which makes the cave
pattern deviate from the ideal branchwork
pattern. Perching of vadose water along partings
in well-bedded rocks, or on insoluble beds,
produces dominantly down-dip passages (Fig. 2).
Most phreatic passages are oriented roughly
parallel to the strike, except where the host rocks
are massive and prominently fractured. All areas
discussed above exhibit these trends, especially
ENCYCLOPEDIA OF CAVES
SPECIES INTERACTIONS
743
the Ridge and Valley Province. Network mazes
are formed either by uniform seepage through
overlying or underlying permeable but insoluble
rock or by severe discharge fluctuations during
floods. Worldwide, many large network caves are
formed by the rise of deep-seated water, with
dissolution performed by sulfuric acid or by
mixing of deep high-CCb water with shallow
I 0 W-CO 2 water. Although some caves in the
Appalachians have patterns and geologic settings
that suggest such an origin, there is so far little
supporting documentation.
• What information do caves provide about past climates
and geologic history? Their close association with
fluvial erosion relates them to the developmental
history of the entire surrounding region. Dating of
sediments and speleothems provides a reliable time
scale, and isotopic analysis of speleothems provides
information on former climates.
Although this section is limited to karst areas of
moderate relief, and describes only caves formed by
the carbonic acid reaction, the conclusions drawn from
it are valid for the majority of karst regions throughout
the world.
See Also the Following Article
Solution Caves in Regions of High Relief
Bibliography
Bretz, JH. (1956). Caves of Missouri (Vol. 39, Second Series). Rolla,
MO: State of Missouri, Division of Geological Survey and Water
Resources.
Davies, W. E. (1960). Origin of caves in folded limestone. National
Speleological Society Bulletin, 22(1), 5—18.
Granger, D. E., Fabel, D., & Palmer, A. N. (2001). Pliocene-
Pleistocene incision of the Green River, Kentucky, determined
from radioactive decay of 26A1 and lOBe in Mammoth Cave sedi¬
ments. Geological Society of America Bulletin, 113(7), 825—836.
Jones, W. K. (1997). Karst hydrology atlas of West Virginia. Charles
Town, WV: Karst Waters Institute.
Kastning, E. H., & Kastning, K. M. (Eds.) (1991). Appalachian karst.
In Proceedings of Appalachian Karst Symposium, Radford, VA.
Huntsville, AL: National Speleological Society.
Palmer, A. N. (2007). Cave geology. Dayton, OH: Cave Books.
Palmer, A. N., & Palmer, M. V. (Eds.), (2009). Caves and karst of the
USA. Huntsville, AL: National Speleological Society.
Sasowsky, I. D., & White, W. B. (1994). The role of stress release frac¬
turing in the development of cavernous porosity in carbonate
aquifers. Water Resources Research, 30(12), 3523—3530.
Stone, R. W. (1953). Caves of Pennsylvania. National Speleological
Society Bulletin, 15, 51—137.
White, W. B., & White, E. L. (Eds.), (1989). Karst hydrology —
Concepts from the Mammoth Cave Area. New York: Van Nostrand
Reinhold.
White, W. B. (2009). The evolution of Appalachian fluviokarst:
Competition between stream erosion, cave development, surface
denudation, and tectonic uplift. Journal of Cave and Karst Studies,
71(3), 159-167.
SPECIES INTERACTIONS
David C. Culver
American University
INTRODUCTION
Species can affect each other in a variety of ways. They
can be predator and prey, they can be competitors, or
they can be mutualists in which case each species has
a beneficial impact on the other. When two species
interact in an ecologically significant way, the interac¬
tion affects population size and population growth.
If species A decreases the number of individuals of
species B and species B decreases the number of indi¬
viduals of species A, then A and B are competitors. If
species A increases population size of species B but
species B decreases population size of species A, then
B is a predator of A. In this demographic sense, para¬
sites also fall into the category of predators. All inter¬
actions can be classified according to the effect that an
encounter of an individual of species A with species B
has (Fig. 1). Two-way interactions are quite familiar—
competition, mutualism, and predation. One-way inter¬
actions are less so—amensalism and commensalism.
Ecologically, amensalism is a very one-sided case of
competition, and commensalism is a very one-sided
case of mutualism. The interactions can occur between
two free-living species, or between individuals, one of
which is permanently attached to or embedded in
another—symbioses. Symbiotic (literally "living together")
organisms can have mutualistic, commensal, or para¬
sitic population interactions.
It is likely that many animals initially enter the sub¬
terranean realm in order to escape predators (including
parasites) and competitors. Visually oriented predators
cannot effectively find their prey in the darkness of a
cave. If a species has a competitor that is more visually
oriented rather than, for example, olfactory oriented,
the visually oriented competitor will be at a disadvan¬
tage in subterranean habitats. Parasites may also find it
difficult to make the transition to subterranean environ¬
ments perhaps because it is more difficult to locate a
host in the absence of light.
Caves have proved to be useful ecological laborato¬
ries for the study of interspecific interactions (Culver,
1982). The number of species in any one cave is quite
small compared to most surface communities. This
makes analysis much simpler because the number of
pairwise interactions is much smaller in smaller
Encyclopedia of Caves.
© 2012 Elsevier Inc. All rights reserved.
744
s
Demographic Impact of
Species B on Species A
Q. <D
t rr\
S ' 00
bi c/>
o|
fc 0)
<U Q
Q CO
0 - +
Amensalism
Commensalism
Amensalism
Competition
Predation
Commensalism
Predation
Mutualism
FIGURE 1 Classification of pairwise interspecific interactions.
When both species have a negative effect on each other this is com¬
petition; when both species have a positive effect on each other this
is mutualism; when one has a positive effect and the other a negative
effect this is predation. Less familiar are the highly asymmetric inter¬
actions of commensalism and mutualism.
communities. In a community of 4 species there are 6
pairs of possible interactions, whereas in a community
of 10 species there are 45 such pairs. (In general, for a
community of S species there are {S(S—1)/2} interac¬
tion pairs.) Many cave streams have only two or three
macroscopic invertebrates compared to dozens in
surface streams. This makes the job of determining
major competitors, predators, and prey much simpler.
Second, cave communities typically have a simple tro¬
phic structure with only two or three trophic levels.
Food webs are simpler and easier to analyze. Third,
because of the patchy nature of the habitat and the
resulting patchy distribution of most species, nearby
caves often have different species combinations pres¬
ent. This allows these "natural experiments" to be
used to study the effects of species additions and
removals. Because of the relatively large number of
caves in cave-bearing areas (there are more than 45,000
known caves in the United States), the number of such
natural experiments can be quite large.
In subterranean environments, predation and compe¬
tition are relatively well known. The other two-way
interaction—mutualism—has been little studied in sub¬
terranean environments. Hobbs (1975) studied ostracods
living on the exoskeletons of cave crayfish. Ostracods
gain an advantage from the interaction because they
feed on microorganisms and detritus that accumulate on
the host exoskeleton. Crayfish may gain some advantage
from the ostracods directly from exoskeleton cleaning.
Hobbs found that the interaction seemed to decrease in
intensity with increasing cave adaptation. In Pless Cave,
Indiana, U.S.A., stygobiotic Orconectes inermis inermis had
an average of 13.5 ostracods per individual. Cambarus
tenebrosus, a stygophile also common in surface streams.
had an average of 31 ostracods in Pless Cave, more than
twice as many. Strictly surface-dwelling species, such as
Cambarus bartoni, have even more ostracod ectocommen-
sals—sometimes more than 100.
Relative to free-living predators, there are few stud¬
ies of parasitic interactions in subterranean environ¬
ments. There are a few parasites that have specialized
on subterranean species, in some ways the ne plus ultra
of extreme specialization. One of the most spectacu¬
lar examples are Temnocephala, parasitic flatworms
which parasitize cave shrimp living in Balkan caves,
probably feeding on the haemolymph of the shrimp that
they access through thin parts of the exoskeleton
(Matjasic, 1994). Seven species and several genera of
Temnocephala are found only on the cave shrimp
Troglocaris schmidti, with each species specializing in a
particular region of the body. For example, Subtelsonia
perianalis is only found around the anus of T. schmidti.
Although many interactions have been remarked
on in the literature, and complete food webs have been
constructed for a few communities, two sets of species
have been studied extensively. One is the predator-
prey system of carabid beetles and their cricket egg
prey, which occurs extensively in caves throughout
the Interior Low Plateaus of Kentucky, Tennessee,
and Alabama and in the Balcones Escarpment of Texas.
The other is the intricate systems of interactions
between amphipods and isopods in cave streams in the
Appalachian Mountains of Virginia and West Virginia.
BEETLE PREDATORS AND THEIR
CRICKET EGG PREY
Because of the scarcity of food in caves, especially
for predators, the great Romanian biospeleologist Emil
Racovitza pointed out that many cave animals are
"carnivores by predilection but saprovores by neces¬
sity." This has the effect of both shortening food webs
and making predatory behavior difficult to observe
because predatory events themselves are usually quite
rare. An exception to both the rarity of observation
and tendency of predators toward omnivory is the
predator—prey pair of cave crickets (in particular,
their eggs) and beetles found in many caves in North
America.
The best studied pair is that of the beetle Neaphaenops
tellkampfi and the cave cricket Hadenoecus subterraneus in
Mammoth Cave, Kentucky, and other nearby caves
(Kane and Poulson, 1976). Very similar predator—prey
interactions occur with other beetles (e.g., Rhadine subter-
ranea ) and cave crickets in Texas (e.g., Ceuthophilus cuni-
cularis) and with other beetles and crickets in other
parts of Kentucky.
ENCYCLOPEDIA OF CAVES
SPECIES INTERACTIONS
745
FIGURE 2 Concentration of cave crickets, Ceuthophilus stygius, on
the ceiling of Dogwood Cave, Hart County, Kentucky, U.S.A. Photo
by H. H. Hobbs. Used with permission.
The key to this predator—prey interaction begins
with the cave crickets (Lavoie et al., 2007). Species
such as Hadenoecus subterraneus are often quite com¬
mon in caves, typically reaching numbers into the hun¬
dreds and even thousands (Fig. 2). Morphologically,
they show signs of both subterranean and surface
life—they have eyes (sometimes reduced) but appen¬
dages are elongated relative to surface species. Many
individuals, especially those near the entrance, leave
the cave at night to forage for food during the warmer
times of year. Some species may leave every night dur¬
ing the summer months but H. subterraneus probably
leaves the cave to feed at 10- to 20-day intervals.
During the day in the summer, and during day and
night in winter months, they stay in caves, where they
typically lay their eggs. Presumably the advantage
they gain from cave life is avoidance of most predators
and some relief from high daytime temperatures in the
summer. Crickets and other relatively large-bodied
orthopterans have many enemies—many birds during
the day and small rodents at night.
Whatever the frequency of outside feeding, it repre¬
sents an important food source entering the cave. Many
species feed on the splattering of cricket guano present
in the caves and a few beetles have specialized on the
eggs crickets bury in the cave. In caves with sand or
uncompacted silt, crickets bury their eggs in this soft
substrate. Cricket eggs, rich in protein and lipids, and
nearly one-third the size of an adult N. tellkampfi
(Fig. 3), are a bonanza to the beetles if they can find
them. Successful cricket egg predation behavior has
evolved several times among cave beetles. Most of
these beetle species are 6.5 to 8 mm long and actively
forage for eggs, locating the holes where the eggs were
laid by a combination of chemoreception and
FIGURE 3 Sketch of Rhadine subterranea eating a cricket egg.
Adapted from a photograph by Robert W. Mitchell.
mechanoreception. Beetles dig a hole to find an egg,
and when they are successful, they remove the egg,
pierce it with their mandibles, and the egg contents are
pumped into the gut. In caves without soft substrates,
beetles are unable to locate and dig up the eggs. The
size of a cave cricket egg relative to that of N. tellkampfi
(other cricket egg predators are of similar size) indicates
the potential energetic importance of cave cricket eggs
(Fig. 3). The dry weight of an H. subterraneus egg
(2.26 g ± 0.03) is nearly ’/i that of the dry weight of an
adult N. tellkampfi (3.02 g ± 0.07). A single egg is suffi¬
cient food for a beetle for weeks and it takes approxi¬
mately 50 days for the beetle to return to its prefeeding
weight.
The intensity of egg predation is high. In sandy areas
in Mammoth Cave beetles consume over 90% of the
cricket eggs laid throughout the year, and more than
half are found and eaten by beetles within 15 days of
being deposited. Beetles dig in the substrate to find
eggs, and when they are successful, they remove the
egg, pierce it with their mandible and the egg contents
are pumped into the gut. The impact of this resource on
beetle population dynamics is striking, and reproduc¬
tion in beetles closely follows the time of maximum rate
of egg depositions by crickets (Fig. 4). The quantitative
impact of egg predation on H. subterraneus populations
is unknown but it must be considerable.
An evolved response on the part of H. subterraneus
has been to increase the ovipositor length in popula¬
tions experiencing egg predation by cave beetles. The
increase in ovipositor length from 12 to 13 mm is suffi¬
cient to reduce egg predation. Beetles often do not
dig deep enough. Most if not all populations of cricket-
egg-eating beetles must switch to other prey because
cricket eggs are not available year-round. In periods of
cricket egg scarcity, N. tellkampfi and other species
become "ordinary" predators, feeding on a variety of
prey. They also must switch habitats since the sandy
ENCYCLOPEDIA OF CAVES
746
s
Month
—♦— Hadenoecus Eggs
—*— Neaphaenops Tenerals
FIGURE 4 Seasonal changes in number of eggs per square meter
of the cave cricket Hadenoecus subterraneus and a visual census of
newly emerging adults (tenerals) of the beetle Neaphaenops tellkampfi
in Edwards Avenue, Great Onyx Cave, Kentucky. Data from Kane and
Poulson, 1976.
substrates and uncompacted silt where cricket eggs are
laid have few other species. Interestingly, they are appar¬
ently not very efficient general predators, much less so
than related species in the genus Pseudanophthalmus as
well as conspecific populations that do not prey on
cricket eggs.
COMPETITION AND OTHER
INTERACTIONS IN APPALACHIAN
CAVE STREAMS
0.5
0.4
0.3
0.2
£ o.i
I
i
S. spinatus S. spinatus with C. holsingeri C. holsingeri
alone C. holsingeri alone with S. spinatus
0.6
~ 0.5
o
(/>
I
4 -
o
>
o
C
CD
3
U"
03
0.4
0.3
0.2
u- 0.1 -
I
\
G. minus alone G. minus with C. holsingeri C. holsingeri
C. holsingeri alone with G. minus
FIGURE 5 Mean frequency, together with standard errors, of
washout rates from riffles of Gammarus minus, Caecidotea holsinger,
and Stygobromus spinatus in various combinations in laboratory
stream experiments.
A thoroughly studied case of interaction within a tro¬
phic level is that of the amphipod and isopod species
that occupy many gravel-bottomed streams in
Appalachian caves (Culver and Pipan, 2009). In these
streams, there is an alternation between deeps (pools)
and shallows (riffles). The amphipods and isopods are
highly concentrated in riffles as a result of the concentra¬
tion of food (especially leaf detritus), increased oxygen,
and the absence of salamander predators, which live in
pools. In this habitat, the three obvious kinds of interac¬
tions are as follows: (1) Species may compete for food,
(2) species may compete for space (the underside of
gravels), and (3) species may serve as food for other spe¬
cies. All can and do occur in particular situations, but
the most universal (and easiest to analyze) is competi¬
tion for space on the underside of riffles (Culver, 1994).
There is a general risk associated with life on the
underside of rocks—the risk of washing out into
the current. Individuals that wash out into pools run
the risk of being eaten by salamanders, the primary
predators living in pools. Individuals also run the risk
of damage from buffeting by the current. As with
many subterranean species, the amphipods and isopods
have long, thin appendages that are easily broken.
Amphipods and isopods with broken appendages are
not easy prey for salamanders but are also attacked and
eaten by other amphipods and isopods. Amphipods
may be at more risk than isopods in current. The cave
stream amphipod species are not good swimmers and
their laterally compressed body shape, compared to the
dorso-ventrally flattened body shape of isopods, is not
hydrodynamically efficient in moving water. It is very
easy to observe the behavioral response to most encoun¬
ters even in a small dish in the lab—one or both indi¬
viduals rapidly move away. More realistic laboratory
experiments can be done in a small artificial riffle,
where the washout rate of individuals put in the riffle
in various combinations can be measured.
An example is shown in Figure 5, involving two
pairs of species from Organ Cave. When the amphipod
Stygobromus spinatus is in an artificial riffle with the
isopod Caecidotea holsingeri, S. spinatus has nearly dou¬
ble the washout rate that it has when it is alone, indi¬
cating that C. holsingeri is its competitor (Fig. 5A).
However, the converse is not true. S. spinatus seems to
ENCYCLOPEDIA OF CAVES
SPECIES INTERACTIONS
747
have no effect on C. holsingeri. Thus the overall interac¬
tion is one of amensalism (Fig. 1), a very one-sided
competition in this case. This asymmetry was not the
result of size differences because these two species are
roughly the same size, about 5 mm. The isopod is the
superior competitor perhaps of its more hydrodynamic
body shape.
Another pair of species, the amphipod Gammarus
minus and the isopod C. holsingeri, displayed competi¬
tive behavior, and both species had a higher washout
rate in the presence of the other (Fig. 5B). This would
seem to be a simple interaction, but it proved to be
more complex than first thought. The complication was
that some C. holsingeri "disappeared" during the day¬
long washout experiments and where they disappeared
to was the gut of G. minus'. The interaction between
these two species had elements of predation (hence the
"disappearing" isopods) and of competition (hence the
mutual avoidance by both species when in a riffle).
The next step was to construct a model of competi¬
tion relating relative washout rate to the interaction
coefficients of the standard competition equations:
dNi/dt — r^NdlC N-] QT2N2)
dN2/dt — r 2 N 2 (X 2 — N2 — 0:21 Ni)
where r is the intrinsic rate of increase of species 1 and
2, N is population size, K is carrying capacity, and ay
represents the effect of species j on species i. For com¬
petition for space in a riffle a y is approximately equal
to the ratio of the washout rate of species i when spe¬
cies j is present to the washout rate of species i when
species j is not present. The resulting estimates of com¬
petition can then be compared with field data on the
amount of overlap among species in caves and within
riffles within a cave.
Three different cave stream communities show the
range of interactions. The first involves a trio of species
found in Thompson Cedar Cave and other nearby caves
in Lee County, Virginia: the amphipod Crangonyx anten-
natus and the isopods Caecidotea recurvata and Lirceus
usdagalun. The intensities of competition measured in
the laboratory were such that, if they accurately reflect
the situation in the field, only one of the three pairs
(Crangonyx antennatus and Caecidotea recurvata) should be
able to persist in the same stream, as well as the triad of
species itself. In a survey of seven cave streams, no
"unstable" species pairs were found, suggesting that lab¬
oratory measurements reflected the situation in the field.
This was further supported by the distribution within
Thompson Cedar Cave (Fig. 6). In various parts of this
small cave stream, all three species occurred together,
Lirceus usdagalun occurred by itself, and Crangonyx anten¬
natus and Caecidotea recurvata occurred together, but in
N
/ 0-30 m
NNSNWWl
FIGURE 6 Map of distribution of Crangonyx antennatus, Caecidotea
recurvata, Gyrinophilus porphyriticus, and Lirceus usdagalun in
Thompson Cedar Cave, Lee County, Virginia.
no place did the pair Lirceus usdagalun and Caecidotea
recurvata or the pair Lirceus usdagalun and Crangonyx
antennatus occur. The final wrinkle is the occurrence of
predaceous larvae of the spring salamander Gyrinophilus
porphyriticus. Its preferred prey, Caecidotea recurvata, is
reduced in abundance, and very few amphipods and
isopods occur in pools. When Gyrinophilus porphyriticus
finds isolated rimstone pools, it decimates any amphi¬
pod or isopod population that might be there.
The third example of species interactions that has
been intensively studied is the isopod community in
Alpena Cave, West Virginia, U.S.A., and one that
demonstrates that competition is not universal in
cave stream communities. Two species occur in the
same stream—the isopods Caecidotea cannulus and
Caecidotea holsingeri. Superficially they would seem to be
competitors. Neither laboratory stream studies nor field
perturbation experiments detected any evidence of com¬
petition between these two species. They both occur in
the same riffles throughout the cave stream. It is possi¬
ble that competition between the two species did exist
in the past. Typically, Caecidotea cannulus is larger than
Caecidotea holsingeri and this difference is enhanced in
Alpena Cave when the two species occur together, thus
reducing competition. Flowever, the size of the isopods
is strongly correlated with the size of the rocks in the
streams and it turns out that Alpena Cave has a
bimodal distribution of gravel sizes. So, it can only be
said that at present competition is not occurring.
COMPETITION AS A RESULT OF
EUTROPHICATION
Species interactions also play a key role in the
replacement of a specialized cave fauna in a cave
ENCYCLOPEDIA OF CAVES
748
s
stream subjected to eutrophication (Sket, 1977). In the
late 1950s, the Pivka River in Slovenia carried a heavy
load of organic pollutants into Postojnska Jama. Where
the Pivka River entered the cave, oxygen concentra¬
tions (a measure of eutrophication) in the river were
10% of saturation. One kilometer into the cave, oxygen
levels recovered to nearly 50% of saturation. Six kilo¬
meters into the cave oxygen concentration was nearly
at saturation and other measures of eutrophication
indicated it had largely disappeared. The very rich
obligate cave-stream fauna, including the very interest¬
ing and intensively studied isopod Asellus aquaticus
cavernicohis, was largely extirpated in the first several
kilometers of stream passage. The fauna was replaced
by surface-dwelling aquatic insects, surface-dwelling
amphipods in the genus Gammarus, and the surface¬
dwelling isopod species Asellus aquaticus aquaticus.
These species were able to invade because of the high
energy and nutrient levels in the eutrophic Pivka
River. As a result of their invasion, the stygobionts
(such as Asellus aquaticus cavernicolus) were pushed far¬
ther into the cave, not because they couldn't survive
under the higher food conditions, but because they are
out-competed. In extreme cases, such as probably
occurred near the entrance to Postojnska Jama, low
oxygen levels can prevent the survival of stygobionts,
but most areas did not have extremely low oxygen
values. Especially interesting are those areas of the
underground Pivka River that came to be dominated
by aquatic insects and Gammarus, both of which
are not species of highly polluted waters, but rather
stygophiles, species that can live in both surface
and subterranean habitats. Stygophiles were probably
better competitors of surface-dwelling species than
were stygobionts in polluted cave streams. Of course,
stygophiles can survive and reproduce in nonpolluted
surface streams.
It also seems likely that competition with surface¬
dwelling species and predation by surface-dwelling
species is a major factor in preventing the movement
of subterranean species onto the surface. Sket showed
that in the absence of competitors and predators, sty¬
gobionts may forage in surface environments. Of
course, other factors prevent movement of subterra¬
nean species to the surface. Many subterranean species
are sensitive to light and also unable to cope with envi¬
ronmental fluctuations.
CONCLUSION
Because of the highly replicated nature of caves,
they can be important ecological laboratories for the
study of species interactions. The above studies
indicate some of the potential for such work. Many
more cave communities await such an analysis.
See Also the Following Article
Natural Selection
Bibliography
Culver, D. C. (1982). Cave life: Evolution and ecology. Cambridge, MA:
Harvard University Press.
Culver, D. C. (1994). Species interactions. In G. Gibert, D. L. Danielopol
& J. A. Stanford (Eds.), Groundwater ecology/ (pp. 271—281). San
Diego: Academic Press.
Culver, D. C., & Pipan, T. (2009). The biology of caves and other subter¬
ranean habitats. Oxford: Oxford University Press.
Hobbs, H. H., Ill (1975). Distribution of Indiana cavernicolous cray¬
fishes and eco-commensal ostracods. International Journal of
Speleology, 7, 273—302.
Kane, T. C., & Poulson, T. L. (1976). Foraging by cave beetles: Spatial
and temporal heterogeneity of prey. Ecology, 57, 793—800.
Lavoie, K. H., Helf, K. L., & Poulson, T. L. (2007). The biology and
ecology of North American cave crickets. Journal of Cave and Karst
Studies, 69, 114-134.
Matjasic, J. (1994). Turbellaria, Temnocephala. In C. Juberthie &
V. Decu (Eds.), Encyclopaedia biospeologica (Tome I) (pp. 45—48).
Moulis, France: Societe Internationale de Biospeologie.
Sket, B. (1977). Gegenseitige Beeinflussung der Wasserpollution und
des Hohlenmilieus. Proceedings of the 6th International Congress of
Speleology, Olomouc, CSSR, 4, 253—262.
SPELEOGENESIS, HYPOGENIC
Alexander Klimchouk
Ukrainian Institute of Speleology and Karstology
INTRODUCTION
Historically, scientific notions of karst evolved from
regions where it is most noticeable, where soluble
rocks are exposed to the surface, and specific geomor-
phic and hydrologic features such as dolines, sinking
streams, and caves are well represented. Karst and
cave development in such regions is driven by
recharge from the surface, either diffused authogenic
or concentrated allogeneic. Dissolution of carbonate
rocks occurs mainly due to carbonic acid generated by
reactions with soil CO 2 .
Karst developed in these settings is termed epigenic.
The traditional paradigm of karst science originally
evolved from studying epigenic karsts. The adoption of
the very scientific term from the Germanized name of
the geographic area in Slovenia further strengthened the
epigenic bias in karst researches, dominating till now.
Traditional concepts and models for the origin of caves
Encyclopedia of Caves.
© 2012 Elsevier Inc. All rights reserved.
SPELEOGENESIS, HYPOGENIC
749
have been mainly developed based on the geomorphic
and hydrogeologic context of epigenic karst.
Epigenic speleogenesis, directly related to the infil¬
tration of meteoric water, is predominantly associated
with local flow systems and/or recharge regimes
of intermediate to regional meteoric flow systems.
Epigenic caves form along paths (connected joints
and/or pores) where the greatest flow occurs between
recharge points at the surface and discharge points in
topographic lows. The recharge regime is characterized
by highly variable input parameters, relatively high
hydraulic heads a decreasing with depth, downward
and divergent flow, chemically aggressive ground-
waters with low dissolved solids content, oxidizing
conditions, and negative anomalies of geothermal heat
and gradient (Fig. 1). The types of recharge, the rela¬
tionships of recharge and discharge points, and the
initial percolation patterns are controlled by the geo¬
morphic settings and distribution of primary porosity
and fractures. The amount of recharge is controlled by
climate. The positive feedback loop between discharge
and the rate of conduit growth is fully operational in
epigenic speleogenesis, which leads to the competitive
development of conduits, progressive concentration of
flow, and domination of branching cave patterns.
Epigenic conduit systems favor lateral flow and
respective porosity diminishes with depth.
Abundant instances of karst cavities encountered by
boreholes and mines at greater depths have been
commonly interpreted as paleo(epigenetic) karst, sim¬
ply because of the overwhelming dominance of the
epigenic concepts of karst and speleogenesis.
However, there are many examples of deep-seated
cavernous porosity in strata, in which epigenic karst
development can definitely be ruled out, for example,
in rocks that were buried continuously after deposition
and have never been exposed. The advancements in
karst and cave science during past 20—30 years led to
the growing recognition of the possibility and wide
occurrence of cave development in deep-seated condi¬
tions, without direct recharge from the surface, by
recharge to the cave-forming zone coming from depth.
This type of speleogenesis is termed hypogenic (or
hypo gene).
The idea that some caves could form at depth by
rising waters was introduced as early as the nineteenth
century with regard to hydrothermal systems. More
significant attention has been given to hydrothermal
speleogenesis since the mid-twentieth century, particu¬
larly in Eastern Europe. Later on, another chemical
mechanism of cave development by rising waters was
recognized, the oxidation of H 2 S and dissolution by
sulfuric acid. By the end of the twentieth century the
term and concept of hypogene speleogenesis had been
firmly established, although remained largely limited
to caves formed by the above dissolution mechanisms
(Ford and Williams, 1989; Palmer, 1991; Dublyansky,
2000 ).
a
a
ill a
b
b
.... b
^ g MT.-4T 9 Eh* Eh' 10 ES 11 HS 12
FIGURE 1 Karst and speleogenesis in the context of diagenetic zones and groundwater flow regimes. The diagram is out of scale and the ver¬
tical dimension is greatly exaggerated. 1 = meteoric, topography-driven regime: a = local systems (unconfined), b = regional and subregional sys¬
tems (confined); 2 = expulsion (exfiltration, basinal) regime, commonly overpressured, driven by compaction and tectonic compression: a = in
newly deposited sediments, b = in older rocks; 3 = interfaces between groundwater regimes and systems: a = meteoric/expulsion regimes,
b = local/regional-subregional meteoric systems; 4 = poorly permeable beds (only a few are shown on the diagram); 5 = meteoric flow paths;
6 = basinal flow paths; 7 = enhanced cross-formational communication; 8 = intense gas inputs; 9 = temperature and gradient anomaly: positive,
negative; 10 = redox conditions: oxidizing, reducing; 11 = epigenic speleogenesis; 12 = hypogenic speleogenesis.
ENCYCLOPEDIA OF CAVES
750
s
The recent past two decades have witnessed a rapid
growth in the number of in-depth studies of different
kinds of hypogene speleogenesis in various places
around the world. It was recognized that cave devel¬
opment by various processes in deep-seated environ¬
ments is much more common than previously thought
(Palmer, 1995; Klimchouk, 2000, 2007; Klimchouk and
Ford, 2009; Stafford et ah, 2009).
At the same time, sedimentologists and industry
geologists concerned with the origin of porosity began
to realize limitations of the heavily used model of sub¬
aerial meteoric diagenesis in freshwater, which implied
that deep-seated porosity in carbonates is related to
unconformities and is mainly due to dissolution in
paleovadose and paleophreatic freshwater zones (i.e., is
paleokarst). They came to realize that deep-burial dia¬
genesis in the mesogenetic environment can contribute
significantly to secondary dissolution porosity and per¬
meability evolution in many carbonate hydrocarbon
reservoirs and ore deposits (e.g., Mazzullo and Harris,
1992; Machel, 1999).
The ongoing burst of interest in hypogene speleo¬
genesis has been largely related to the establishment of
a hydrogeological rather than geochemical approach to
its definition. This approach highlighted the common
hydrogeological genetic background and explained the
multifaceted similarity of caves formed by upwelling
flow, previously seen as unrelated because of their
attribution to different chemical processes involved.
The recent reviews, special symposia, and paper col¬
lections (Klimchouk, 2007; Klimchouk and Ford, 2009;
Stafford et ah, 2009) have demonstrated that hypogene
speleogenesis is not just an aberrant curious phenome¬
non within the otherwise predominantly epigenic karst
paradigm but is one of the fundamental categories of
karst, at least of equal importance with epigene karst.
BASIC CONCEPT AND DEFINITION
Speleogenesis is a result of interaction between
groundwater and its environment, determined by the
various components and attributes of the two systems
seeking equilibration. Geochemical mechanisms that
cause the dissolution effect on the host rocks are many
and varied. To cause speleogenetic development, the dis¬
solution effect of disequilibria has to accumulate over
sufficiently long periods of time and/or to concentrate
within relatively small rock volumes or areas. The sys¬
tematic transport and distribution mechanism that cre¬
ates and supports the required disequilibrium conditions
and hence controls the distribution of the dissolution
effect is the groundwater flow, which suggests that the
latter has to be a primary consideration.
Sedimentology considers (among other alteration
processes) any dissolution in the rock during diagene¬
sis, most of which occurs in a dispersed mode in the
pore media. Speleogenesis deals with the creation
and evolution of organized permeability structures on
the macro-scale (>ca. 5—15 mm) due to dissolutional
enlargement of an earlier porosity.
The concept of hypogene speleogenesis does not nec¬
essarily mean cave development at great depth but,
rather, refers to the origin of the cave-forming agency
from depth. Hypogene speleogenesis is defined as the
formation of solution-enlarged permeability structures by
water that recharges the cavernous zone from belozv, indepen¬
dent of recharge from the overlying or immediately adjacent
surface (Ford, 2006; Klimchouk, 2007). Hypogene caves
form by a variety of geochemical mechanisms including
those that do not rely on acids (such as dissolution of
evaporites). At the same time, hypogene caves formed
in different lithologies and by different dissolution pro¬
cesses demonstrate remarkable similarity in their pat¬
terns, morphologies, hydrostratigraphic occurrence, and
current or inferred hydrogeologic functioning, which
suggests their common hydrogeologic backgrounds.
Hypogene speleogenesis is associated with upwell¬
ing flow characteristic of discharge regimes of deeper
(subregional and regional) meteoric flow systems,
either terminal or intervening, or those of basinal flow
systems (Fig. 1). It also tends to be related with inter¬
faces between different flow systems. An upwelling
flow inherently implies a certain degree of hydrogeo¬
logical confinement, which is almost always charac¬
teristic of deep flow systems because of the almost
ubiquitous layered heterogeneity in sedimentary
sequences. When hypogenic solution porosity struc¬
tures are shifted to the shallower, unconfined situa¬
tion due to uplift and denudation but their further
development continues to depend on upwelling flow
from deeper systems, this is still hypogenic develop¬
ment although now unconfined. Unconfined hypo¬
gene development can be regarded as an extinction
phase of hypogene speleogenesis.
DIAGENETIC ENVIRONMENTS,
HYDROGEOLOGICAL CYCLES, AND
HYDRODYNAMIC ZONES OF
THE EARTH’S CRUST
Studies of hypogene speleogenesis have emphasized
the need to approach the development of karst porosity
and permeability from a perspective that goes far
beyond contemporary geomorphological epochs. To set
a framework for characterization and understanding of
hypogene speleogenesis, one has to view karst from the
perspective of the whole geologic evolution of soluble
ENCYCLOPEDIA OF CAVES
SPELEOGENESIS, HYPOGENIC
751
deposits, particularly of the evolution of diagenetic and
hydrogeologic environments.
Diagenetic Environments
The realm of diagenesis comprises all alterations of
sedimentary deposits that occur since ultimate deposi¬
tion till the onset of metamorphism. Being the result of
dissolution, speleogenesis can be viewed as a specific
kind of diagenetic changes. However, in discussing
aspects of diagenesis in this account we are concerned
mainly with localized effects of dissolution processes
rather than with dissolution that may occur diffusely
throughout the micropore media of sediments.
Various classifications have been proposed to distin¬
guish diagenetic environments and respective stages
and zones. Interested readers should consult the abun¬
dant existing literature. The scheme of Choquette and
Pray (1970) is one of the most widely used, which
divides diagenetic history into three stages:
1. E ogenesis, when the shallow marine sediment is first
laid down and may be altered by the effects of
periodic drying or exposure to rainwater. The
eogenetic zone grades downward into the
mesogenetic zone where the newly deposited
sediment is buried to depth where near-surface
processes become ineffective.
2. Mesogenesis denotes alteration of sediments in the
conditions of increasing burial depth (and hence
increasing temperature and pressures) by the
processes unrelated to the surface. Probably the
most of hypogene speleogenesis occurs in the
mesogenetic domain.
3. Telogenesis replaces mesogenesis when buried
deposits, being shifted to a shallower position due
to uplift and denudation, are affected again by
processes related to the exposed surface, particularly
to invading meteoric groundwater. This notion
implies that the rock has been subjected to burial
diagenesis to some degree. The lower limit of
the telogenetic realm is not clearly defined.
Some researchers draw this limit at or near water
table (thus, telogenesis would roughly coincide with
the realm of epigenic karst in this connotation)
whereas others put it to the lower boundary of
penetration of meteoric groundwaters. With the
latter meaning, hypogene speleogenesis in confined
deep but still meteoric flow systems falls in the
domain of telogenesis.
Machel (1999) classifies the diagenetic realm into
near-surface, shallow-, intermediate-, and deep-burial
diagenetic settings, hydrocarbon-contaminated plumes,
and fractures. Near-surface settings and the upper part
of the shallow burial settings in this scheme roughly
correspond to the eogenetic zone of Choquette and
Pray (1970) whereas intermediate and deep burial set¬
tings, with a boundary between them at depths vary¬
ing from 2000 to 3000 m (the top of liquid oil window),
represent the mesogenetic realm. Deep-burial diage¬
netic settings merge down into the metamorphic realm
at temperatures around 200°C (at depth of about
6—9 km). Hydrocarbon-contaminated plumes and frac¬
tures, which are particularly important for hypogene
speleogenesis, may crosscut other settings.
Most sedimentary rocks have been buried to various
depths (typically several hundreds or thousands of
meters) for geologically long periods. Some of them
experienced subsequent uplift, exposure, and new
burial (with possible repetitions of the cycle) while
others have never been exposed. The occurrences of
large cavities at great depths in the rocks that have
never experienced exposure are common and consti¬
tute a strong evidence for hypogene speleogenesis in
deep-seated settings.
Aqueous fluids in the mesogenetic environment can
be of different origins and they mix in varying propor¬
tions to constitute formation waters. Connate waters
dominate in pores in shallow and intermediate settings
of subsiding basins. With increasing burial depth they
are expelled from clayey sediments into coarsely
grained reservoirs. During the continental exposure,
meteoric waters can penetrate to depths of several km in
geologically and topographically favorable conditions.
In deep burial settings, the increasing temperature
causes dehydration of minerals containing combined
water and the consequent release of substantial
volumes of dehydration waters that partially replace con¬
nate waters in the sediments. Water is also generated
during maturation of organic matter and subsequent
hydrocarbon degradation, including thermochemical
sulfate reduction. Prograde metamorphism and devola¬
tilization reactions at still greater depth can liberate
water as well ( metamorphic waters), and volcanogenic
( juvenile) waters can rise into the sedimentary cover
from the deeper parts of the Earth's crust.
Groundwater Systems and Regimes
In subsiding basins the dominant flow drive in pro¬
gressively buried strata is compaction due to the
increasing load, which requires an expulsion of the
pore waters from the sediments. This is the expulsion
regime, also termed elision or exfiltration in the East
European literature. Flow is directed upward, and on
the regional scale from areas of greatest subsidence to
the margins of basins. Varying susceptibility of sedi¬
ments to compaction and pressure solution, as well as
ENCYCLOPEDIA OF CAVES
752
s
to chemical water-rock interactions, lead to increasing
inhomogeneity in porosity and permeability. In more
homogeneous strata and strata comprising evaporites,
large-scale density-driven (convection) flow systems
may develop. With still deeper burial and further
rise of temperature and lithostatic load a thermobaric
regime develops in which the fluid pressures are
caused by the thermal expansion of water or by the
release of water by mineral dehydration in a low-per¬
meability environment. Yet another flow regime can
be generated by tectonic strain in the vicinity of colli¬
sion and uplift areas (a compression regime). Because of
the ubiquity of some pressure-generating mechanisms
in the progressive burial environment and the com¬
monness of others, the elision regime is commonly
characterized by the abnormal (higher than hydro¬
static) pressures.
Following uplift and continental exposure, the mete¬
oric circulation regime driven by topography evolves,
also termed infiltration regime in the East European lit¬
erature. Meteoric waters increasingly flush out the for¬
mation waters from a basin, and the meteoric regime
substitutes the expulsion in the upper part of the crust,
although expulsion circulation may still predominate
in deep environments (Fig. 1). The pattern of the mete¬
oric circulation is controlled by the basin geometry
and relief, and by geological inhomogeneities that
determine permeability distribution.
In the meteoric regime, hierarchical flow systems of
increasing scales and depths are distinguished: local,
intermediate (subregional), and regional, the last two
being almost universally confined due to the layered
heterogeneity of strata. Despite the confinement, basin-
scale hydraulic continuity is a ubiquitous feature of
the meteoric regime (Toth, 1995). Cross-formational
hydraulic communication between the systems of dif¬
ferent scales occurs through both the pore media in
low-permeable strata and (in more localized and
intense mode) stratigraphic windows and transverse
disruptions such as major fractures and conductive
faults. This communication in the meteoric regime is
directed downward beneath highlands and upward
below topographic depressions. The upward commu¬
nication is generally much more intense as topographic
depressions often coincide with fault zones and other
conductive geologic weaknesses.
The meteoric water circulation (regional systems)
can penetrate to depths of several km affecting even
deeply buried strata in geologically and topographi¬
cally favorable conditions, especially in tilted and
faulted basins adjacent to mountainous areas and in
intermountain basins where the deep plunge of mete¬
oric waters is favored by descending limbs of large
convection circulation systems. In several regions
fresh, meteoric groundwater flow was found by ocean
drilling far offshore. However, the water flux through
deeper systems is often significantly less than that
through shallower systems.
The meteoric regime is perched on ubiquitously
ascending waters of the expulsion regime, commonly
overpressured (Fig. 1). Zones of interaction between
the two regimes, either crosscutting or lateral, are partic¬
ularly favorable for hypogene speleogenesis. Hypogene
speleogenesis is commonly a part of mixed flow sys¬
tems, where topography-driven flow interacts with
the deeper compaction- or density-driven regimes. The
nature and the geometry of the transition between the
two regimes (and hence of the transition from mesogen-
esis to telogenesis, in the extended connotation of the
latter) are controlled by respective fluid potentials and
geological heterogeneities, especially sedimentary win¬
dows and conductive faults. The boundary may be
blurred but it is more distinct when it coincides with
some low-permeability strata of a regional extent. In the
course of the geological evolution, the deeper strata
migrate upward through this boundary with the onsets
of uplift and denudation, and the nature and the geom¬
etry of this transition adjusts to the structure of rising
strata and changing potentials of the two regimes. The
meteoric systems at shallower depths are often pierced
by upwelling cross-formational plumes of basinal
waters.
In many generalizing schemes the hydrogeody¬
namic circulation on the continents (especially in era-
tons) is described as displaying two distinct stories.
The upper story has active meteoric water circulation
signified by the presence of considerable amounts of
dissolved oxygen and nitrogen of meteoric origin. It
corresponds to local and subregional flow systems of
Toth (1995). The notion of telogenesis should probably
be restricted to this story. The lower story, which cor¬
responds to regional flow systems, is dominated by
formation waters lacking such gases.
Hydrogeologic Cycle
The hydrogeologic cycle starts with subsidence and
marine sedimentation, continues through uplift, denu¬
dation, and exposure, and closes with the new marine
transgression. During the second cycle, the new
marine waters may enter any older lithified rocks in
the mature basin, partly replacing meteoric or forma¬
tion waters in them inherited from the previous conti¬
nental regime (Fig. 2). In deeper parts of a basin the
meteoric regime can be "trapped" by the new sedi¬
ment pile, and the expulsion regime from the first
burial stage may still be continuing. There can be sig¬
nificant time lags in the adjustment of the flow
systems to the changing conditions and mingling
ENCYCLOPEDIA OF CAVES
SPELEOGENESIS, HYPOGENIC
753
FIGURE 2 Evolution of geohydrodynamic systems and hydrogeological structures in the upper part of the hydrosphere (A) and karst
development in the context of this evolution (B—D). B = direction of changes of evolutionary types of karst in newly deposited rocks in the
course of diagenetic and geomorphic evolution; C—D = relative intensity of karst development in the course of evolution (the superposition of
bars also demonstrates a potential for inheritance of preformed karst features by evolving karst systems): C = hypogenic karst; D = epigenic
karst; E = tectonic settings; F = dominant geodynamic/geomorphic regimes.
groundwaters of differing ages and origins. The com¬
plexity is further added by increasing deformation,
disruption, and subdivision of originally unified im¬
mature groundwater basins through their history. This
leads to the directed evolution of large-scale hydrogeo¬
logic structures (Fig. 2) from artesian basins with pre¬
vailing porous stratiform groundwater bodies (layered
aquifers) through disrupted basins and layered massifs
to block-fault massifs with a prevalence of fracture
and vein (conduit) types of porosities, which are often
crosscutting.
Evolution of Karst Settings
A useful framework to characterize the changes in
karst and speleogenetic environments in the upper
part of the Earth's crust is provided by the evolution¬
ary classification of karst types (Klimchouk and Ford,
2000; Fig. 3). It distinguishes major stages in the geo¬
logic and geomorphic history of a soluble rock as the
types of karst, marked by distinct combinations of the
diagenetic and structural prerequisites for groundwa¬
ter flow and speleogenesis and different potential for
inheritance from earlier conditions. These types are (in
the order they potentially evolve) syngenetic/eogenetic
karst in freshly deposited rocks; deep-seated karst, which
develops during mesogenesis, particularly during its
ascending limb (when the rocks are being shifted
toward the surface); subjacent karst, where the cover is
locally breached by erosion; entrenched karst, in which
valleys incise below the bottom of the karst aquifer
and drain it, but where the soluble rocks are still cov¬
ered by insoluble formations for the most part; and
denuded karst, where the insoluble cover materials have
been completely removed. If karst bypasses burial, or
if the soluble rock is exposed after burial without hav¬
ing experienced any significant karstification during
burial, it represents the open karst type. Later on, karst
may become mantled by a cover that develops contem¬
poraneously with the karst, or reburied under younger
rocks to form paleokarst, and be reexposed ( exhumed
karst).
Although this classification does not directly specify the
origin of caves, it characterizes dominant speleogenetic
environments and their evolutionary changes. The types of
karst correlate well with the three major types of speleoge¬
netic settings distinguished now (Klimchouk et al, 2000).
Coastal and oceanic speleogenesis in diagenetically imma¬
ture rocks falls into the syngenetic/eogenetic karst domain.
ENCYCLOPEDIA OF CAVES
754
V U
burial with no
noticeable
speleogenesis
cover is older than karst
m cover is contemporal with karst
[x§|fj cover is younger than karst
□ soluble unit
| underlying formation
j epigene speleogenesis domain
i hypogene speleogenesis domain
FIGURE 3 Evolutionary types of karst and speleogenetic environments (modified from Klimchoiik and Ford, 2000). Background colors indicate
the domains of hypogenic and epigenic speleogenesis.
Deep-seated karst is represented exclusively by hypogenic
speleogenesis. In subjacent karst both hypogenic and epi¬
genic speleogenesis may operate, depending on the domi¬
nant groundwater regime and interaction of flow systems,
but hypogene speleogenesis still dominates. Exposed karst
types (entrenched and denuded) are overwhelmingly epi¬
genic, with an inheritance of hypogenic features that can
be reworked by epigenic processes or get fossilized. In
both karst types, however, ceasing hypogene systems may
still operate. Open karst is marked by exclusively epigenic
speleogenesis.
Very Deep Perspective: The Endokarst Realm
The realm of mesogenesis grades downward into
the metamorphic realm. Water and CO 2 are gener¬
ated there through prograde metamorphism and
devolatilization reactions, and deeper juvenile fluids
rise from still greater depths. Little is known about
fluid flow and dissolution pattern at great depth.
There is mounting evidence for extensive fluid circu¬
lation in very deep boreholes in many regions, for
example, large flows of hot brine detected at depths
to greater than 9 km in the deep borehole of the Kola
Peninsula, which proves the existence of fluid-filled
fracture zones at these depths in the crust. In many
instances, mineralization of water at great depths is
comparatively low (5—10 gL~* and less) and the total
amount of dissolved gases, mainly CO 2 and H 2 S, is
high (up to 60—70 vol %).
An important conceptual framework about the
hydrodynamics of the Earth's crust has been devel¬
oped by Russian scientists in the 1980s—1990s (S.N.
Ivanov, Y.A. Ezhov, G.A. Lysenin, and others; see a
summary publication by Ivanov and Ivanov, 1993),
based on observations from very deep boreholes, theo¬
retical calculations and experimental results of rock
deformation at high temperatures and pressures, and
seismic data (Fig. 4). These ideas were discussed from
the perspective of karst in Dublyansky (2000),
Klimchouk (2000), and Andreychouk et al. (2009).
ENCYCLOPEDIA OF CAVES
SPELEOGENESIS, HYPOGENIC
755
A BCD
FIGURE 4 Vertical hydrodynamic zoning (A) and karst stories (D) of the Earth's crust. (B) and (C) show, respectively, dominating flow
regimes and the origins of groundwater in different zones and stories. (A) also shows changes of some important parameters with depth: solid
line—parameters before the impulse breakthrough of fluids through buffer subzone II-A; dotted line—parameters after the breakthrough. See
text for definitions of zones and subzones. (A) is from Ezhov and Lysenin, 1990, as reproduced in Andreychouk et ah, 2009.
According to this model, the almost complete closure
of fissures, cavities, and open pores in rocks at a cer¬
tain depth (commonly of about 7—15 km but some¬
times less), creates a dense "buffer" interval of
maximum compaction almost impermeable to fluids.
This interval (subzone II-A; the upper part of the
mesozone) serves as a planetary-scale regulator of
defluidization of the deepest parts. Above it lies a
hydrostatic zone (I), in which fluid pressures are close
to the hydrostatic pressures, not exceeding them by
more than 10—25%.
Below the buffer interval is a transition subzone (II-
B) marked by variable suprahydrostatic pressures,
decreasing mineralization, reduced rock density, and
high porosity and fracturing. This grades downward
into the deepest zone where liquid-vapor fluids
released from thermal breakdown of hydrous minerals
and arriving from the lower crust and mantle are
under lithostatic pressure (lithostatic zone III in Fig. 4).
The buffer subzone is the base of a variety of tectonic
structures, including the overwhelming majority of
deep faults and hydrothermal conduit systems. Under
seas and oceans, due to additional pressure from the
water column, it is closer to the sea floor than it is to
the land surface on the continents. When pressure of
fluids beneath the buffer subzone reaches values suffi¬
cient for hydrofracturing, the fluids break through it
into the above hydrostatic zone. Such periodic break¬
throughs of pulses of deep-seated fluids are believed to
be responsible for local aquifer pressure anomalies in
the hydrostatic zone and associated thermal.
hydrochemical, and gas anomalies. In favorable struc¬
tural conditions, such plumes may transect the hydro¬
static zone to various heights and up to the surface,
causing respective anomalies in all shallower ground-
water systems (Fig. 5). Apparently, these aggressive
fluxes play the important role in generating hypogene
speleogenesis in the hydrostatic zone along and around
these paths, especially hydrothermal speleogenesis.
Based on the above model, a new notion of endokarst
has been suggested (Ezhov and Lysenin, 1990; Andrey¬
chouk et ah, 2009) that radically differs from the conve¬
nient usage of this term in the karst literature (i.e., simply
meaning underground karst) but correlates well with a
broader geologic usage. The endokarst realm encom¬
passes the deep parts of the crust below the buffer sub¬
zone. Endokarst processes involve metamorphic and
igneous fluids and act at elevated temperatures
(>80—100°C) and pressures much exceeding the hydro¬
static ones. In such conditions the fluids are highly
aggressive with respect to many sedimentary, metamor¬
phic, and igneous rocks, not only to conventional soluble
rocks such as carbonates and sulfates. In these conditions,
solubility of quartz, for instance, becomes comparable
to that of gypsum or anhydrite in near-surface condi¬
tions (Fig. 6H). Cavities may exist in the endokarst
story because fluids there are under lithostatic pres¬
sures so that effective pressure is zero. However, they
can be preserved while passing through the above
buffer interval only if filled with some secondary min¬
eral such as calcite. This could be removed in the
course of further dissolution in the mesogenetic zone.
ENCYCLOPEDIA OF CAVES
756
s
FIGURE 5 Highly idealized conceptual representation of the cross-formational nature of hypogene speleogenesis in the upper part of the
Earth's crust (A), and of typical hypogene speleogenesis settings: (B) artesian settings; (C) deep-rooted settings. Hypogene speleogenesis is
related to upwelling plums of water originated at varying depths, extending upward from the source of the potential along zones of enhanced
vertical permeability. The upwelling plums interact with waters in any upper aquifers and develop zones of influence, which lateral extension
varies depending on permeabilities and the groundwater regime in the aquifers crossed. (C) is from Andreychouk et ah, 2009.
At least some of the known caves in silicate rocks could
owe their origin to endokarst processes (Andreychouk
et al, 2009).
The above cited works also suggested the term exo¬
karst be used for all karst processes operating above
the buffer interval, that is, in the entire mesogenetic
and telogenetic zones. The present author, however,
recommends limiting the use of this term by the mean¬
ing equivalent to epigenic karst, that is, karst that
develops in the shallowest groundwater systems with
direct meteoric recharge. The zone between the buffer
interval and the exokarst zone, the arena of hypogenic
speleogenesis by upwelling fluids in the hydrostatic
zone, is termed here the hypokarst zone (Figs. 4D and 5),
the main focus of this article.
HYPOGENE SPELEOGENESIS
Chemical Mechanisms
There was a long belief in the karst literature that
confined conditions generally offer limited dissolution
potential for karstification as slow artesian and basinal
flow would be chemically equilibrated with the soluble
rocks. A new perspective emerged from the recognition
of the variety of deep-seated sources of the aggres¬
siveness and of widespread leakages across strata that
were supposed to be "impermeable" within the old
artesian paradigm. The fundamental feature of hypo¬
gene speleogenesis is that it is mainly driven by trans¬
verse flow across boundaries between different strata,
diagenetic zones, and flow systems (Fig. 5), whereas the
boundaries commonly coincide with major contrasts in
water chemistry, gas composition, and temperature.
This causes disequilibrium conditions and supports
diverse mineral reactions, including those of dissolution
with the spatially concentrated effects. The variety of
processes that can potentially cause the aggressiveness
and dissolution in deep-seated settings has been charac¬
terized in Mazullo and Harris (1992), Surdam et al.
(1984), Palmer (1991, 2007), Klimchouk (2000, 2007), and
Dublyansky (2000). The most important processes are
briefly overviewed below (Fig. 6).
Carbonic acid dissolution dominating epigenic carbon¬
ate speleogenesis also operates as a hypogenic agent.
ENCYCLOPEDIA OF CAVES
SPELEOGENESIS, HYPOGENIC
757
(B)
calcite,
(C) individual calcite + dolomite,
minerals gypsum + gypsum
added effect of 2 cm 3 /L NaCI
FIGURE 6 Diagrams illustrating the potential for aggressiveness of groundwater in different conditions with respect to various rocks. (A)
The effect of mixing of water differing in CO 2 and H 2 S content on dissolution of carbonates (from Palmer, 1991). (B) Gypsum solubility as a
function of concentration of other salts. Magnesium chloride and nitrate have a particularly strong effect on gypsum solubility. (C)
Dissolution in mixed carbonate/sulfate strata ( from Palmer, 2007). The diagram shows solubilities of calcite, dolomite, and gypsum: (a) as indi¬
vidual minerals at 15°C and Pco 2 = 0.01 atm; (b) in a mixture of calcite and gypsum (both become less soluble, although the effect on gypsum
solubility is minor); (c) where water first encounters calcite and dolomite, and then gypsum (both dolomite and gypsum become more soluble
as calcite is forced to precipitate). The shaded areas show the effect of 2 cm 3 L -1 of dissolved salt (NaCI). In (c), calcite is held at SI = 0.1 to
account for the residual supersaturation required to precipitate it. (D) Change in calcite solubility when H 2 S in solution oxidizes to sulfuric
acid ( from Palmer, 1991). Dissolution is greatest if the C0 2 generated by the reaction is retained in solution. (E) and (F) Solubility of C0 2 (E)
and CaC0 3 (F) in ascending thermal water as function of temperature. Temperature changes from depth toward the surface: on (E) from 200
to 25°C; on (F) lines indicate the following ranges: 1—from 225—100°C, 2—from 225—50°C, 3—from 225—25°C, 4—from 200—100°C for solu¬
tion with ionic strength 1 = 1. (G) and (H) Dissolution of Si0 2 in water in different conditions: (G) the dissolution rate of amorphous Si0 2 at
T = 60°C and varying pH; (H) solubility of different varieties of silica with change of pH in standard conditions ( from Andreychouk et al., 2009).
ENCYCLOPEDIA OF CAVES
758
s
though the origin of the acidity is different. CO 2 is pro¬
duced at depth mainly by prograde metamorphism and
devolatilization reactions at great depths, thermal degra¬
dation and oxidation of deep-seated organic compounds
by mineral oxidants, and by several other reactions.
Efflux of CO 2 of the deep origins into upper aquifers
can be massive, point-wise or disperse, depending on
geologic/structural conditions. CO 2 from oxidation of
deep-seated organic compounds is especially common
in the vicinity of hydrocarbon fields, where waters
characteristically contain high CO 2 concentrations, and
in the areas of young volcanism.
Dissolution by cooling thermal water can occur along
ascending flow paths, even at constant CO 2 levels. In
a closed system, the solubility of calcite increases with
decreasing temperature, so that more calcite can be
dissolved (Fig. 6F). Near the surface, C0 2 concentra¬
tion decreases and the solubility of calcite drops
(Fig. 6E). Carbonates dissolve at depth along the ris¬
ing thermal flow paths but calcite precipitates near
the surface. The dissolution effect increases with
increasing CO 2 partial pressure. This mechanism is
known as hydrothermal speleogenesis, occurring in high-
gradient zones where ascending flow is localized
along some highly permeable paths (Dublyansky,
2000). However, dissolution by cooling of thermal
water alone can produce substantial caves only where
high thermal gradients are sustained through long
periods of time. Most caves thought to be thermal
actually owe their origin to the mixing of high-CC >2
thermal water with I 0 W-CO 2 water of shallower flow
system (Palmer 1991, 2007).
Hydrogen sulfide is another common hypogene source
of acidity, abundant in deep meteoric groundwater and
basinal brines, generated at depth by microbial or ther¬
mal reduction of sulfates in the presence of organic car¬
bon. When dissolved in water, hydrogen sulfide forms
a mild acid, but it can cause carbonate dissolution if
this water flows elsewhere and mixes with water that
has little or no H 2 S content (Palmer, 1991, 2007), or
escapes from reducing zones as a gas and is reabsorbed
in freshwater. H 2 S can also react with dissolved metals,
such as iron, to produce sulfide ores and dissolve car¬
bonate rocks.
As both CO 2 and H 2 S are commonly present
together in deep groundwater systems, and their
proportion changes in the course of basinal evolution,
the particular contribution of the above dissolution
mechanisms in hypogene speleogenesis is difficult to
discriminate.
Mixing of waters of contrasting chemistry, particularly
those differing in CO 2 or H 2 S content or salinity, can
renew or enhance solutional aggressiveness to carbo¬
nates due to the concavity of solubility curves
(Fig. 6A), the effect widely referred to in the karst
literature as mixing corrosion. Gases have a greater
effect than salinity in controlling mixing dissolution,
especially when differences in equilibrium gas con¬
tents are large (Palmer, 1991). Mixing corrosion is
highly relevant for hypogene systems, where water
that rises from depth encounters shallower meteoric
water along paths of the cross-formational flow.
Dissolution by sulfuric acid, a very strong speleogenetic
agent (Fig. 6D), occurs in shallower conditions where
H 2 S-bearing waters rise to interact with oxygenated shal¬
lower groundwaters. It is recognized as the main speleo¬
genetic process for certain large caves ( e.g., caves in the
Guadalupe reef complex in the United States and in
the Apennines in Central Italy) and many smaller ones.
The dissolution by sulfuric acid is most readily observable
in caves that are now in unconfined settings but continue
receiving rising H 2 S-bearing water from depth. Intense
production of sulfuric acid near the water table and in the
subaerial (above the water table) conditions causes pro¬
nounced speleogenic and mineralogical effects, which led
some researchers to believe that most of the cave origin is
due to subaerial sulfuric acid dissolution. Substantial sul¬
furic acid dissolution can also be caused by oxidation of
iron sulfides such as pyrite and marcasite, localized in
ore bodies or along certain horizons or bedding planes in
carbonates (Lowe and Gunn, 1997).
Organic acids generated in great quantities during
the maturation of organic matter in the mesogenetic
environment can significantly enhance the solubility
of carbonates (Surdam et al., 1984; Palmer, 2007) and
therefore be a powerful agency for hypogenic speleo¬
genesis. Their dissolution effect appears to be the most
pronounced under high (above 80° C) temperatures, as
at lower temperatures they are readily metabolized by
bacteria. At still higher temperatures, organic acids
break down to CH 4 and CO 2 , and the latter forms car¬
bonic acid to dissolve more carbonates. Although
some of the reactions involved may lead also to precip¬
itation, the onsets of dissolution and precipitation can
be spatially separated.
Dissolution of evaporites is a simple dissociation of
the ions and diffusional mass transport from the sur¬
face into the solution. The solubility of gypsum
increases significantly (up to 3—6 times) with the pres¬
ence of other salts in the solution (Fig. 6B). Gypsum
dissolution can be rejuvenated by reduction of sulfates,
which removes sulfate ions from the solution and
allows more sulfates to dissolve, and by dedolomitiza-
tion, which generates further dissolutional capacity
with respect to gypsum because Ca 2+ is removed from
solution and the sulfate ions react with Mg (Fig. 6C).
Because mixed evaporite (gypsum/anhydrite and salt)
and evaporate/carbonate formations are common in
sedimentary successions, the above effects are
extremely relevant to hypogene speleogenesis.
ENCYCLOPEDIA OF CAVES
SPELEOGENESIS, HYPOGENIC
759
Dissolution in mixed carbonate/sulfate strata is a kind
of synergetic process, more complex than in each
lithology alone (Palmer, 2007), which does not require
an acid source. In a mixed system, up to 1.5 times
more gypsum can dissolve, and up to 7 times more
dolomite, compared to the solubilities of gypsum and
dolomite alone (Fig. 6C). The effect increases with
salinity of water. In a simple carbonate-sulfate system,
the volume of precipitated calcite is less than 30% of
the combined total volume of dissolved gypsum and
dolomite, which results in considerable solution poros¬
ity. The process is mediated by the relatively slow dis¬
solution of dolomite, so that it is a viable mechanism
for the development of dissolution paths in gypsum in
deep-seated settings.
There may be other chemical conditions and reac¬
tions operative at depth, poorly known so far or still
overlooked in karst science. Low-mineralization ground-
water anomalies, frequently documented at great depth,
especially in the vicinity of hydrocarbon fields, bring
about chemical disequilibrium conditions that can be
localized and extended upward along cross-forma-
tional flow paths. Chemical interaction of matrix pore
and conduit waters is still poorly studied. Radiolysis
of groundwater can locally modify redox conditions
and therefore the speciation and the solubility of the
compounds. It possibly accounts for the recorded
instances of the presence of oxygen in considerable
amounts (up to 50 vol % of dissolved gases) in water
at depths of 2 to 3 km, which can generate reactions
involving oxidation of hydrocarbons, sulfides, and
other compounds.
There is mounting evidence suggesting that more
than one process could be involved in most cases of
hypogene speleogenesis, operating either in combination
or sequentially in time.
Distribution and Settings
Recent studies made it apparent that hypogene spe¬
leogenesis is much more widespread in nature than pre¬
viously thought. It operates in various geological and
tectonic settings, at different depths (ranging from a few
tens of meters to several kilometers), due to different dis-
solutional mechanisms operating in various lithologies.
According to the scale and depth of flow systems that
drive upwelling flow, hypogene speleogenesis settings
can be classified into three types:
1. Artesian hypogene speleogenesis, related to upwelling
limbs of regional or subregional (intermediate)
meteoric systems, roughly corresponding to the
subzone I-B (Figs. 4 and 5B). This type of speleogenesis
is most common for the post-orogenic tectonic
regime marked by massive and deep penetration of
topography-driven meteoric waters.
2. Deep-rooted hypogene speleogenesis, related to
upwelling plumes generated by deep basinal
systems in the subzone I-C or fluxes from the
mesozone II-B through the buffer interval
(Figs. 4 and 5C). It is common for both the
extensional tectonic regimes of continental rifting
and the compressional regimes of active margins and
collision fold-thrust zones, favored by intense and
deep dislocations and related thermal and
geochemical cross-formational anomalies.
3. Combined artesian/deep-rooted speleogenesis, where the
settings 1 and 2 interact intensely. It typically occurs
in the wedge-top and forebulge settings of foreland
basins, as well as in disturbed cratonic zones.
A common misconception about hypogenic karst is
that it is thought to be peculiar to arid climates and
less common in humid regions. In reality, hypogenic
speleogenesis is largely independent of climate, but
the epigenic overprint of hypogenic caves, inherited by
entrenched and denuded karst types, strongly depends
on climate. The overprint is particularly strong in
regions of moderate and high runoff. In arid and semi-
arid regions epigenic speleogenesis and surface karst
morphologies are normally subdued due to limited
infiltration and short supply of CO 2 and organic acids
from the soil, but hypogenic features are often abun¬
dant in the subsurface, resulting in a strong contrast in
the degree of karstification between the surface and
subsurface.
Hydrogeologic and Structural Controls
Cross-formational flow in a heterogeneous (layered,
deformed, and fractured) sequence may have very
complex patterns. They are determined by complex
interplay of driving forces of different flow systems
and highly variable permeabilities, which vary with
classes of heterogeneities {e.g., layered and discontinu¬
ous) and types of elementary porosity structures {e.g.,
pore, vugs, and fractures) and their connectivity.
Soluble rocks, with their own variety of geological het¬
erogeneities and primary porosity structures, may
have different positions and play different functions
within the overall hydrostratigraphic framework and
flow patterns.
Distinct integrated stratiform permeability struc¬
tures, relatively independent of each other (hydro¬
stratigraphic units, or hydrofacies), can be related to
single beds or be comprised of several rock units. They
can be separated by beds of distinctly lower perme¬
ability (aquitards, or hydrostratigraphic barriers) or by
less obvious boundaries along which the vertical
ENCYCLOPEDIA OF CAVES
760
FIGURE 7 Transverse flow across several hydrostratigraphic units and interfaces in a soluble bed (A) and the morphologic suite of rising
flow (MSRF), diagnostic of hypogenic transverse origin of caves (B). In (A) a n denotes exchange terms between different permeability struc¬
tures. Modified from Klimchonk, 2007.
hydraulic connection between contrasting and discor¬
dant permeability structures in the adjacent beds is
limited (hydrostratigraphic interfaces). Leakage across
hydrostratigraphic barriers occur along major crosscut
fractures and sedimentary windows. Leakage through
hydrostratigraphic interfaces occurs through points
where permeability features of adjacent units (strati¬
form pores and vugs, or minor fractures) vertically
intersect (Fig. 7A).
Zones of enhanced vertical permeability that cross
thick sequences and interfaces between major flow sys¬
tems and diagenetic subdomains (the fracture diage-
netic setting of Machel, 1999) are rarely comprised of
single fractures. More common are combinations of
several major fractures in echelon, associated feather¬
ing fracture systems and lithological variations within
different intervals. The vertical permeability of such
zones varies at different levels. Flow along such struc¬
tures is invariantly vertical within hydrostratigraphic
barriers, but it may diverge laterally where high-
permeability beds are intersected, especially if the
guiding vertical structure terminates upward at one
location but continues into the upper strata in the other
nearby location.
Though both insoluble and soluble units can ini¬
tially (prior to speleogenesis) serve as hydrostrati¬
graphic barriers in an aquifer system, their further
behavior is different. The vertical permeability of the
former remains conservative, while in the latter the
conduit permeability develops in the course of speleo¬
genesis, so that they lose their separating function
(Fig. 7). The same happens with hydrostratigraphic
interfaces within the heterogeneous soluble sequence
(Fig. 7A). In this way, hypogene speleogenesis serves
to enhance cross-formational hydraulic communication
across initially less pervious soluble beds or hydrostra¬
tigraphic interfaces. The net result is that multiple
aquifer systems achieve greater integration.
In epigenic karst (exokarst), directly recharged from
the surface, the mode of recharge evolves in coordi¬
nation with the development of solution conduits.
The evolution is characterized by a switch from the
hydraulic control of flow (by size of a conduit itself) to
the catchment control (available recharge) when the
conductivity of a growing conduit becomes larger than
available recharge (Palmer, 1991).
In contrast, in hypogene settings, recharge to a soluble
sequence comes from the adjacent strata below and dis¬
charge occurs through the overlying strata. Recharge is
independent from the surface conditions. Both recharge
and discharge can be diffuse or point-wise, laterally coin¬
ciding or shifted, which determines to a large extent the
pattern of speleogenesis in the soluble sequence.
The mode of recharge is conservative, although its
amount and chemical characteristics may change on
the long run of the geologic time scales as the deep
driving force or the regional flow system changes.
Discharge conditions are more dynamic as they
change with erosion and breaching of the cover and the
major confinement. The breaching can be by external
geomorphological processes, such as erosional entrench¬
ment, or by the internal development of the hypogene
cave system, such as vertical stoping of breakdown
structures above cavities across the confining strata.
Where there is no major geologic confinement and
semiconfined conditions are caused by the layered het¬
erogeneity within the soluble sequence itself, changes
in the discharge conditions are most dynamic and occur
mainly through the internal development of a cave
system.
ENCYCLOPEDIA OF CAVES
SPELEOGENESIS, HYPOGENIC
761
As recharge to and discharge from the soluble rock
occurs through the adjacent strata with relatively con¬
servative permeabilities, flow rates are controlled by
the permeability of the least permeable member.
Thus there is an external conservative hydraulic con¬
trol on the amount of flow through the evolving con¬
duits. When conduits have grown to the breakthrough
conditions and the hydraulic gradient across the unit
is diminished, their further growth is not accelerated
significantly as flow across the soluble bed is now con¬
trolled not by the hydraulic resistance of the conduit
system but by the permeability of the least permeable
member, and by the boundary conditions of the sys¬
tem. This suppresses the positive flow-dissolution
feedback and speleogenetic competition, the main
mechanism acting in unconfined (epigenic) speleo-
genesis, and promotes more pervasive and uniform
enlargement of initial permeability structures in the
cave-forming zone where recharge diffuses.
The presence of hydrostratigraphic interfaces within
the soluble formation, as well as the restricted outflow,
also favors pervasive conduit development. If recharge
and discharge are both through a single major crosscut¬
ting fracture, the conduit development in the soluble
unit is localized, although it can get dispersed below the
upper insoluble barrier that restricts outflow despite the
growth of the conduit in the soluble unit. In hydrother¬
mal speleogenesis in carbonates there is another specific
mechanism, caused by the thermal coupling between
the fluid and rock, that also suppresses speleogenetic
competition.
Large, deep, and complex transverse structures sup¬
porting cross-formational upwelling flow are com¬
monly geologically long-lived, although periods of
intense and abated activity may alternate. Geochemical
environments may migrate along them and change
many times, promoting alternating dissolution and pre¬
cipitation conditions. Hypogene speleogenesis at differ¬
ent depths along such structures can operate through
various processes, either simultaneously or sequen¬
tially. The resultant solution porosity structures,
aligned to the same transverse zone, can have multi¬
phase evolution and complex spatial organization,
although they can be represented by certain elementary
patterns within particular intervals.
Porosity Patterns
Solution porosity patterns produced by hypogenic
speleogenesis are the result of the complex interaction of
structural, hydraulic, and geochemical conditions, all
varying in the course of geological evolution. The overall
position of cave-forming zones is controlled by the 3D
distribution of soluble rocks, their position and hydrau¬
lic function within the flow systems, the nature and
geometry of major crosscutting fractures, and the pattern
of geochemical environments in a given setting. Overall
patterns of cave systems are strongly guided by the ini¬
tial (prespeleogenetic) permeability features in a sedi¬
mentary sequence, that is, by the spatial distribution of
the permeability structures and hydrostratigraphic inter¬
faces within the soluble and adjacent units; by the mode
of water input to, and output from, cave-forming zones;
and by the overall recharge-discharge configuration in
the multiple aquifer system. The presence of crosscut¬
ting permeability features such as major fractures can
exert strong effects on cave patterns through their
inflow, throughflow, and outflow controls. In contrast to
epigene settings where initial effective permeability
structures are exploited by speleogenesis in a very selec¬
tive manner, hypogene speleogenesis tends to exploit
most of them within cave-forming zones, provided the
aggressiveness is maintained. Geochemical interactions
of flow components guided by transverse and lateral
permeability pathways determine zones of pronounced
speleogenetic development and influence the resultant
patterns. General evolutionary factors, such as regional
tectonic and geomorphic developments that control rates
and architecture of flow and timing of speleogenesis,
also affect cave patterns forming in hypogene settings.
Hypogene caves demonstrate a variety of patterns,
as classified and briefly described below. For more
extended discussion refer to Palmer (1991, 2007), Ford
and Williams (2007), Klimchouk (2000, 2007, 2009), and
Audra et al. (2009). Same patterns are known to form in
different lithologies and by different dissolution mecha¬
nisms. Certain patterns common for hypogene speleo¬
genesis can also be formed locally in epigenic
environments. Conversely, branchwork patterns, the
most common one for epigenic speleogenesis, never
form in hypogenic settings. This reflects the funda¬
mental difference between the mechanisms of epi¬
genic speleogenesis, largely competitive, and of
hypogenic speleogenesis, in which the competition
between alternative flow paths is subdued.
The following elementary cave patterns are typical
(although not necessarily exclusive) for hypogene
speleogenesis:
• Single passages or rudimentary networks of passages
• Cavernous edging along transverse hypogene
conduits
• Network maze
• Spongework maze
• Irregular isolated chambers
• Rising, steeply inclined passages or shafts
• Collapse shafts over large hypogenic voids and
breccia pipes
Single isolated passages or rudimentary netzvorks of pas¬
sages are probably the most common cave structures of
ENCYCLOPEDIA OF CAVES
762
s
hypogene origin. They form by transverse flow across
a single soluble bed where guiding cross-bed fractures
are isolated or have only limited lateral connection. The
passages are typically slot-like, sometimes with rounded
sections in the middle or upper parts of the cross-sec¬
tion, single or combining in small clusters of few inter¬
secting passages according to the pattern of guiding
fractures. Their notable characteristic is that they are
dead-end in every direction. They can be observed
directly when exposed in cliffs or encountered by
underground mines. The cross dimensions vary from
several meters to a few centimeters (dissolutionally
enlarged fractures), the latter being the most common.
Cavernous edging is represented by zones of cavern¬
ous porosity comprised of closely spaced irregular
touching vugs and small conduits (up to 2—5 cm in
diameter but sometimes much larger) which frequently
border walls of crosscutting fracture conduits up
to depths of a meter or more. The size of voids
diminishes in the direction normal to walls, and farther
into the rock they pinch out. The cavernous edging can
be stratiform or distributed by clusters along the frac¬
ture/ conduit walls. The formation of this type of caver-
nousity is apparently related to the mixing corrosion
effect, where pore water in more diffusely permeable
beds or zones converges toward a conductive fracture.
Network maze patterns are common for hypogenic
caves. Passages are strongly fracture-controlled and
form more or less uniform networks, which may dis¬
play either systematic or polygonal patterns, depend¬
ing on the nature of the fracture networks. Systematic,
often rectilinear, patterns are most typical, displaying
networks of regular cross-bed fractures. Networks of
different density and orientation may be present at dif¬
ferent stories of a single cave. Network maze caves of
hypogene origin are known in limestones, dolomites,
and gypsum, in mixed limestone-dolomite-gypsum
strata, and also in conglomerates. A common feature
of network mazes is a very high passage density.
Spongework maze patterns are much less common
than networks. Highly irregular passages develop
through enlargement and coalescing of pore- and
vuggy-type initial porosity in those horizons of the
cave-forming zone that have no fractures but intercon¬
nected pores and vugs. Clusters of spongework-type
mazes are commonly combined with other patterns in
the adjacent intervals to form complex cave structures.
Isolated chambers commonly occur at the bottoms of
soluble sequences but they are also frequent within
particular beds in stratified carbonate sequences. They
form in two situations: (1) by buoyant dissolution at
the bottom of a soluble formation, commonly evapor-
ites, where there is a major underlying aquifer; (2)
where aggressive recharge from below is localized
along major crosscutting fractures and aggressiveness
is enhanced by mixing of rising deep flow with a
shallower flow system. Similar chambers form where
there is a point-wise input of geogenic gases into the
basal aquifer. Irregular chambers of the hypogene origin
can attain very large dimensions, such as directly docu¬
mented cavities in evaporates of southern Harz,
Germany (cavities of the schlotten type), the Big Room
in Carlsbad Cavern, or a huge cavity encountered by
boreholes in the Archean and Proterozoic marbles in
southern Bulgaria with a maximum vertical dimension
of 1340 m and an estimated volume of 237.6 million m 3 ,
probably the largest known, although not accessible,
cave chamber on Earth. Smaller isolated chambers are
abundant everywhere. Where large irregular chambers
are complicated by blind branches extending outward,
the pattern is termed ramiform.
Rising, steeply inclined passages or shafts are outlets of
deep hypogene systems in which the "root" structure
remains unknown in most cases. A type example is the
392-m-deep Pozzo del Merro near Rome, Italy, pre¬
sumably formed by rising thermal water charged with
C0 2 and H 2 S.
Collapse shafts over large hypogenic voids can form
mega-sinkholes at the surface, locally termed cenote
(Mexico) and obruks (Turkey). The type examples are
Sistema El Zacaton in Mexico (several features with
depth up to 329 m from the surface and ~317 m
below the water table), obruks in the Central Anatolia,
Turkey (71 features with depths up to 125 m from the
surface), and cenotes near Mt. Gambier in southeastern
Australia. These collapse features indicate giant cham¬
bers formed at the base around localized inputs of
fluids charged with volcanogenic C0 2 in the areas of
young volcanism.
Fossil vertically extended collapse features known
as breccia pipes, collapse columns, or geologic organs, and
also termed vertical through structures, are a common
by-product of hypogene speleogenesis in a variety of
geological settings. They originate from collapse of
large cavities, such as described above, and propagate
upward by stoping across sedimentary successions
that include soluble rocks. Such features may reach
many hundreds of meters in the vertical extent. They
are not merely breakdown structures but complex
hydrogeologic structures whose development depends
on (and favors) focused cross-formational groundwater
circulation and continuing dissolution of intercepted
soluble beds and infallen clasts.
Composite 3D systems are comprised of various elemen¬
tary patterns at different levels, such as irregular cham¬
bers, clusters of network or spongework mazes, and
rising subvertical conduits and other morphs connecting
them. Composite 3D systems include many of the world's
largest documented caves. Their organization reflects ver¬
tical heterogeneity in distribution of initial permeability
ENCYCLOPEDIA OF CAVES
SPELEOGENESIS, HYPOGENIC
763
structures that guided, redistributed, or impeded rising
flow within a soluble formation, and the interplay bet¬
ween structural, hydrogeological, and chemical factors.
Composite 3D cave structures may develop within a
rather thin bed and be laterally extended ( e.g., two- to
four-story giant mazes in the western Ukraine con¬
fined within the 16—30-m-thick gypsum strata), or be
vertically extended through a range of several hundred
meters {e.g., Monte Cucco system, central Italy: 930 m;
Lechuguilla Cave in the Guadalupe Mountains, New
Mexico, U.S.A.: 490 m).
Multistory mazes are "layered" variants of compos¬
ite 3D systems. In multistory mazes, stories (up to 5—6)
are commonly comprised of stratiform networks differ¬
ing in patterns, connected via rising conduits or chim¬
neys, or cross-cutting rift passages. The stories can
superpose each other within the same area or display a
staircase arrangement within a system, with cave areas
at different stories shifted relative to each other because
areas of discharge through the cover are shifted with
respect to areas of preferential recharge from below.
Some vertically extensive 3D caves have prominent
feeders as large isolated, steeply ascending passages, or
clusters of rift-like passages connecting to some master
levels of passages and chambers, and prominent outlet
segments rising from the latter. Many such caves are com¬
posed of network and spongework mazes at various
levels connected through rising vertical conduits, coalesc¬
ing with large chambers and passages. Other notable var¬
iants include complex bush-like upward-branching
structures, composed of rising sequences of chambers and
large spherical cupolas, and network maze clusters at the
base of a soluble sequence with ascending staircase limbs
of vertical ascending pits and subhorizontal passages at
different levels (outlet components). Composite 3D struc¬
tures dominated by large irregular chambers, from which
blind branches extend outward, are distinguished as rami-
form patterns.
Because of the transverse nature of hypogene spe-
leogenesis, caves and their parts tend to have a clus¬
tered distribution in the plan view, although clusters
may merge to extend over considerable areas.
Laterally extensive multistory maze caves such as in
the western Ukraine, or in the Black Hills, South
Dakota, or vertically extensive 3D structures such as in
the Guadalupe Mountains and the Apennines, are in
fact combinations of many clusters representing rela¬
tively independent transverse flow subsystems.
Mesomorphology Features
In spite of the great variability in geological and
chemical conditions of hypogene speleogenesis and
in size and patterns of hypogene caves, their
mesomorphological features exhibit the remarkable simi¬
larity at all scales. This arises due to several reasons: (1)
generally sluggish conditions of rising flow in which
hypogene speleogenesis occurs; (2) the great role of
buoyant convection due to density differences and the
related abundance of upward-oriented morphs; and (3)
the clustered nature of cavities resulting from transverse
flow along multiple closely spaced paths, and occasional
lateral merging of individual clusters.
Individual occurrences of some features characteris¬
tic for hypogene caves, such as cupola-form solution
pockets or ceiling half-tubes, also occur in epigenic
caves where they form in the unconfined phreatic zone
or subaerial conditions. Specific to hypogene speleo¬
genesis is, however, that different morphs commonly
occur in spatially and functionally related groups
where fluid flow paths, including distinct buoyant con¬
vection components, can be traced from rising inlet
conduits, through transitional wall and ceiling features
(rising wall channels and ceiling half-tubes), to outlet
features (cupolas and domepits). This regular combina¬
tion has been distinguished as the morphologic suite
of rising flow (MSRF; Klimchouk, 2007, 2009; Fig. 7B).
It provides diagnostic evidence for hypogene speleo¬
genesis. MSRF has been recognized in hundreds of
hypogene caves across the globe. The diagram on
Fig. 7B is generic and elastic; it can be stretched verti¬
cally, and complexity can be added to allow for multi¬
ple stories. The pattern of the forms will repeat itself
on each story. Hypogene caves may consist of a few
elementary segments such as the one depicted in
Fig. 7, or combine hundreds and thousands of laterally
merged segments within a single system.
Among elementary morphs in this suite, inlet fea¬
tures (termed feeders, vents, or risers in recent publica¬
tions) are the most indicative of recharge from below
and hence the hypogene origin for a cave. Original
feeders are basal input points to hypogenic cave sys¬
tems, the lowermost components, vertical, or subverti¬
cal conduits, through which fluids rise from the source
aquifer. The vertical extension of feeders can be a few
meters, such as in the gypsum mazes in the western
Ukraine fed by a basal artesian aquifer, but they can
rise through several hundred meters, as documented
in mines that intercepted feeder conduits below a
currently shallow-lying relict multistory maze cave
in dolomites of the Neoproterosoic Vazante Group,
Minas Gerais, Brazil.
Where the water table is established at a given level
within a former confined hypogenic cave system, sig¬
nificant lateral widening and merging of passages into
large chambers may occur, with characteristic speleo-
gens such as horizontal notches and corrosion tables.
This is typical for caves in evaporites and limestones
subjected to back-flooding from a nearby river, and for
ENCYCLOPEDIA OF CAVES
764
s
limestone caves where deep recharge by sulfidic deep
waters continues during the unconfined development.
In the latter case, distinct upward-oriented morpholog¬
ical effects of subaerial sulfuric-acid dissolution (i.e.,
condensation corrosion above the water table) may
also develop, such as air-convection cupolas, which
are difficult to distinguish from speleogens formed
under submerged conditions. Their distribution, how¬
ever, is restricted to the zone immediately above the
water table, whereas in composite 3D systems the ele¬
ments of the MSRF are systematically distributed
through various stories in large vertical ranges.
For a more extended recent discussion of the mor¬
phology of hypogene caves the reader is referred to
Klimchouk (2007, 2009), Palmer (2007), and Audra
et al. (2009).
Hypogene Speleogenesis and Paleokarst
Within the conventional, predominantly epigenic,
karst paradigm, instances of deep-seated karst have
commonly been interpreted as paleo-(epigenetic) karst
because the possibility of karstification in deep environ¬
ments without recharge from the immediately overlying
or adjacent surface has been neglected for a long time.
Paleokarst is not a particular type of karst but is,
rather, a fossilized condition. Features become paleo¬
karst as they get hydrologically decoupled from contem¬
porary systems, in contrast to relict features that exist
within contemporary systems but are removed from
the environment in which they developed (Ford and
Williams, 2007). True paleokarst is buried karst (see
Fig. 2), which is a complete infilling and burial of
exposed karst by later materials such as transgressive
marine sediments. Paleokarst horizons are reliably rec¬
ognized where they underlie unambiguous stratigraphic
unconformities related to subaerial exposure.
With growing recognition of hypogenic speleogenesis,
it becomes increasingly obvious that in many cases fea¬
tures previously interpreted as paleo-(epigenetic) karst,
including coastal/oceanic karst, can be better explained
as active or relict hypogenic features. Review of the inter¬
national literature, especially concerned with carbonate-
hosted hydrocarbon reservoirs, reveals that in many
cases, because of their occurrence beneath distinct forma-
tional contacts, paleokarst features have been dubiously
interpreted as evidence of subaerial exposure by a pro¬
cess of reciprocal reasoning. Other common cases of
problematic paleokarst are stratiform breccia horizons
which are commonly the ultimate result of hypogenic
speleogenesis, namely, the collapse of laterally extensive,
stratigraphically conformable maze caves.
Hypogene speleogenesis tends to operate over long
time spans, intermittently or being repeatedly suspended
and reactivated. It may even continue uninterruptedly
from one hydrogeologic cycle through another (Fig. 2).
Hypogenic features may become relict but still remain
within the contemporary systems, for example, in a sys¬
tem where original confinement was breached and the
flow pattern reversed from upwelling to descending.
Hypogenic features are not paleokarst unless their evolu¬
tion is completely halted by the removal of the cave¬
forming units from the geological section and their sub¬
stitution by stratiform breccia horizons, complete sealing
by cementation (mineralization), or lithification of the fill
material. Establishing the paleostatus of hypogenic fea¬
tures and their distinction from epigenic paleokarst
requires additional discussion.
Many important hydrocarbon and mineral deposits are
karst-related, and commonly thought to be related to
paleo-(epigenic) karst. Better understanding of hypogene
speleogenesis is of paramount importance for their proper
genetic interpretation, which in turn is crucial for the
development of more adequate approaches to the predic¬
tion, prospecting, and development of these resources.
See Also the Following Articles
Speleogenesis, Telogenetic
Sulfuric Acid Caves
Bibliography
Andreychouk, V., Dublyansky, Y., Ezhov, Y., & Lysenin, G. (2009).
Karst in the Earth’s crust: Its distribution and principal types.
Sosnowiec—Simferopol: University of Silesia—Ukrainian Institute
of Speleology and Karstology.
Audra, P., Mocochain, L., Bigot, J.-Y., & Nobecourt, J.-C. (2009).
Morphological indicators of speleogenesis: Hypogenic speleo¬
gens. In A. B. Klimchouk & D. C. Ford (Eds.), Hypogene speleogen¬
esis and karst hydrogeology of artesian basins (pp. 17—22).
Simferopol: Ukrainian Institute of Speleology and Karstology
(Special Paper 1).
Choquette, P. W., & Pray, L. C. (1970). Geologic nomenclature and
classification of porosity in sedimentary carbonates. American
Association of Petroleum Geologists Bulletin, 54, 207—250.
Dublyansky, Y. V. (2000). Hydrothermal speleogenesis—Its settings
and peculiar features. In A. B. Klimchouk, D. C. Ford, A. N.
Palmer & W. Dreybrodt (Eds.), Speleogenesis: Evolution of karst aqui¬
fers (pp. 293—297). Huntsville, AL: National Speleological Society.
Ezhov, Y. A., & Lysenin, G. A. (1990). Vertical zonation of karst
development. Izvestija AN SSSR Serija Geologii, 4, 108—116 (in
Russian)
Ford, D. C. (2006). Karst geomorphology, caves and cave deposits: A
review of North American contributions during the past half cen¬
tury. In R. S. Harmon & C. W. Wicks (Eds.), Perspectives on karst
geomorphology, hydrologi/ and geochemistry (pp. 1—14). Geological
Society of America (Special Paper 404).
Ford, D.C., & Williams, P.W. (1989). Karst geomorphology and hydrol¬
ogy. London, Unwin Hyman.
Ford, D. C., & Williams, P. W. (2007). Karst hydrogeology and geomor¬
phology. Chichester, U.K.: Wiley and Sons.
ENCYCLOPEDIA OF CAVES
SPELEOGENESIS, TELOGENETIC
765
Ivanov, S. N., & Ivanov, K. S. (1993). Hydrodynamic zoning of the
Earth's crust and its significance. Journal of Geodynamics, 17(4),
155-180.
Klimchouk, A. B. (2000). Speleogenesis under deep-seated and con¬
fined settings. In A. B. Klimchouk, D. C. Ford, A. N. Palmer &
W. Dreybrodt (Eds.), Speleogenesis: Evolution of karst aquifers
(pp. 244—260). Huntsville, AL: National Speleological Society.
Klimchouk, A. B. (2007). Hypogene speleogenesis: Hydrogeological and
morphogenetic perspective. Carlsbad, NM: National Cave and Karst
Research Institute.
Klimchouk, A. B. (2009). Morphogenesis of hypogenic caves.
Geomorphology, 106, 100—117. doi: 10.1016/j.geomorph.2008.09.013.
Klimchouk, A. B., & Ford, D. C. (2000). Types of karst and evolution
of hydrogeologic settings. In A. B. Klimchouk, D. C. Ford, A. N.
Palmer & W. Dreybrodt (Eds.), Speleogenesis: Evolution of karst
aquifers (pp. 45—53). Huntsville, AL: National Speleological
Society.
Klimchouk, A. B., & Ford, D. C. (Eds.), (2009). Hypogene speleogenesis
and karst hydrogeology of artesian basins. Simferopol: Ukrainian
Institute of Speleology and Karstology.
Klimchouk, A. B., Ford, D. C., Palmer, A., & Dreybrodt, W. (Eds.),
(2000). Speleogenesis: Evolution of karst aquifers. Huntsville, AL:
National Speleological Society.
Lowe, D. J., & Gunn, J. (1997). Carbonate speleogenesis: An inception
horizon hypothesis. Acta Carsologica, 26(2), 457—488.
Machel, H. G. (1999). Effects of groundwater flow on mineral diagen¬
esis, with emphasis on carbonate aquifers. Hydrogeology Journal, 7,
94-107.
Mazzullo, S. J., & Harris, P. M. (1992). Mesogenetic dissolution: Its
role in porosity development in carbonate reservoirs. American
Association of Petroleum Geologists Bulletin, 76, 607— 620.
Palmer, A. N. (1991). Origin and morphology of limestone caves.
Geological Society of America Bulletin, 103(1), 1—21.
Palmer, A.N. (1995). Geochemical models for the origin of macro¬
scopic solution porosity in carbonate rocks. In D. A. Budd,
P. M. Harris & A. Sailer (Eds.), Unconformities in carbonate strata:
their recognition and the significance of associated porosity
(pp. 77—101). American Association of Petroleum Geologists Memoir 63.
Palmer, A. N. (2007). Cave geology. Dayton, OH: Cave Books.
Stafford, K., Land, L., & Veni, G. (Eds.), (2009). Advances in hypogene
karst studies. Carlsbad, NM: National Cave and Karst Research
Institute (NCKRI Symposium 1).
Surdam, R. S., Boese, S., Ik Crossey, C. (1984). The chemistry of sec¬
ondary porosity. American Association of Petroleum Geologists
Memoir, 37, 127—149.
Toth, J. (1995). Hydraulic continuity in large sedimentary basins.
Hydrogeology Journal, 3(4), 4—15.
SPELEOGENESIS, TELOGENETIC
F rand Qabrovsek
Karst Research Institute ZRC SAZL7, Slovenia
INTRODUCTION
We describe basic principles of meteoric (as opposed
to hypogenic) speleogenesis in telogenetic (as opposed to
eogenetic) rocks. In other words, we discuss cave gene¬
sis in mature rocks with dominant fracture porosity
caused by normal circulation of meteoric water. A
dividing line between each of the above-mentioned
systems is often not clear, for example, speleogenesis
in the telogenetic stage can use horizons developed in
the early eogenetic stage. Also a clear distinction
between meteoric and hypogenic cave development is
often not possible. We primarily discuss speleogenesis
in limestone, although many of the principles are valid
for other soluble rock.
Speleogenesis is a result of dissolution of soluble
rocks by groundwater. It is a complex interplay of
many factors and processes leading to the evolution
of conduit networks in karst aquifers. The dynamics of
speleogenesis and the resulting geometry of the con¬
duit network strongly depend on local hydrology,
lithology, structure, geochemical settings, and topogra¬
phy. Each of these imposes different parameters which
can, as well, change in time.
Cave genesis is a major topic in speleology, with
implications in other fields, particularly in karst
hydrology. Many important papers discuss speleogen¬
esis in various settings. They compose an invaluable
base of knowledge of which only a part is summarized
in this entry. An interested reader is referred to a few
textbooks that contain most of what we know today on
speleogenesis and contain many references to original
works. The book edited by Klimchouk et al. (2000) pro¬
vides information about history, basic principles, and a
broad review of speleogenesis in different settings. For
a comprehensive review of numerical modeling of spe¬
leogenesis in the past two decades, the volume by
Dreybrodt et al. (2005) is recommended. Ford and
Williams (2007) is considered to be the major reference
in karst by many; it dedicates one of its 12 chapters to
speleogenesis. Probably the best source to learn about
speleogenesis is the book by Palmer (2007), which suc¬
cessfully links the basic principles with many field
examples.
BASIC DYNAMICS OFJSPELEOGENESIS
The simplest scenario includes a single flow path
via one or a series of fractures connecting the points
of recharge and discharge at different hydraulic heads
(Fig. 1). Hydraulic head, h, is defined as h = z + p/
( pg ), where the first term, z, presents elevation head
and the second, p/ipg), the pressure head, p is the
density of water and g is the gravitational accelera¬
tion. Water tends to flow from higher to lower head.
Water that is aggressive to rock-forming minerals
(e.g., calcite) dissolves fracture walls. The concentra¬
tion of ionic species, and correspondingly the satura¬
tion ratio of water with respect to the mineral,
increases along the flow path.
Encyclopedia of Caves.
© 2012 Elsevier Inc. All rights reserved.
766
FIGURE 1 A single pathway evolving from the point of recharge
elevated by h above the spring.
If dissolution rate (the rate of wall retreat is directly
proportional to dissolution rate) decreases linearly
with saturation ratio, it decreases exponentially with
the flow distance along the fracture. If the rate law is
nonlinear, the dissolution rate drops hyperbolically
along the fracture. One can introduce penetration
length. A, which defines how rapidly from the entrance
the dissolution rates drop below some value. For the
nonlinear rate law, the penetration length A is propor¬
tional to
QP " ('-;£)
where P is the fracture perimeter, Q is the flow rate, c m
is the concentration at the input, c eq is the equilibrium
concentration, and n is the kinetic order. Because the
fracture is being dissolutionally widened, the flow rate
and penetration length increase with time. Both quanti¬
ties are locked in a feedback loop, that is, an increase of
one causes an increase of the other. They accelerate in
time until an abrupt increase of both occurs. The terms
breakthrough and breakthrough time (denoted as T B ) have
been used for the event and the time when it happens.
After the breakthrough, A becomes large compared to
the fracture length L, and the whole fracture widens
uniformly at dissolution rates defined mainly by the sat¬
uration state at the entrance.
Such growth cannot last long, as there is no catch¬
ment large enough to sustain a constant head differ¬
ence between the input and the output. The flow rates
no longer depend on fracture resistance, but on the
amount of water that a catchment can supply to the
input. If the position of the flow path is above the base
level the full pipe flow conditions will change to open
channel flow and a transition to canyon will start.
Figure 2 schematically presents some of the features
discussed above. The evolution of a fracture is divided
into three stages: initiation, breakthrough, and growth.
The curves present the evolution of a 1-km-long frac¬
ture with initial an aperture of 0.02 cm and width 1 m.
Flydraulic head at the input is 50 m higher than at the
x/L
FIGURE 2 Evolution of a single fracture. (A) Flow rates for the
case of limited (black dashed line) and unlimited (full black line)
recharge. Total flow passed through the fracture since t = 0 (dashed
gray line). The ratio between output and input dissolution rates (gray
dashed line with scale on the right axis). (B) Hydraulic head along
the fracture at different times in units of T B (breakthrough time).
spring. Black lines show the evolution of flow rates
when the recharge to the system is unlimited (full
lines) or limited to 1 m 3 /s (dashed). The gray dashed
line shows the total flow that has passed through the
fracture from the beginning to a certain time. Note that
only about 20 L of water passes the fracture until the
breakthrough occurs. The dotted line (scale on the
right side) presents the ratio between dissolution rates
at the output and dissolution rates at the input of the
fracture. This ratio is initially small and asymptotically
approaches 1 at the breakthrough.
The hydraulic head distribution along the evolving
fracture shows interesting dynamics. Initially the head
drop is linear. As the fracture evolves, the high head
from the input is shifted toward the tip of the fracture,
and a high gradient is built up between the tip and the
output (solid lines on Fig. 2B). After the breakthrough
(dashed lines on Fig. 2B), the fracture is widened uni¬
formly, which makes the head drop increasingly uni¬
form. When recharge dominates the evolution, the
head at the input decreases, and when vadose canyon
starts to evolve, the head becomes equal to elevation.
ENCYCLOPEDIA OF CAVES
SPELEOGENESIS, TELOGENETIC
767
The dynamics of head distribution as presented is
important for understanding the evolution of more
complex systems.
The natural conditions are usually far more complex
than those described above. However, by having in
mind a feedback mechanism and the concept of pene¬
tration length, the processes in more complex settings
can be more easily understood.
EVOLUTION OF CONDUIT NETWORKS
Conduit systems in nature are far more complex
than a single flow path. Caves rarely look linearly, and
even if they do, it is likely that most of their complex¬
ity is hidden to explorers. Local geological and
hydrological conditions determine the dynamics of
speleogenesis and the final geometry of the conduit
system. Cavers have mapped thousands of kilometers
of caves around the world, exhibiting a variety of
geometries in various settings. However, a careful look
into cave maps allows us to organize cave patterns
into several categories, which can be related to the
type of recharge and geological structure (Palmer,
2007). In telogenetic settings branchwork and maze pat¬
terns are most common.
To demonstrate simple scenarios that lead to the
formation of one or the other of the two above-men¬
tioned patterns, we apply a numerical model of a 2D
fracture network as shown in Figure 3. The domain
consists of a network of 100 X 100 fractures with initial
apertures distributed statistically between 0.007 cm
and 0.015 cm. Into this network, a coarse network of
10 X 10 fractures, 100 m long with initial apertures of
0.03 cm, is embedded. Only fractures with aperture
widths equal to or larger than 0.03 cm are shown. Four
inputs of calcite aggressive water on the left side are
denoted as 11—14. On the right side are the outputs
denoted by 01—04. The head difference between
input and output is 50 m. The upper and lower bound¬
aries are impermeable. Aperture widths are presented
by the line thickness and flow rates by colors, as
shown on the legend. Flow rates are shown in relative
units Q/ Q max , where Q max is the flow rate in the frac¬
ture with the maximal flow in the net. Line thicknesses
are shown in logarithmic scale. Lines of equal heads
with values in cm are also shown.
During the initiation stage several competing flow
paths have grown from the inputs to the outputs
(Fig. 3A). The flow path from 14 to 03 has evolved most
efficiently. Shortly before the breakthrough, about 80%
of its length is widened considerably. Its wide part
offers low resistance to flow, resulting in only a small
head drop along its length. Almost complete head drop
occurs between the tip of the evolved part and the
outputs. The head distribution in the leading pathway
influences the head distribution in its vicinity. It causes
low gradients in the wider upstream region of the net¬
work and thus prevents the evolution of the competing
pathways emerging from inputs 11—13.
After the breakthrough (Fig. 3B), the flow path
14-03 is widened by uniform rates defined by the solu¬
tion at the input. Consequently the head drop along it
becomes increasingly uniform. The gradients in the
upstream part reestablish so that other flow paths start
growing again. Meanwhile, the new distribution of
hydraulic heads diverts some of the flow from inputs
II to 13 toward the flow path 14—03.
From this point, further evolution of the network can
take diverse routes depending on the recharge.
Figure 3C shows the situation at 2 T B when the recharge
can sustain constant head at the inputs. II first connects
to the output Ol and later to the conduit system evolved
from the input 14, to which 12 also connects. As long as
the gradients in the network are maintained, a system of
bypasses evolves, shortcutting and connecting the prom¬
inent flow paths. This results in a maze pattern
(Fig. 3C). Input 13 is isolated from the prominent net¬
work and cannot develop its own flow path.
In most relevant situations, the catchment cannot
supply enough water to the input. Figure 3D presents
the outcome when the recharge to the inputs is limited
to 100 L/s. After the breakthrough, the head at 14 and
along the flow path 14—03 drops. After 1.2 T B , the head
at 14 falls below 10 m and drops along the flow path
toward 13. This results in high gradient between the
flow paths originating from II to 13, which are still at
high heads, and the flow path 14—03. Therefore, inputs
II and 12 integrate to 14—03. Later on (not shown), 13
also integrates, because of the extreme gradients in its
vicinity. Although some bypasses have evolved, the sce¬
nario clearly leads to a branchwork conduit system.
Although the presented model is simple and just
one case, it clearly shows some basic rules of pattern
development, which can be observed in the field and
in other modeling results:
Maze patterns evolve when the recharge is high
enough to sustain high head in the network. In this
case the maze evolves by the successive growth of
bypasses emerging from the leading pathways. The
whole conduit network thus evolves simultaneously
with high dissolution rates.
When recharge is limited, the head at the input and
along the whole path decreases over time. This
attracts the flow and growth of the neighboring
flow paths emerging from other recharge points.
The branchwork pattern evolves. The branchwork
pattern continues to develop when one or more of
the passages become vadose.
ENCYCLOPEDIA OF CAVES
768
FIGURE 3 Evolution of a 2D fracture network with inputs 11—14 on the left side of the domain and outputs 01—04 on the right side.
(A) Initiation phase, (B) situation shortly after the breakthrough, (C) growth phase (post-breakthrough) with recharge sustaining initial heads
at the inputs, (D) growth phase with limited recharge to the inputs. Aperture width is represented by line thickness. Color presents the flow
rates in units Q/Q max , where Q max is the maximal flow rate in the whole network.
FIGURE 4 Upward evolution of branchwork pattern. See text for discussion.
The recharge conditions in natural systems are not
constant. A branchwork pattern can develop during
low flow and a maze pattern during high flow,
when high head conditions allow expansion of the
network by bypassing. These conditions give rise to
the so-called floodwater mazes (Palmer, 2007).
The mechanisms revealed by this simple model are
not the only ones leading to one or the other pattern.
Maze caves can develop in settings where the input
solutions are close to saturation. In this case the
long penetration lengths, proportional to (1 — c„ 2 /c e? ) 1_ ".
result in slow but simultaneous widening of many
alternative flow paths. Maze caves also develop when
recharge occurs through a thin insoluble layer to many
openings in limestone or directly into soluble rock, as
in epikarst (Palmer, 2007).
However, in most telogenetic cases the maze devel¬
opment is not sustained. If and when a flow path
evolves to the stage when the head along it drops, the
neighboring flow paths (still at high heads) will grow
toward it and a branchwork will develop.
Figure 4 shows a conceptual model where a branch-
work pattern develops upward from the spring. In the
ENCYCLOPEDIA OF CAVES
SPELEOTHEM DEPOSITION
769
example shown, input is distributed through the sink¬
holes on a plain (denoted 1 to 5 in Fig. 4A,B,C). The
flow path connecting Sinkhole 1 evolves first. Sinkhole
2 initially evolves toward the valley, but is later redir¬
ected to the flow path 1—0, when its head becomes
low enough (Fig. 4A). The evolution proceeds upward
from the spring when the head in the evolved flow
paths drops or they become vadose. Sinkholes 3 to 5
connect to one of the two primary flow paths (1—0,
2 - 0 ).
The parameter space that determines speleogenesis
has many dimensions. It seems impossible to embrace
even the most common speleogenetic scenarios in a
short text. Conditions during speleogenesis continu¬
ously change; valleys are down cut and/or filled up;
glaciations and deglaciations, transgressions and
regressions, surface denudation and other processes
continuously alter the evolution of cave systems. The
role of sediments, which is important, has also not
been discussed here. For a comprehensive picture, the
reader is once more referred to the cited textbooks.
Flowever, there are many open questions in speleo¬
genesis, which, when resolved, will surely add many
new aspects to the present state-of-the-art.
See Also the Following Articles
Speleogenesis, Hypogenetic
Solution Caves in Regions of Moderate Relief
Solution Caves in Regions of High Relief
Bibliography
Dreybrodt, W., Gabrovsek, F., & Romanov, D. (2005). Processes of spe¬
leogenesis: A modeling approach. Ljubljana, Slovenia: ZRC Publishing.
Ford, D. C., & Williams, P. (2007). Karst geomorphology and hydrology.
London: John Wiley & Sons.
Klimchouk, A., Ford, D. C., Palmer, A., & Dreybrodt, W. (Eds.),
(2000). Speleogenesis: Evolution of karst aquifers Huntsville, AL:
National Speleological Society.
Palmer, A. N. (2007). Cave geology. Dayton, OH: Cave Books.
SPELEOTHEM DEPOSITION
Wolfgang Dreybrodt
University of Bremen, Germany
INTRODUCTION
Precipitation of calcite from supersaturated H 2 0—
C0 2 —CaCC >3 solutions sculptures fascinating cave
decorations of stalagmites, stalactites, and other spe-
leothems and also creates flowstone and sinter ter¬
races. Rainwater seeping through vegetated soils on its
way down to the limestone bedrock can absorb large
amounts of carbon dioxide, which is present in the soil
in much higher concentrations than in the atmosphere.
When this water moves down, it dissolves limestone
in the soil and in the fissures of the bedrock and comes
very close to saturation with respect to calcite. If such
a solution enters a cave, outgassing of C0 2 creates a
supersaturated solution from which calcite is depos¬
ited. This causes the growth of calcite speleothems in
their various shapes (White, 1988). The chemical com¬
position of this solution depends on the C0 2 content of
the soil atmosphere, but also on the conditions under
which C0 2 is absorbed.
Two extremes can be envisaged. If the soil is free of
calcite, only C0 2 is absorbed there, and dissolution
of limestone later proceeds exclusively in the fractures
of the rock under conditions closed to carbon dioxide.
For each CaCC >3 unit dissolved, one C0 2 molecule is
consumed. When this solution enters a cave, its C0 2
concentration may have become so low that carbon
dioxide is taken up from the cave atmosphere and the
solution becomes aggressive. This can cause the forma¬
tion of Rinnenkarren extending below bedding planes
in many caves. If, however, the partial pressure of car¬
bon dioxide in the soil atmosphere is sufficiently high,
outgassing of C0 2 creates a supersaturated solution pre¬
cipitating calcite. The other extreme occurs when the
soil contains limestone particles, and rainwater seeping
down dissolves limestone under conditions open to the
atmosphere such that for each CaCCh unit dissolved,
one molecule of C0 2 is absorbed from the soil air. If
such a solution reaches equilibrium with respect to cal¬
cite and later on encounters a cave, it will deposit cal¬
cite, provided the Pco 2 of the cave atmosphere is lower
than that in the soil above. Intermediate cases, for exam¬
ple, when dissolution first proceeds in the soil under
open conditions and then continues in the fissures of
the rock under closed conditions, are also possible
(Dreybrodt, 1999). These considerations show that super¬
saturation and consequently growth rates of speleothems
are not simply related to soil Pco 2 , but also in a complex
way to the hydrochemical pathway of the water on its
way down to the cave.
THE CHEMISTRY OF PRECIPITATION
AND DISSOLUTION OF CALCITE
To address the issues of how fast speleothems grow
and which shape they take, one has to know the chemis¬
try of a carbonate solution. Figure 1 illustrates the chem¬
ical reactions in the system H 2 0—C0 2 —CaC0 3 . Gaseous
C0 2 from the air enters into the water and is dissolved
as aqueous C0 2 , which reacts with water to form
Encyclopedia of Caves.
© 2012 Elsevier Inc. All rights reserved.
770
FIGURE 1 Visualization of the chemical reactions and transport
by diffusion during dissolution/precipitation of calcite.
carbonic acid H 2 CO 3 . This is a slow reaction. The follow¬
ing reactions are fast. The complete reaction chain is:
C0 2 tm + H 2 0 $3 H 2 0 + CC> 2 q b H 2 C0 3 b H +
+ HC0 3 b H + + H + + CO 2-
COj q is in equilibrium with the partial pressure,
P- 2 , of CO 2 in the soil air by Henry's law
(C0 2 ) = Kh ■ P("o 2 . The mass action laws control the
concentrations of the other species:
K 0 • (H 2 CO 3 ) = (C0 2 ); K, • (H 2 C0 3 ) = (H + ) • (HCO 3 -);
K 2 • (HCO 3 ) = (H + ) ■ (CO 2- )
The Ks are the corresponding mass action constants.
Parentheses denote concentrations. Furthermore dis¬
sociation of water, H 2 0 <-> H + + OH - with (H + ) ■
(OH - ) = K w , must also be considered.
Ca 2+ ions and CO 2 ions are released from the calcite
surface by three simultaneous reactions:
CaC0 3 + H 2 0 o Ca 2+ + CO 2 + H 2 0; CaC0 3 + H +
<^> Ca 2+ + HC0 3
and
H 2 C0 3 + CaC0 3 + H 2 O^Ca 2+ + 2HC0 3 + H z O
These can be summarized by the global reaction
C0 2 + CaC0 3 + H 2 0 <^> Ca 2+ + 2HC0 3 . Stoichiometry
requires that for each Ca 2+ released from the calcite
surface, one molecule of C0 2 is converted to HC0 3
and H + and is thus removed from the solution. On the
other hand, if Ca 2+ is deposited, one molecule of C0 2
is delivered into the solution. Since calcite dissolution
and precipitation are ubiquitous in nature, this plays
an important role in the global carbon cycle. During
the chemical reactions at the surface all participating
reactants must be transported to the surface or away
from it by molecular diffusion. In this way concentra¬
tion gradients arise and the concentrations at the cal¬
cite surface differ from those in the bulk of the
solution.
The dissolution/precipitation rates F (mol cm -2 s -1 )
at the surface have been found experimentally by
Plummer et al. r 1978. They are given by
F = k, • (H + ) s + k 2 • (H 2 C0 3 ) s + k 3 -k 4 • (Ca 2+ ) s • (HC0 3 - ) s
( 1 )
The suffix s denotes the concentrations at the react¬
ing surface, which are unknown. The rate constants of
dissolution k 4 , k 2 , k 3 depend on temperature only,
whereas the rate constant k 4 , which comprises the
three backward reactions of precipitation, depends
also on the C0 2 concentration in the solution. The state
of saturation of the solution is given by the ratio
O = (Ca 2 + )(C0 3 - )/Kc. K c is the solubility constant.
For Q < 1 the solution is undersaturated and calcite
is dissolved; for fl > 1 the solution is supersaturated
and calcite precipitates. Chemical equilibrium is
attained if 0 = 1. The equilibrium concentration of
Ca 2+ is C eq = (Ki ■ K C • K h /4 • K 2 ■ 7ca • j 2 hco ^ ■
(Dreybrodt, 1988). The q are ionic activity coefficients
and Pg^is the partial pressure of C0 2 in equilibrium
with the solution by Henry's law.
PRECIPITATION RATES FROM THIN
WATER LAYERS
To calculate the rates, one needs the values of the con¬
centrations at the surface. To obtain them from the bulk
concentrations, which can be measured, one has to solve
a system of coupled reaction—diffusion equations. These
must take into account the reaction rates F as a boundary
condition at the surface, the transport of reactants by dif¬
fusion, the slow reaction H 2 0 + C0 2 <-> H 2 C0 3 , and the
condition that in a system open to C0 2 for each calcium
released one molecule of C0 2 must de-gas from the solu¬
tion. Buhmann and Dreybrodt (1985) have solved the
transport reaction equations numerically. In their numer¬
ical model they assumed a planar calcite surface covered
by a stagnant water layer of thickness 8 . The free surface
of this layer is in contact with an atmosphere with partial
pressure, Pq>>, °f carbon dioxide, and equilibrium
between the C0 2 concentration in the water and its sur¬
rounding atmosphere is established. Figure 2 shows the
precipitation rates for various thicknesses of the water
film as a function of the Ca concentration in the solution
at a temperature of T = 10°C and in equilibrium with a
ENCYCLOPEDIA OF CAVES
SPELEOTHEM DEPOSITION
771
FIGURE 2 Precipitation rates of calcite from thin solution films of
various depth: A: 6 = 0.005 cm; B: 6 = 0.0075 cm; C: 6 = 0.01 cm;
D and E: 0.02 cm < 8 < 0.04 cm. The right-hand scale gives growth
rates in mm yr” 1 .
cave atmosphere with Pco 2 = 3 • X 10 -4 atm. All the
curves in Figure 2 can be approximated within an error
of less than 10% by the equation:
F = ol ■ (C — C eq ) (mol cnT 2 s _1 ) (2)
a (cm s _1 ) is a reaction constant; C is the calcium concen¬
tration (mol cm -3 ) in the solution, and C eq its equilibrium
concentration with respect to the partial pressure of
C0 2 in the cave air. Values of a have been obtained
numerically (Buhmann and Dreybrodt, 1985; Baker et al.,
1998) and verified experimentally by Dreybrodt et al.,
1997. The values of a depend on the film depth 8. For
6 < 0.005 cm they decrease with decreasing 6, because
the slow reaction H 2 C 03 <-> H 2 0 + C0 2 becomes rate
controlling. The number of C0 2 molecules forming per
time due to precipitation of calcite is V ■ Rco 2 , where
the rate function Rco 2 depends only on the carbonate and
H + concentrations in the solution with volume V.
Stoichiometry requires this number to be equal to the
number of Ca 2+ ions deposited per time to the surface
area A, which is A-F. This gives F = Rco> -(V/A). For
planar water films V/A = 8 and for sufficiently low
values of 8 the precipitation rates decrease linearly with 8.
By use of the enzyme carbon anhydrase, Rco 2 can be
enhanced. Therefore adding this enzyme to a calcite-
precipitating solution enhances precipitation rates by an
order of magnitude (Dreybrodt et al., 1997). As seen from
Figure 2 the rates are only weakly dependent on 8 for
depths from 0.0075 cm to 0.04 cm. This is the thickness of
films covering stalagmites. It turns out that a is almost
independent of the Pco 2 in the cave air in the range from
3 X 10~ 4 atm up to 5 X 10 3 atm, which covers most of
the natural conditions. Values of a in this range between
0°C and 30°C are reported by Baker et al. (1998). These
can be fitted by the relation
a = (0.52 + 0.04-T c + 0.004 -T 2 )-10~ 7 m/s (3)
where T c is the temperature in centigrade (Romanov
et al., 2008). a increases by a factor of ten with increasing
temperature from 0°C (a = 5.2 X 10 -6 cm s -1 ) to 30°C
(a = 5.3 X10 -5 cm s _1 ). For C = 2 X 10 -b mol cm -3 and
C eq = 0.5 X10 6 mol cm 3 one obtains growth rates
between 0.2 mm yr -1 up to 2 mm yr -1 . Such values have
been observed on many actively growing stalagmites
(Baker et al., 1998; Genty et al., 2001), measuring growth
rates by annual lamina counting, using stalagmites
where drip-water samples were analyzed for Ca. When
comparing theoretically predicted values, they observed
good agreement between theoretical and measured
stalagmite growth rates.
PRECIPITATION FROM WATER FILMS
IN TURBULENT FLOW
Deposition of cave sinter in many cases occurs from
turbulently flowing water. Relatively thin films of a
few millimeters flow down from stalactites or stalag¬
mites when water supply from drip points is suffi¬
ciently high. Supersaturated water flowing down cave
walls precipitates flowstone. Finally cave rivers can
shape sinter terraces and rimstone dams.
When water flow is turbulent, mass transport is
enhanced by eddies in its bulk. This turbulent bulk,
however, is separated from the mineral surface by a
laminar diffusion boundary layer (DBL) with thickness
s, which depends on the hydrodynamic conditions,
such as flow velocity, surface roughness, and depth of
the flowing fluid. In this layer, mass transport pro¬
ceeds by molecular diffusion solely. Because diffu-
sional mass transport increases with decreasing e,
deposition rates are strongly dependent on the hydro¬
dynamics of flow. In all the cases of interest here,
e is about a tenth of a millimeter (Dreybrodt and
Buhmann, 1991).
Rates also depend on the depth 8 of the flowing
water, because 8 = V/A determines the influence of
HC0 3 on the C0 2 reaction. For small 8 < 0.2 cm this
reaction becomes rate limiting. As in laminar flow, the
rates exhibit a linear relation of the Ca concentration as
given by Eq. (2). Values of a have been listed by Liu
and Dreybrodt (1997) for various values of 8 , e, Pco 2 ,
and temperature.
Figure 3 depicts precipitation rates calculated by
use of these values and Eq. (2) for various thicknesses
c and 8 as listed on the curves for a partial pressure
ENCYCLOPEDIA OF CAVES
772
FIGURE 3 Precipitation rates of calcite from thin films of solution in turbulent flow with film depth 6 of 0.1 cm, 1 cm, 10 cm, and 100 cm.
The numbers on the lines depict the values of the thickness e of the DBL.
Pcc >2 = 1 X 10 3 atm of CO 2 in the solution at 10°C.
These conditions apply to many cave rivers, from
which flowstone is deposited.
In all cases one visualizes a strong decrease of the
rates with increasing e. The uppermost curves show
the rates for s = 0. In this case, due to complete mixing
concentrations at the mineral surface equal those in the
bulk. Due to the rate-limiting slow reaction of HCO 3
conversion, the rates increase with increasing 8 . There
is a thickness 8 (10 cm in Fig. 3) above which the rates
become independent of 8 . At that thickness the V/A
ratio is sufficiently large and HC0 3 conversion is no
longer controlling the rates. This thickness increases
with decreasing PCO 2 in the solution. Its value is about
1 cm at 5 X 10 _t atm, 10 cm at 1 X 10 -3 atm, and 100 cm
at 3 X 10 -4 atm. For e = 0 and 8 >10 cm, neither mass
transport nor HCO 3 conversion is rate limiting and
the rates can be obtained from the PWP equation,
Eq. (1). These are the maximal possible rates for a
given chemical composition of the solution.
The values of a depend on the PCO 2 in the solution.
Figure 4 depicts as an example values of a in their
dependence on PCO 2 for 8 = 1 and 10 cm, and e = 0.01
and 0.02 cm. a depends almost linearly on temperature
increasing by about a factor of 3, from 5°C to 20°C.
Details can be found in Liu and Dreybrodt (1997).
In summary, rates from turbulently flowing water
can be significantly higher than those from thin lami¬
nar films. This is the case when flow is highly turbu¬
lent and DBL thickness is less than 0.005 cm. When a
DBL with e *=» 0.01 cm is present rates drop, and for
water films less than 1 mm deep they are close to rates
in laminar or stagnant films of about 0.01 cm, because
in both cases diffusion proceeds through a layer (DBL)
or a laminar film of comparable thickness, and also
HCO3—CO 2 conversion occurs mainly in the DBL.
This has been verified by field observations (Liu et al.,
1995; Bono et al., 2001).
MORPHOLOGY OF REGULAR
STALAGMITES
Franke (1965) was the first who recognized the prin¬
ciple of the growth of ideal regular stalagmites. Fie pro¬
posed the following mechanism assuming growth
conditions constant in time. A drop of solution, super¬
saturated with respect to calcite falling to the ground
spreads out radially when it impinges at the surface.
Precipitation rates are maximal at the center and
decrease with distance outward. This way a thin layer
of precipitated calcite is created. It has a lenticular
shape, thick at the center and thinning out at the edges.
The next layer is built in the same way on the top of
the previous one and a vertical candle-shaped stalag¬
mite grows upward. After it has grown to a height,
which is equal to its diameter its shape on the top
becomes stable and also the diameter of the stalagmite
does not change any more. In other words, the growth
rates in the vertical direction are constant everywhere
and the shape of the stalagmite is shifted vertically in
its growth direction. Figure 5 shows this on a natural
stalagmite, where the vertical distances between two
growth layers are constant everywhere. Dreybrodt (1988,
1999) and Romanov et al. (2008) derived this growth
ENCYCLOPEDIA OF CAVES
SPELEOTHEM DEPOSITION
773
FIGURE 4 Values of the rate constants a in dependence on Pco 2
at 10°C for various values of the thickness 6 of the DBL and the
depth 5 of a solution in turbulent flow. Filled circles are calculated
points. Redrawn after Bono et al., 2001.
principle from three simple generally valid assumptions
by use of numerical modeling, but also by strict mathe¬
matical proof. These are (1) the point of drop impact, the
drip interval, and temperature are constant in time; (2)
growth by deposition of calcite is directed normal to the
actual surface of the stalagmite; (3) growth rates W(l)
(cm yr ') decrease monotonically with distance 1 from
drop impact along the actual surface and become zero at
some large but finite distance 1 0 .
The basic principle of stalagmite growth can be for¬
mulated as follows: under constant growth conditions
in time the top of a stalagmite attains an equilibrium
shape, which during further growth is shifted up verti¬
cally. The part of the stalagmite below is a vertical cyl¬
inder with constant equilibrium radius R eq . This is
illustrated schematically in Figure 6, which shows two
growth surfaces of a stalagmite in equilibrium, which
are separated by some distance Wo-At-Wo (cm yr -1 ) is
the growth rate at the apex and At (year) the time dis¬
tance when the two layers originated. The total volume
V c between these two layers is composed of circular
rings (yellow) with radius x and width Ax, all with
equal height W 0 ■ At. They form a circular disc with
radius R eq and height W 0 ■ At with volume:
V c = W 0 'rcRg q -At (4)
W 0 is the growth rate at the apex in cm yr -1 . The pre¬
cipitation rates F are given by Eq. (2).
They are converted to a growth rate W in cm yr -1
(Dreybrodt, 1988) by
W = 1.17-10 9 q:(c — c eq ) cmyr -1 (5)
FIGURE 5 Longitudinal section through an ideal stalagmite. The
black lines between growth lines, representing earlier surfaces, are of
equal length. The stalagmite has grown by shifting its surface into
the vertical direction. Photo courtesy of D. Genty. Used with permission.
Now the volume Vd deposited to the stalagmite in 1
year (At = 1) must be equal to the volume Vl of the
amount of calcite lost during this time from the water,
which enters with flow rate Q (cm 3 s -1 ) and concentra¬
tion Q n at the apex and leaves the stalagmite with
equilibrium concentration C eq :
V L = ^ • (Cta - C eq ) • 1.168 • 10 9
t 4 (6)
V d =W 0 irR 3 q =1.168 • 10 9 a ■ (C in - C eq ) ■ nR 3 q
V d rop is the volume of the drop and t the drip interval.
Therefore by equating V L = V d one finds
Req y/Vd rop / T O' T (7)
This result relates in a simple way the equilibrium
radius to the kinetic constant a and the drip interval t.
Note that the radius does not depend on the initial
supersaturation (C in — C eq ).
As an example, with a = 1.3 X 10~ 5 cm s -1 , corre¬
sponding to a temperature of 10°C, a drop volume
Vdrop = 0.1 cm 3 , and t = 30 s the equilibrium radius is
Req = 10 cm, a very realistic number.
In this approach one assumes that the water sup¬
plied to the apex spreads out to a water film, which
expands radially, flowing down the stalagmite in lami¬
nar flow. A water film covers the stalagmite with
radial symmetry, as shown in Figure 7. One has to
ENCYCLOPEDIA OF CAVES
774
FIGURE 6 Computer simulation of a regular stalag¬
mite. The red lines show the surface every 200 years
during the initial stage. The black lines depict the equi¬
librium shape. The top four lines are separated by 1000
years. The green lines depict the growth W(l) ■ dt normal
to the actual surface decreasing with distance 1 from the
apex. The new surface after time step dt is generated by
approximating the actual surface by a polygon P!—Pj
and calculating the new positions of the points P7 from
the points Pi by adding the vector ->W(lj)dt. The red
lines between the top two surfaces show vertical growth
at all points equal to growth W(0) ■ dt = W 0 ■ dt at the
apex.
.Q
FIGURE 7 Water impinging on the apex with flow rate Q spreads
out radially forming a water film with depth 8. An annulus of this
film between distance f.\ and ( i+1 is shown. The water supply Q
flows radially from the apex. At distance q along the growth surface
the water enters into the annulus with concentration q and leaves it
at distance f i+1 with concentration c i+ i.
consider an annulus between the points i and i + 1 at
the surface with radial distance Ri and Rj+i, respec¬
tively. The distance between these two points along
the surface is l i+1 — 1, = Af. Mass balance requires that
the amount of calcite deposited to the area of the annu¬
lus between points i and i + 1 must be equal to the loss
of CaCC >3 in the solution when passing over the annu¬
lus. Because the total flow Q = V drop /T enters the
annulus at R, with concentration Q and leaves it at
R i+1 with concentration C i+1/ the mass of calcite lost
per second from the solution covering the surface of
the annulus is
M S = Q(Q-C i+1 ) (8)
M s must be equal to the mass M p of calcite precipi¬
tated per second to the surface S a = 2ir ■ Al; • R ; of the
annulus. Note that M§ and M p in this context are in
mol s _1 .
M p = 2ty • Ah • Ri ’ a • (Q - C eq ) (9)
Equating M p = M s/ one obtains
Q+, = Q - 2-iv • a- (Q - C eq ) • R; ■ Al,/Q (10)
From this the deposition rates W i+1 can be calculated
by subtracting C eq from both sides of the equation and
then multiplying both sides by a
Wi+ a = W ; (l — 2ir-a-Ri'Ah/Q) (11)
Remembering that R eq = ^/V/it err = v/Q/W Eq. (11)
can be rewritten as
W i+1 = Wi • (1 — 2 • Ri • Ah/R^) (12)
ENCYCLOPEDIA OF CAVES
SPELEOTHEM DEPOSITION
775
This recursive equation shows that growth rates
decline with distance 1,. Equation 11, which contains
only defined parameters and the somewhat idealistic
assumption of continuous flow spreading out radially,
can be used for numerical modeling. This has been
performed as FLOW-model by Romanov et al. (2008).
They employed the following procedure: the actual
surface of the stalagmite at time t is approximated by
the polygon Pj to P n , where Pi is the point of drip
impact. The surface at later time t + At is obtained by
drawing lines with lengths W(l ; ) ■ At from points Pi
normally to the actual surface (green lines in Fig. 6).
This way the three basic assumptions on stalagmite
growth are fulfilled and the new surface is approxi¬
mated by the polygon Pi to P„, as shown in Figure 6.
The results are also shown in Figure 6. It depicts
growth layers of a stalagmite growing on a plane sur¬
face. Note that the x and y axes have different scales.
Therefore the picture presents a shape compressed in
height. Clearly the shape of the stalagmite changes in
its initial stage of growth (red lines) until an equilib¬
rium shape is attained (black lines). From then on this
shape is maintained during further growth and the
vertical distance between any two equilibrium growth
layers remains constant on all points of the surface.
The vertical red lines, all of equal lengths, between the
two upper surfaces indicate this. The equilibrium
shape is attained after a growth height close to its
diameter. Furthermore it is independent of the shape
of the surface from where it started to grow
(Dreybrodt, 1988).
A SIMILARITY RULE OF REGULAR
STALAGMITES
An important rule can be obtained from Eq. (12). By
introducing dimensionless coordinates x = x/R eq and
y = y/Req one finds
W i+1 =Wi-(l-2-Ri- Ah) (16)
This, however, means that any regular stalagmite
scaled by the factor 1 /R eq into this new coordinate sys¬
tem must exhibit a unique shape independent of its
former dimensions. This proves the similarity rule:
scale the stalagmites in both coordinates by the factor
1/R eq and choose their growth axis as y-coordinate
and their apex as origin. Then all regular stalagmites
have the same shape.
To prove this by observation one has to select
regular stalagmites, which show vertical walls about
1 diameter below their apex. An example is shown in
Figure 5. Seven longitudinal sections of stalagmites
exhibited in the Deutsche Archiv fur Sinterchronologie
(DASC) in the Schillat-cave, Germany, are used.
FIGURE 8 Rescaled shapes of all stalagmites compared to the
shape obtained from the FLOW-model (fat red line) and a modified
FLOW-model (fat black line). The numbers give the diameters of the
stalagmites.
Furthermore four stalagmites have been taken from
the literature. For details see Dreybrodt and Romanov
(2008). They have also performed a search of regular
large stalagmites on the Internet. The largest stalagmite
in the world is reported from Cueva San Martin
Infierno in Cuba. This giant stalagmite exhibits a
height of about 70 m and 20-m diameter (www.good-
earthgraphics.com/virtcave/largest.htm). Other ones
have been found in Carlsbad Caverns, New Mexico,
U.S.A., with a diameter of 4 m (www.fingerlakesbmw
.org/visual/az/cccolumnl.jpg), and finally from Lost
World Caverns, West Virginia, U.S.A., the Bridal
Veil with 2-m diameter (www.wonderfulwv.com/
archives/juneOl/feal.cfm). Suitable growth layers, or
the surface of those stalagmites, have been digitized.
To compare the shapes all stalagmites were rescaled to
the same radius R eq = 1 • Figure 8 shows the plots of
the complete selection of samples and compares them
to the theoretically obtained rescaled shape of the
FLOW-model depicted in Figure 7. Although not per¬
fect, the theoretical shape (fat red line) is quite close to
the natural shapes. The deviation could be explained
in the following way: whenever a drop impinges to
the apex of a stalagmite, splashing occurs in a water
layer covering the central part of the stalagmite's sur¬
face. This acts like a reservoir from which outward
flow is supplied. In this layer due to mixing, when a
new drop impinges, a constant calcium concentration
will be maintained and the growth rate W(l) is con¬
stant within the radius. To model the shape of this sta¬
lagmite requires Wj = Wo for all points i with Rj < R c .
For Rj > R c = 0.33 • R e q the recursive equation, Eq. (11),
is valid. The shape obtained this way is closer to
nature and is shown by the fat black line.
Also ideally shaped cone-like stalactites obey an anal¬
ogous similarity rule (Short et al., 2005). These authors
ENCYCLOPEDIA OF CAVES
776
s
have modeled growth of stalactites, considering a thin
film of water flowing evenly down the stalactite. Because
of the decreasing diameter of the stalactite, its depth
increases downward. As stated already, for very thin
water films precipitation rates increase with film thick¬
ness. Consequently growth rates increase downwards
the stalactite. Using this their model predicts similar
shapes for all ideal stalactites. It has been verified by
comparing the rescaled shapes of stalactites of various
sizes in Kartchner Caverns (Benson, AZ, U.S.A.) to the
predicted shape.
GROWTH OF STALAGMITES UNDER
CONDITIONS VARIABLE IN TIME
Only a few stalagmites grow in their ideal shapes
under conditions constant in time. Changes in temperature
affect the rate constant a and may also influence mete¬
oric precipitation, thus altering the drip intervals.
In many caves one finds cone-shaped stalagmites,
which form when the equilibrium radius Req(t) =
^/V/TTT(t)a(t) decreases in time. Club-like stalagmites
grow when R eq (t) increases in time. From this, one
understands why, in spite of the simple growth mech¬
anism, stalagmites exhibit such a variety of shapes.
As an example, one assumes a periodic change of
drip intervals t as shown in Figure 9B. The temperature
remains constant. The shape of a stalagmite growing
under such conditions can be modeled by adjusting in
each time step a(t) and T(t). The result is shown in
Figure 9A. The stalagmite grows with periodic change
of its diameter. To illustrate which effect climatic
changes could imprint into the shape of a stalagmite,
one can use the temperature variations from the Vostok
ice core rescaled to moderate climates. Additionally one
assumes a linear relation between temperature and
meteoric precipitation, which cause low drip intervals
at warm climatic conditions and higher ones for cooler
climate. This is illustrated in Figure 9D. Furthermore
Pcc >2 is low at cold climates and because of stronger
vegetation it becomes high at warmer temperatures.
The resulting stalagmite is shown in Figure 9C. The
growth lines are drawn every 1000 years. At high tem¬
peratures t is small and a is moderately large. For high
Pcc >2 the equilibrium concentration C eq is high, causing
large growth rates. Large diameters and wide distances
between the growth lines reflect this. The regions where
growth is slow and growth lines become dense (black
parts) exhibit small diameters because drip intervals are
long. This rather crude simulation shows as an example
how climatic changes might affect the shapes of stalag¬
mites. More details can be found in Kaufmann (2003).
Recently Tan et al. (2006) have measured the width
of annual lamina, as they occur in many stalagmites,
on 2000 lamina in a stalagmite of China. They suggest
that the time series of growth rates obtained this way
may be used as a proxy of paleoclimate. The theoretical
FIGURE 9 Model of stalagmite growth under
conditions varying in time. (A) and (B) Growth
at constant temperature; drip rate varies periodi¬
cally. a = 1.3 ■ 10~ 5 cm s -1 . (C) and (D) Growth
under climatic conditions (temperatures) as
obtained from the Vostok ice core.
ENCYCLOPEDIA OF CAVES
SPELEOTHEMS: GENERAL OVERVIEW
777
and experimental knowledge on calcite precipitation is
an important tool to decipher such proxies.
Bibliography
Baker, A., Genty, D., Dreybrodt, W., Barnes., W. L., Mockler, N. J., &
Grapes, J. (1998). Testing theoretically predicted stalagmite
growth rate with recent annually laminated samples: Implications
for past stalagmite deposition. Geochimica et Cosmochimica Acta,
62, 393-404.
Bono, P., Dreybrodt, W., Ercole, S., Percopo, C., & Vosbeck, K. (2001).
Inorganic calcite precipitation in Tartare karstic spring (Lazio,
central Italy): Field measurements and theoretical prediction on
depositional rates. Environmental Geology, 41, 305—313.
Buhmann, D., & Dreybrodt, W. (1985). The kinetics of calcite dissolu¬
tion and precipitation in geologically relevant situations of karst
areas: 1. Open system. Chemical Geology, 48, 189—211.
Dreybrodt, W. (1988). Processes in karst systems. Physics, chemistry and
geology. Berlin: Springer.
Dreybrodt, W. (1999). Chemical kinetics, speleothem growth and cli¬
mate. Boreas, 28, 347—356.
Dreybrodt, W., & Buhmann, D. (1991). A mass transfer model for
dissolution and precipitation of calcite from solutions in turbulent
motion. Chemical Geology, 90, 107—122.
Dreybrodt, W., Eisenlohr, L., Madry, B., & Ringer, S. (1997). Precipitation
kinetics of calcite in the system CaC0 3 —H 2 0—C0 2 : The conversion
to C0 2 by the slow process H + + HC0 3 ~ -> CO 2 + H 2 0 as a rate lim¬
iting step. Geochimica et Cosmochimica Acta, 61(18), 3897—3904.
Dreybrodt, W., & Romanov, D. (2008). Regular stalagmites: The the¬
ory behind their shape. Acta Carsologica, 37, 175—184.
Franke, H. W. (1965). The theory behind stalagmite shapes. Studies in
Speleology, 1, 89—95.
Genty, D., Baker, A., & Vokal, B. (2001). Intra- and inter-annual
growth rate of modern stalagmites. Chemical Geology, 176,
191-212.
Kaufmann, G. (2003). Stalagmite growth and paleoclimate: The
numerical perspective. Earth and Planetary Science Letters, 214,
251-266.
Liu, Z., & Dreybrodt, W. (1997). Dissolution kinetics of calcium car¬
bonate minerals in H 2 0—C0 2 solutions in turbulent flow: The
role of the diffusion boundary layer and the slow reaction
H 2 0 + C0 2 ->H + + HCO 3 . Geochimica et Cosmochimica Acta, 61,
2879-2889.
Liu, Z., Svensson, U., Dreybrodt, W., Yuan, D., & Buhmann, D. (1995).
Hydrodynamic control of inorganic calcite precipitation in
Huanglong Ravine, China: Field measurements and theoretical pre¬
diction of deposition rates. Geochimica et Cosmochimica Acta, 59(15),
3087-3097.
Plummer, L. N., Wigley, T. M. L., & Parkhurst, D. L. (1978).
The kinetics of calcite dissolution in C0 2 -water systems at 5°C to
60°C and 0.0 to 1.0 atm C0 2 . American Journal of Science, 278,
537-573.
Romanov, D., Kaufmann, G., & Dreybrodt, W. (2008). Modeling
stalagmite growth by first principles of chemistry and physics
of calcite precipitation. Geochimica et Cosmochimica Acta, 72,
423-437.
Short, M. B., Baygents, J. C., Beck, J. W., Stone, D. A., Toomey, R. S., &
Goldstein, R. E. (2005). Stalactite growth as a free-boundary prob¬
lem: A geometric law and its platonic ideal. Physical Review Letters,
94, 018501.
Tan, M., Baker, A., Genty, D., Smith, C., Esper, J., & Cai, B. (2006).
Applications of stalagmite laminae to paleoclimate reconstructions:
Comparison with dendrochronology/climatology. Quaternary/ Science
Reviews, 25(17—18), 2103—2117.
White, W. B. (1988). Geomorphology and hydrology of karst terrains.
New York: Oxford University Press.
SPELEOTHEMS: GENERAL OVERVIEW
William B. White
The Pennsylvania State University
Speleothems are secondary deposits in caves. Speleothem
is a made-up word. The term was invented in 1952
to replace the obsolete term cave formations, which could
be easily confused with bedrock formations. Speleothems
are the primary attraction in show caves and provide
much of the esthetic attraction for cave explorers. This
article is intended as a broad overview of speleothems.
More detail is provided in other articles on specific
speleothem groups. For a comprehensive discussion
of cave minerals and speleothems see Hill and Forti
(1997).
SPELEOTHEM MINERALS
Although speleothems can be composed of many
minerals, in fact the vast majority of speleothems is
composed of one (or a mixture) of only three minerals:
calcite, CaC0 3 ; aragonite, also CaC0 3 ; and gypsum,
CaSC >4 ■ 2H 2 0. Of these, calcite is, by far, the most com¬
mon. Each of these minerals has a different crystal
structure and also different habits of crystal growth.
Calcite is trigonal, space group R3c. The structure
consists of alternating layers of Ca 2+ ions and CO 2
ions stacked perpendicular to the c-crystallographic
axis. Calcite occurs in nature in a great variety of forms
but the most common in caves is the scalenohedron.
The c-axis is the fast growth direction in calcite and this
is responsible for much of the observed microstructure
in speleothems.
Aragonite is orthorhombic, space group Pmcn.
Although it has the same chemical composition as
calcite, the calcium and carbonate ions are more densely
packed in aragonite, giving the mineral a somewhat
higher density. Crystal growth in the [001] direction is
much faster than other directions, giving many arago¬
nite speleothems an acicular habit.
Gypsum is monoclinic, space group A2/a (or 12/a).
Layers composed of Ca 2+ ions and SO 4 ions are sepa¬
rated by a layer of H 2 0 molecules so that these layers
are held together only by hydrogen bonds (Fig. 1). The
result is that gypsum has a perfect cleavage on [ 010 ].
Large crystals can be easily cleaved into transparent
slabs, a form of gypsum known as selenite. The stron¬
gest bonding is in the [ 001 ] direction which is also the
fast crystal growth direction. Gypsum grows readily as
[ 001 ] fibers, a habit that produces the form of gypsum
Encyclopedia of Caves.
© 2012 Elsevier Inc. All rights reserved.
778
STRUCTURE OF GYPSUM
{ 010 }
[ 001 ]
I Fast growth
* direction
FIGURE 1 The crystal structure of gypsum, CaSC >4 ■ 2H 2 0, showing the direction of perfect cleavage and also the fast growth (fiber)
direction.
known as satin spar and is responsible for the widely
occurring gypsum flowers in caves.
MECHANISMS FOR SPELEOTHEM
DEPOSITION
The essential condition for the deposition of any spe-
leothem is that water bearing mineral matter in solution
enters the cave out of chemical equilibrium with the
cave environment. Some reaction will then take place in
which the water chemistry strives toward equilibrium
with the concurrent deposition of minerals.
Carbonate minerals, both calcite and aragonite,
are formed when water, supersaturated with respect to
these minerals is exposed to the cave atmosphere.
Ca 2+ + 2HC0 3 <-*• CaC 03 + C0 2 + H z O
If the C0 2 pressure of the aqueous solution is less
than the C0 2 pressure of the cave atmosphere, C0 2 will
de-gas and calcite (or aragonite) will precipitate. This is
the overall mechanism. The details are much more com¬
plicated with multiple subreactions depending on tem¬
perature and other ions present in the water. However,
the most important variable is the C0 2 pressure in the
infiltrating water compared with the background C0 2
pressure of the cave atmosphere. Calcite and aragonite
can also be precipitated by evaporation of the solutions.
Evaporation becomes important in well-ventilated caves
and near cave entrances. In general, evaporation pro¬
duces very small carbonate crystals; the largest and
best-developed crystals are found in water-saturated
caves where C0 2 loss is the only mechanism.
In contrast, the sulfate minerals, gypsum, and also
mirabilite, epsomite, and others, are deposited mainly
by evaporating water. Dissolved sulfate ions recombine
with Ca 2+ , Mg 2+ , Na + , and other cations as evaporation
drives the solution to supersaturation.
SPELEOTHEM SHAPES
The shapes of speleothems are the result of a com¬
petition between the hydraulics of the fluids carrying
the nutrient to the growing speleothem and the ten¬
dency of the crystals to take on their natural habits.
Because of the great variety of combinations of flow
regimes and growth habits there are a very large
number of speleothem shapes. These shapes have been
given a great variety of names and there have been
numerous attempts at classifications, none of them
entirely satisfactory. Some classifications are very elab¬
orate, attempting to provide a slot for every possible
speleothem. Others consist only of an alphabetical
ENCYCLOPEDIA OF CAVES
SPELEOTHEMS: GENERAL OVERVIEW
779
TABLE 1 A Classification of Speleothems
A. Forms created by flowing water
1. Flowstone
2. Rimstone dams
3. Botryoidal forms (globulites, cave popcorn, cave coral)
B. Forms created by dripping water
1. Stalactites
2 . Stalagmites
3. Draperies
4. Shields
C. Forms created by seeping water
1 . Crusts
2. Fielictites
3. Oulopholites (gypsum flowers)
4. Anthodites/frostwork
D. Forms created in standing water
1. Pool spar
2 . Concretions (including cave pearls)
listing of the names. The classification given in Table 1
is a compromise based on the flow characteristics of
the fluids that deposit the speleothems.
Speleothems Created by Flowing Water
Flowing water in caves may occur as streams flowing
along the cave floor and it may take the form of thin
sheets of water flowing down cave walls. Of these, thin
sheets of flowing water, draining through the vadose
zone from the epikarst, contain higher concentrations
of dissolved carbonates, and are at higher CO 2 pres¬
sures. Thin sheets of flowing water are responsible
for a great deal of speleothem deposition. In contrast,
caves streams are usually undersaturated with respect
to carbonates and speleothem deposition on stream
beds is relatively uncommon although there are
exceptions (Fig. 2).
CO 2 degassing from moisture films on cave walls
produces sheets of calcite known as flowstone (Fig. 3).
Flowstone consists of superimposed layers of calcite
crystals which typically grow with their c-axes perpen¬
dicular to the sheet.
Cave streams near chemical saturation have a
remarkable ability to pond themselves in the form
of rimstone dams (Fig. 4). The mechanism for the forma¬
tion of dams is not known with certainty and it is pos¬
sible that microorganisms play a role. They are found
in small and large flowing streams where the chemis¬
try is correct and are also found in surface rivers of
which the Krka River and the Plitvice Lakes in Croatia
are outstanding examples.
More enigmatic speleothems are the nodular or
bead-like forms variously known as cave beads, cave
coral, cave grape, cave popcorn, or globulites. The
speleothems are small, typically a few millimeters to a
FIGURE 2 Flowstone formed on the bed of a small in-feeder
stream. Marvel Cave, Missouri. Photo by the author.
FIGURE 3 Flowstone in Furong Dong, Chongqing Province,
China. The height of the flowstone is about 10 meters. The colors are
due to the show cave lighting. Photo by the author.
few centimeters, and grow from cave walls and from
other speleothems. Some consist entirely of calcite and
form on rough cave walls that have been etched by
descending streams of water (Fig. 5) or on the sides
of stalactites (Fig. 6). These are layered structures with
the growth originating from small projections on the
cave wall (Fig. 7). Some nodular speleothems contain
aragonite and other minerals and are not as obviously
associated with flowing water.
Speleothems Created by Dripping Water
Water infiltrating downward from the epikarst
follows available pathways along fractures until it
emerges from the ceiling of an underlying cave. If the
seepage rates are low, drops form and hang for a
period of time during which CO 2 is degassed from the
drop, thus supersaturating it with respect to calcite.
ENCYCLOPEDIA OF CAVES
780
FIGURE 4 Rimstone dams formed in a shallow pool, a form
sometimes called "lily pads." Photo by the author.
FIGURE 6 Nodular speleothems on the sides of stalactites in
Swago Pit, West Virginia. Photo by the author.
FIGURE 5 Cave beads formed on a bare bedrock wall in
Harlansburg Cave, Pennsylvania. The top scale is in inches; the bot¬
tom scale in centimeters.
The calcite deposits as a thin ring on the surface from
which the drop is suspended. Eventually, the drop
falls, leaving the ring of calcite behind and a new drop
forms. This process is repeated and the deposit takes
the form of a straw-shaped stalactite (Fig. 8). Straw
stalactites are hollow tubes, about 5 mm in diameter,
and are often very coarsely crystalline with the straw
axis corresponding to the c-axis direction of the calcite
crystals. Stalactites continue to grow laterally and
FIGURE 7 Cross-section of nodular speleothem showing banded
structure. The diameter is about 1 cm. Photo by the author.
FIGURE 8 Straw stalactites and other speleothems in Whisper
Rocks (Lincoln Caverns), Huntingdon County, Pennsylvania. Note
the fracture that provides the pathway for the seeping vadose water.
Photo courtesy of Lincoln Caverns. Used with permission.
ENCYCLOPEDIA OF CAVES
SPELEOTHEMS: GENERAL OVERVIEW
781
vertically from water flowing down the outside as well
as through the central canal (Fig. 9). Stalactites can
grow very large and indeed reach a weight that
exceeds the tensile strength of their attachment and
they spontaneously break loose and fall. Cross-sections
of stalactites typically consist of concentric rings of cal-
cite around the central canal.
Water drops from cave ceilings or from stalactite tips
usually fall before depositing their entire load of calcite.
Additional deposition takes place when these falling
drops hit the floor, thus forming upward-growing stalag¬
mites. Stalagmites grow as a sequence of superimposed
layers. There is no central canal. Stalagmites growing
beneath a single drip site with a uniform drip rate and
uniform chemistry take the form of vertical cylinders
with varying diameters depending on the drip rate
and drip water chemistry (Fig. 10). Because stalagmites
rest on cave floors, they can grow to sizes limited only
by the size of the cave passage.
Stalactites and their counterpart stalagmites eventu¬
ally grow together to form columns. Multiple drip sites
and shifting drip sites combine to generate complex
speleothems that can become very massive (Fig. 11).
Water dripping along fractures can produce a stalactite
form known as the drapery, an intermediate form
between flowstone and an isolated stalactite.
Related to dripstone and flowstone forms are the
relatively rare speleothems known as shields (Fig. 12).
Shields appear to be formed by waters moving along
fractures under pressure so that shields can form at
any angle to the vertical. Shields contain a medial
crack rather than a medial canal. Shields often have a
fringe of stalactites formed by water seeping from the
medial crack and dripping off the edges. Because
shields require unusual hydraulic conditions, they are
found on only a few caves such as Grand Caverns,
Virginia, and Lehman Caves, Nevada.
FIGURE 9 Carrot-shaped stalactites grown by enlargement of
straw stalactites in Ohio Caverns, near West Liberty, Ohio. Photo by
the author.
FIGURE 10 Stalagmites in Postojna Cave, Slovenia. Photo by the
author.
FIGURE 11 Massive stalactite-stalagmite combination. Buckhill
Cave, Virginia. Photo by the author.
ENCYCLOPEDIA OF CAVES
782
s
FIGURE 12 Shields located in Grand Caverns, Virginia. Note that
the medial planes are at different oblique angles to the vertical. Photo
by the author.
Speleothems Created by Seeping Water
The carbonate and gypsum rocks in which caves
form tend to have a very low permeability so that
most of the groundwater movement is through joints
and bedding plane partings. However, low permeabil¬
ity is not zero permeability and there can be seepage
through the bulk rock in addition to seepage along
fractures. Fluids seeping through pores in the rock
carry mineral matter which is deposited as crusts or
more fully developed speleothems when the solutions
evaporate into the cave passage. The high supersatura¬
tions achieved by these evaporating solutions often
result in the deposition of aragonite rather than calcite
and thin crusts are revealed under the microscope
to be composed of mats of very small acicular crystals
of aragonite. The growth of larger aragonite crystals
in the form of frostwork or anthodites is not well under¬
stood but they appear to be formed from seeping
fluids. These are masses of aragonite fibers or radiating
sprays of aragonite crystals that may reach lengths of
many centimeters.
Helictites are an erratic stalactite-like form that occurs
when seepage rates are too slow for complete drops to
form. Instead, the speleothem follows the whims of cal¬
cite growth (Fig. 13). Helictites have a central canal and
grow primarily from the tip.
Sulfate-bearing solutions seeping from pores in the
bedrock deposit crusts of gypsum and other sulfate
minerals. Continued seepage produces additional crys¬
tal growth. If there is a zone of higher permeability,
gypsum growth takes the form of fibrous crystals
which push outward. Because the fastest growth rate
is at the center, the fibers curl to the outside producing
the gypsum flower (Fig. 14). "Flowers" can also form
FIGURE 14 Gypsum flower from Martin Ridge Cave System,
Kentucky. Photo by Joe Kearns. Used with permission.
from other sulfate minerals such as mirabilite,
Na 2 S0 4 ■ 10H 2 O, and epsomite, MgS0 4 ■ 7H 2 0.
Speleothems Created in Standing Water
There is a marked contrast between speleothems
that grow at the interface between air and a water
surface and speleothems that grow entirely immersed
in water. Subaerial speleothems have smooth rounded
surfaces dictated by the characteristics of the flowing
water. Subaqueous speleothems exhibit the growth
habits of their component minerals.
Cave pearls are near-spherical, unattached speleothems
that form in pools agitated by water dripping from con¬
siderable heights (Fig. 15). They form in concentric layers
around a nucleus which can be a sand grain or even a
bat bone. Drip pools in former show caves sometimes
contain cave-pearl—like objects formed as calcite coatings
built up over bits of gravel that had been used in the
ENCYCLOPEDIA OF CAVES
SPELEOTHEMS: GENERAL OVERVIEW
783
FIGURE 15 A pool with cave pearls. Barberry Cave, Bath
County, Virginia. Photo by Nevin W. Davis. Used with permission.
FIGURE 16 Pool spar from Helictite Cave, Highland County,
Virginia. Photo by Philip C. Lucas. Used with permission.
tourist trail. Sufficient agitation from dripping water
is needed to keep the pearls from becoming attached
and to roll them into roughly spherical forms. There are
a few rare examples of cave pearls in the form of cubes
tightly nested in the drip pool.
Standing water in cave pools can become super¬
saturated by the gradual degassing of CO 2 from the
pool surface. Calcite crystals may precipitate on the
surface of the pool and aggregate to form calcite rafts.
These may be suspended by the surface tension of
the water until they finally accumulate sufficient mass
to sink. They collect on the bottoms of pools as loosely
aggregated plates.
Crystals forming in pools are unconstrained and can
adopt the natural habit of the precipitating mineral.
Calcite is the most common and takes the form of pool
spar. Cave pools are frequently crystal-lined, often with
a coarse mass of small crystals but sometimes with near
perfect scalenohedra (Fig. 16).
FIGURE 17 Solubility curves for calcite and aragonite as a func¬
tion of C0 2 partial pressure. Calculated from thermodynamic data.
SPELEOTHEM PROPERTIES
The Calcite/Aragonite Problem
Calcite is the thermodynamically stable form of
CaCC >3 at atmospheric pressure over the entire tem¬
perature range from absolute zero to the decomposi¬
tion temperature. The thermodynamic stabilization of
aragonite at ambient temperature requires pressures
in excess of 3000 atmospheres. But oblivious to ther¬
modynamics, aragonite is a common mineral in caves
where it is precipitated from seepage and drip waters
at cave temperature and atmospheric pressure. How
to explain this observation has occupied many researchers
for many years and is known as the "calcite/aragonite
problem."
Aragonite is more soluble than calcite at any given
temperature and CO 2 pressure but only 10—15% more
soluble (Fig. 17). What is needed are mechanisms that
will either inhibit the precipitation of calcite so that
the concentration of ions in solution can increase
past the aragonite solubility curve or that will enhance
the nucleation of aragonite which can then grow meta¬
stability in competition with calcite. Two important
mechanisms that have been identified involve the
common minor elements in karst waters: Mg 2-1 " and Sr 2+ .
Mg 2+ ions have been shown to poison the growth cen¬
ters on calcite crystal, thus inhibiting calcite growth.
SrCC >3 has the same crystal structure as aragonite. Early
precipitation of SrCC >3 provides a nucleus on which
aragonite can grow. There may be other mechanisms
and other contributing ions.
ENCYCLOPEDIA OF CAVES
784
s
In many caves, the carbonate speleothems consist
entirely of calcite. But when aragonite speleothems
are found, there are usually many of them. This
suggests that the presence (or absence) of aragonite
is controlled by local chemical composition of the
bedrock or chemical composition of the vadose seep¬
age water. Aragonite is often identified by its crystal
habit since it has a tendency to grow as tufts of
radiating acicular crystal. Flowstone and dripstone
forms of aragonite do occur but are uncommon
(Fig. 18).
FIGURE 18 Massive column of aragonite in Carroll Cave,
Missouri. Note color contrast between the white aragonite and the
yellow and brown calcite speleothems. Photo by the author.
Color and Luminescence of Speleothems
Chemically pure calcium carbonate and calcium sul¬
fate are colorless. Most calcite speleothems are colored
various shades of tan, orange, and brown. Although a
few speleothems are pigmented by incorporated iron
oxides and hydroxides, most speleothem colors arise
from humic substances extracted from the overlying
soil and incorporated in the speleothems. The large
and complex humic and fulvic acid molecules seem to
fit into the calcite structure but not the aragonite struc¬
ture. Aragonite and gypsum speleothems tend to be
colorless (Fig. 18).
Calcite speleothems are usually luminescent under
excitation by ultraviolet light. The emission occurs
as a broad band in the range of 400—450 nm,
giving the luminesence a blue-white color (Fig. 19A).
Occasionally, a more yellow emission is observed
on organic-rich speleothems (Fig. 19B). The lumines¬
cence, like the color, is due to humic substances.
The fulvic acid fraction produces luminescence; the
humic acid fraction is responsible for the color. For
this reason, nearly white speleothems tend to have
the strongest luminescence.
SPELEOTHEMS IN VOLCANIC CAVES
In solution caves, those in carbonate rocks and gyp¬
sum, the deposition of speleothems is the reverse
of the process that formed the caves. Speleothem depo¬
sition in volcanic caves is a more complicated sequence
of events (Forti, 2005). Lava tube caves result from the
draining of tubular flows of lava beneath the surface.
Such caves form very rapidly and many of the explored
lava tube caves are only years to centuries old. The
speleothems that form in lava tube caves develop in a
sequence as the tube drains, cools, and becomes subject
to rainwater infiltration.
In the final stages of the draining of the tube, still-
molten lava can drip from the tube ceiling and then
FIGURE 19 (A) Typical blue-white luminescence from a segment of stalactite. (B) Oily yellow luminescence from a stalactite fragment
from Cueva del Guacharo, Venezuela. Photos by the author.
ENCYCLOPEDIA OF CAVES
SPELEOTHEMS: GENERAL OVERVIEW
785
freeze to produce lava stalactites. Lava stalagmites
and other forms can also develop. Some have ques¬
tioned whether these lava forms should be called
speleothems. Purists have argued that speleothems
are secondary deposits and that lava stalactites are
primary features.
As cooling progresses, the still hot lava can be
impacted by water either from rainfall or—in coastal
areas—by saltwater from wave action or storm surges.
The tubes and associated fractures can be infused with
superheated steam which can extract and transport
heavy metals from the lava. The resulting mineral
deposits take the form of coatings and small nodules
and stalactites sprayed on the tube walls as brightly
colored silicates of copper, vanadium, and other
metals. The chloride content of seawater is particularly
effective in mobilizing metals. Elemental sulfur is also
found.
The most dramatic speleothems form in tubes only
a few years old that have cooled enough to permit
some circulation of liquid water. The few such caves
that have been examined ( e.g ., Porter, 2000) contain
suites of complex sulfate minerals, many of which are
highly hydrated and water soluble. The sequence of
mineralization is highly transitory. Further cooling of
the tube and infiltration of rainwater quickly removes
the water-soluble minerals.
The final stage is a lava tube at ambient temperature
with the same dripping water found in other caves.
The resulting speleothems are also very similar to those
found in solution caves. Gypsum is common as crusts
and nodules. Calcite occurs as nodules and rarely as
dripstone.
SPELEOTHEMS AS PALEOCLIMATE
ARCHIVES
Although speleothems and the processes that form
them are interesting in their own right, the most
important speleothem study is their role as archives
of paleoclimate (Fairchild et al., 2006). Stalagmites,
in particular narrow cylindrical stalagmites, are built
up layer by layer and can be regarded as a drill core
available without the necessity of drilling. A longitu¬
dinal section through such a stalagmite provides a
microstratigraphy that can span thousands to tens
of thousands of years of time. U-Th dating provides
an absolute chronology than can, in principle, extend
back to the mid-Pleistocene.
Profiles can be measured along the stalagmite axis
of oxygen isotope ratios, H/D ratios, carbon isotope
ratios, strontium isotope ratios, and the concentration
profiles of chosen trace elements such as Mg, Sr,
and any other trace elements that might be climate
FIGURE 20 Sequences of speleothem deposition in Ruddle Cave,
Pendleton County, West Virginia. Note first-deposited dark brown
speleothems that were coated by a later flooding event, followed by
more recent deposition of the orange stalagmite. Photo by the author.
signatures. Likewise, the color banding, luminescence
banding, and crystal growth textures can be measured.
Because the water that deposits speleothems begins its
journey as rainfall which then passes through the soil,
the epikarst, and the vadose zone to reach the site
of deposition in the cave, it should carry a signature
of the surface climate. The climate signatures at each
time interval are stored in the growth layers of the sta¬
lagmite. There are certainly signals present and much
current research is focused on the interpretation of
these signals.
Because paleoclimate studies require removing sta¬
lagmites from caves and making measurements in the
laboratory, attention has focused on climate changes
during the closing phases of the last ice age and on cli¬
mate changes during the Holocene. What awaits are
investigations of entire caves and of larger and older
speleothems. Consider Figure 20. There was clearly an
earlier phase of speleothem deposition, following by a
period of flooding that coated the speleothems with
silt and clay, and finally the most recent stage when
the clear orange stalagmite was deposited. The size of
the final stalagmite suggests that the previous events
are quite old. There are many cave sites similar to
Figure 20, just waiting for the investigators.
See Also the Following Articles
Gypsum Flowers and Related Speleothems
Helictites and Related Speleothems
Paleoclimate Records from Speleothems
Speleothem Deposition
Stalactites and Stalagmites
Uranium Series Dating of Speleothems
ENCYCLOPEDIA OF CAVES
786
s
Bibliography
Fairchild, I. J., Smith, C. L., Baker, A., Fuller, L., Spotl, C., Mattey, D.,
et al. (2006). Modification and preservation of environmental
signals in speleothems. Earth Science Reviews, 75, 105—153.
Forti, P. (2005). Genetic processes of cave minerals in volcanic envir¬
onments: An overview. Journal of Cave and Karst Studies, 67, 3—13.
Hill, C., & Forti, P. (1997). Cave minerals of the world. Huntsville, AL:
National Speleological Society.
Porter, A. (2000). The initial exploration of Lower Lae'apuki Cave
System, Hawai'i Volcanoes National Park. National Speleological
Society News, 58, 10—17.
SPIDERS AND RELATED GROUPS
James R. Reddell
The University of Texas at Austin
INTRODUCTION
The class Arachnida is second only to the Insecta class
in number of species and diversity (Savory, 1977). It is
also among the more important groups of invertebrates
occurring in caves, with 9 of the 11 extant orders con¬
taining troglobites. The windscorpions (order Solifugae)
do not occur in caves except as accidentals. The whips-
corpions (order Uropygi) rarely occur as trogloxenes in
tropical caves.
DISTRIBUTION
The class Arachnida is distributed worldwide and,
with the exception of many Acarina, is entirely terrestrial.
Of the orders containing troglobites, three (Schizomida,
Amblypygi, and Ricinulei) are predominantly tropical
and subtropical in distribution. Troglobitic Ricinulei occur
only in the New World tropics and subtropics and troglo¬
bitic Schizomida are known only from North America,
South America, and Australia. With the exception of one
species each in Israel and Sarawak, unquestionably
troglobitic Scorpiones occur only in Mexico and Ecuador.
The remaining orders include troglobites in both Old and
New Worlds and in tropical and temperate regions.
With some exceptions, genera containing troglobites
belong to families also known from the surface. The
spider family Holarchaeidae is known only from two
troglobitic species in New Zealand. Several families of
aquatic mites are apparently restricted to subterranean
waters. Although some genera contain only troglobites,
they typically are closely related to other genera occur¬
ring on the surface in the same area. Relict species are
known in several groups. The troglobitic telemid spider
Telema tenella from Europe is the only species in the
family occurring in temperate regions. Other isolated
species now inhabit caves in areas outside the general
range of the genus or family. These include species that
have survived in the moist habitat of caves in desert
regions. Examples of this type of distribution include
several species of Schizomida, Ricinulei, and Opiliones
in northern Mexico and the southern United States.
Early studies of the distribution of troglobites in gen¬
eral indicated that troglobites were extremely rare in
tropical lowlands. In recent years extensive studies in the
tropics have demonstrated a rich troglobitic fauna.
Troglobitic arachnids are especially abundant in areas
previously thought to contain a depauperate fauna. It is
typical of most tropical areas for the insect fauna to
include few troglobites. For example, in the Yucatan
Peninsula of Mexico, only one true insect (a phalangop-
sid orthopteran) is considered to be troglobitic, whereas
about 18 species of arachnids are troglobites. Arachnids
to a large extent appear to occupy the predatory positions
held by beetles and other insects in temperate caves.
Most species of troglobitic Arachnida have very
limited distributions, with many occurring in a sin¬
gle cave. A few widely distributed species may rep¬
resent newly evolved troglobites that have not been
isolated long enough for morphological differentia¬
tion to have occurred. In some cases, more detailed
analyses of populations have revealed the existence
of cryptic species.
BIOLOGY
The biology of few species of troglobitic arachnids
has been studied. In numerous cases only a single sex
or even specimen is known and the rarity of most
troglobites precludes anything more than cursory
observations. Except for many mites that may feed
on living or decaying organic matter or are parasites,
most arachnids are predators. Some Opiliones are
known to also feed on organic material.
Most troglobitic arachnids are preadapted for life in
caves. All species of Ricinulei and Palpigradi are eye¬
less and in the Schizomida the ocular structures consist
of one small eyespot on each side of the head.
Arachnid families containing troglobites typically live
in leaf litter, soil, or other sheltered habitats. Scorpions
of the genus Typhlochactas are all blind, but the troglo¬
bitic species have longer appendages than the litter¬
dwelling surface species.
Adaptations for the cave environment include loss
or reduction of eyes and pigment, elongation of appen¬
dages, and an increase in sensory structures. For the
few species where data are available, fewer and larger
eggs are produced and metabolic rates are lower. They
may also have longer life spans. Troglobitic species of
the dictynid spider genus Cicurina have been
Encyclopedia of Caves.
© 2012 Elsevier Inc. All rights reserved.
SPIDERS AND RELATED GROUPS
787
maintained in the laboratory for as much as 2 years
before reaching sexual maturity.
Epigean Amblypygi are nocturnal, spending the
day under rocks, ground debris, and crevices. Many
cavernicolous species have extremely elongate pedi-
palps and exhibit various degrees of eye reduction.
A few species are totally eyeless.
A study of the Hawaiian troglobitic spider Lycosa
howarthi Gertsch revealed greater water loss and a lower
metabolic rate than a related epigean species. Oxygen
consumption in the troglobitic species remained con¬
stant over a 12-hour period, whereas it increased in the
epigean species during darkness.
The biology and environment of the Australian troglo-
bite Dmculoid.es vinei (Harvey) have been studied in
detail. The troglophilic species Stenochrus portoricensis
Chamberlin of southern Mexico and the West Indies is of
special interest in that it is facultatively parthenogenetic.
With only a few exceptions, all insular populations are
parthenogenetic. In the Yucatan Peninsula of Mexico
there is an equal ratio of males and females from the sur¬
face, but essentially all specimens from deep within
caves are females. This presumably explains the extreme
abundance of this species in the caves of the peninsula.
Epigean Ricinulei inhabit leaf litter and the under¬
side of rocks. They are primarily predators, but caver-
nicole species have been found feeding on dead
arthropods and the feces of bats and millipedes.
Endogean Palpigradi are found under rocks and in
leaf litter. All are eyeless, depigmented, and with slen¬
der appendages. Troglobitic species are distinguished
by longer appendages and larger numbers of trichobo-
thria and other sensory structures.
CONSERVATION
Troglobitic Arachnida, as in all groups of cave organ¬
isms, are particularly susceptible to loss of populations
and even extinction. Small population sizes, long life
span, slow reproductive rates, and a high degree of
endemicity all place troglobitic arachnids at risk.
Deforestation in many parts of the world has resulted
in the loss of habitat for many species of invertebrates
and bats that provide essential nutrients into the cave
ecosystem. The loss of vegetation may result in environ¬
mental degradation, especially of small, shallow caves.
Modifications of natural vegetation in the vicinity of the
cave may result in invasion of introduced species.
Urbanization in many areas has resulted in the destruc¬
tion of many caves and impacted many others.
The actual number of species threatened by extinc¬
tion is unknown but it must be a very high number.
The U.S. Fish and Wildlife Service has placed 11 spe¬
cies of troglobitic arachnid on the official list of
endangered species. Ten of these are from two limited
areas of Central Texas where active urbanization
threatens essentially every cave with destruction.
Furthermore, the red imported fire ant Solenopsis
( Solenopsis ) invicta Buren has invaded hundreds of
caves with devastating effect on the cavernicole fauna.
The listed species are the neobisiid pseudoscorpion
Tartarocreagris texana (Muchmore); the phalangodid har¬
vestmen Texella cokendolpheri Ubick and Briggs, T. red-
delli Goodnight and Goodnight, and T. reyesi Ubick and
Briggs; and the dictynid spiders Cicurina ( Cicurella) bar-
onia Gertsch, C. (C.) madia Gertsch, C. (C.) venii
Gertsch, and C. (C.) vespera Gertsch; and the leptonetid
spiders Neoleptoneta microps (Gertsch) and N. myopica
(Gertsch). In addition to these species, the Hawaiian
lycosid spider Adelocosa anops Gertsch is also consid¬
ered endangered.
The listing of these species has slowed destruction of
caves in these very limited areas, but only an aggressive
program of conservation throughout the world can pre¬
serve the remarkable diversity of troglobitic arachnids.
SYSTEMATICS
The classification of the Arachnida is far from
settled, especially with respect to the phylogenetic rela¬
tionships of the orders. Many specialists now consider
the Acari to constitute a separate subclass of arachnids,
with some placing the order Ricinulei within that sub¬
class. Others consider the Acari to be polyphyletic and
provide a different arrangement for arachnid orders
and suborders. For convenience the Acari is considered
here as an order and the classic suborders maintained.
Order Scorpiones (Scorpions)
This order of about 1300 species in nine families is
worldwide in distribution (Fig. 1) (Volschenk and
Prendini, 2008). Scorpions where they occur in caves
are the top predator in the system. There are now 22
unquestionable troglobites known worldwide. An
additional six species known only from caves show
slight reduction in pigmentation or eyes but are proba¬
bly not troglobitic. There have been no studies, how¬
ever, on the biology or ecology of any of the species.
All of the troglobitic species are extremely rare, most
known only from a few specimens.
Family Akravidae. The family includes the troglobi¬
tic Akrov israchanani Levi from Israel.
Family Buthidae. The only unquestioned troglobite
in the family is Troglorhopalurus translucidus Lourenfo
et al. from Brazil. Troglotityobuthus gracilis (Fage) from
Madagascar has reduced pigment and may be troglobitic.
ENCYCLOPEDIA OF CAVES
788
s
FIGURE 1 Alacran tartarus from Sotano de San Agustin, Oaxaca,
Mexico. Photo courtesy of Robert and Linda Mitchell. Used with
permission.
Family Chactidae. Broteochactas trezzii (Vignoli
and Kovarik) from Venezuela has reduced eyes and
pigmentation and may be a troglobite.
Family Chaerilidae. Two species, Chaerilus chapmani
Vachon and Lourengo from Sarawak and C. sabinae
Lourenco from Indonesia, are troglobites. A third spe¬
cies, Chaerilus cavernicola Pocock from Indonesia, is
probably a troglophile.
Family Diplocentridae. The genus Diplocentrus
includes six species from caves in southern Mexico, of
which three are unquestioned troglobites (D. actun
Armas and Palacios-Vargas, D. anophthalmus Francke,
and D. cueva Francke). Diplocentrus mitchelli Francke con¬
tains reduced pigmentation but is probably not
a troglobite. Heteronebo clareae Armas from the
U.S.-owned Navassa Island in the Caribbean has reduced
eyes and pigmentation but is probably not troglobitic.
Family Euscorpiidae. Two species of the genus
Troglocormus ( T. willis Francke and T. ciego Francke)
have been described from caves in Mexico.
Family Liochelidae. Liocheles polisorum Volschenk
et al. from Christmas Island has reduced eyes and pig¬
ment and is certainly a troglobite.
Family Troglotayosicidae. The only troglobite in
the family is Troglotayosicus vachoni Lourenco from
Ecuador. A closely related species is known from lit¬
ter in adjoining Colombia. The European Belisarius
xambeui Simon was once considered troglobitic
because of its reduced eyes. It is now known to be a
soil-dwelling form occasionally found in caves.
Family Typhlochactidae. This family includes seven
troglobites. The Mexican genus Typhlochactas includes
four troglobites (T. cavicola Francke, T. poncho Sissom,
T. reddelli Mitchell, and T. rhodesi Mitchell) from the
Mexican states of Tamaulipas and Veracruz. This
genus also contains an eyeless species from leaf litter
in southern Mexico. Sotanochactas elliotti (Mitchell) is
known only from San Luis Potosi. The genus Alacran
includes two troglobitic species A. tartarus Francke and
A. chamuco Francke from Oaxaca. Alacran tartarus is
especially interesting in that it has been found more
than 900 m below the surface and has been collected
from under-water on at least one occasion.
Family Urodacidae. Aops oncodactylus Volschenk and
Prendini from Barrow Island, off the coast of Western
Australia, is the only troglobite in the family.
Family Vaejovidae. The only unquestioned troglo¬
bite in the family Vaejovidae is Vaejovis gracilis Gertsch
and Soleglad from Veracruz, Mexico. Many other spe¬
cies of Vaejovis have been found in caves but most are
at best considered troglophiles. Uroctonus grahami
Gertsch and Soleglad with reduced eyes is known only
from caves in California, U.S.A., and is probably a trog¬
lobite. The genus Pseudouroctonus includes the troglobi¬
tic P. sprousei Francke from Mexico. Other species of
troglophile are known from Mexico. Pseudouroctonus
reddelli (Gertsch and Soleglad), is an extremely abun¬
dant troglophile in the caves of Texas, U.S.A.
Order Schizomida (Short-Tailed Whipscorpions)
This small order contains about 200 species in two
families (Reddell and Cokendolpher, 1995). About 80
species, some awaiting description, have been recorded
from caves. They range in size from 2 to 12.4 mm in
total length. Epigean species are found under rocks, in
leaf litter, and some species are associated with termites
and ants.
Family Hubbardiidae. This family occurs world¬
wide in the tropics and subtropics. A large number of
species of Hubbardiidae have been recorded from
caves, but unquestioned troglobites are known only
from caves in Cuba, Jamaica, Belize, Mexico, California
(U.S.A.), Ecuador, and Australia. Two monotypic gen¬
era, Cokendolpherius and Reddellzomus, have recently
been described from caves in Cuba. Two species of
Draculoides have been found in Western Australia.
Hubbardia from California and Arizona is a genus most
closely related to the Asian fauna. The only troglobite
is H. shoshonensis (Briggs and Horn) from a single
cave surrounded by desert. The genus Rowlandius
includes numerous species, including several troglo¬
philes. Troglobites are known from caves in Cuba and
Jamaica. The genus Sotanostenochrus contains only two
troglobitic species from Mexico. The large genus
Stenochrus contains eight described and several unde¬
scribed troglobites from caves in Mexico. The genus
Stewartpeckius contains only the troglobite S. troglobius
(Rowland and Reddell) from Jamaica. The genus Tayos
ENCYCLOPEDIA OF CAVES
SPIDERS AND RELATED GROUPS
789
contains only the troglobite T. ashmolei (Reddell and
Cokendolpher) from Ecuador.
Family Protoschizomidae. Eleven species have been
described from Mexico. Protoschizomus includes four
species of troglobite from the states of San Luis Potosi
and Tamaulipas; an undescribed species is from Texas
(U.S.A.). Agastoschizomus with five species from caves
in Guerrero, Hidalgo, San Luis Potosi, and Tamaulipas
includes A. lucifer Rowland, at 12.4 mm in length the
largest species in the order.
Order Amblypygi (Tailless Whipscorpions)
This order of about 100 species in three families is
distributed worldwide in tropical and subtropical habi¬
tats (Weygoldt, 1994). Many species are large (up to
4.5 cm in body length) with extremely elongate pedi-
palps and antenniform first legs (Fig. 2). Cavernicolous
species occur in all families and virtually every species
found in karst regions also inhabits caves. About 40
species have been recorded from caves.
Family Charontidae. This family is largely Asian in
distribution but the genus Charinides includes three tro-
globitic species in Cuba. Five or six species of the genus
Stygophrynus inhabit caves in Malaysia, Burma, Borneo,
and Java. An undescribed eyeless Charinus is known
from caves in Belize. Other genera containing cavernic¬
olous species include Phrynichosaurus and Sarax from
Malaysia, Tricharinus from Jamaica and Surinam, and
Charinus from Africa, Venezuela, and Cuba.
Family Phrynichidae. This tropical family contains
cavernicolous species in the genera Phrynichus (Africa),
Damon (Africa), and Trichodamon (Brazil). All are prob¬
ably troglophiles.
Family Phrynidae. The family Phrynidae is the
dominant family in the New World tropics and
FIGURE 2 Paraphnjnus chacmool from Actun Ziz, Yucatan,
Mexico. Photo courtesy of Robert and Linda Mitchell. Used with
permission.
subtropics and is well represented in caves. All genera
contain cavernicolous species, but the genus
Paraphrynus is notable for containing six described tro-
globites and several troglophiles in Mexico. A large
undescribed troglophile (previously recorded as P. rap-
tator) is sympatric with the troglobitic P. chacmool
(Rowland) in Yucatan, Mexico. Both species are found
on cave walls, but P. chacmool has been found only in
total darkness far from the cave entrance. The genus
Acanthophrynus includes A. coronatus Kraepelin from
southern Mexico. One species of Heterophrynus has
been found in Colombian caves, but other species of
the genus may occur in caves. Phrynus contains several
troglophiles in Mexico and Texas (U.S.A.)
Order Araneae (Spiders)
This is the largest order of Arachnida with more
than 50,000 described species (Ribera and Juberthie,
1994). Thousands of species, including large numbers
of troglobites and troglophiles, have probably been
recorded from caves. The diversity of troglobitic spi¬
ders is great with numerous families being recorded in
essentially every karst region worldwide.
Suborder Mesothelae
Family Liphistiidae. This primitive family is tropi¬
cal in distribution. Troglobitic species of the genus
Liphistius are known from Malaysia and Thailand.
Suborder Mygalomorphae
Family Barychelidae. The only troglobite in this
family is Troglothele coeca fage from Cuba.
Family Cyrtaucheniidae. The only troglobitic spe¬
cies in this family is Acontius stercoricola (Denis) from
Guinea.
Family Dipluridae. Troglobitic and troglophilic spe¬
cies in the genus Euagrus have been found in Mexico.
The genus Masteria includes troglobites in Jamaica and
the Philippines. Troglodiplura lowryi Main is a troglo¬
bite in Australia.
Family Hexatelidae. The only troglobite in this fam¬
ily is Hexathele cavernicola Forster from New Zealand.
Family Microstigmatidae. The only troglobite in this
family is Spelocteniza ashmolei Gertsch from Ecuador.
Family Theraphosidae. This family of tarantulas
includes six troglobites in the genus Hemirrhagus in
Mexico. One species is eyeless while the others have
reduced eyes and pigment.
Suborder Araneomorphae. At least 25 families of
Araneomorphae are considered to contain troglobites,
although some are better represented in caves than
others.
Family Agelenidae. The genus Tegenaria includes
troglobitic and troglophilic species in Europe and the
United States.
ENCYCLOPEDIA OF CAVES
790
s
Family Anapidae. This largely tropical family
includes troglobites in the genera Cruzetulus in South
Africa and Conculus from Korea.
Family Austrochilidae. This primitive tropical fam¬
ily contains only one troglobite, Hickmania troglodytes
(Higgens and Peters) from Tasmania.
Family Cybaeidae. This family includes troglobitic
species of the genus Cybaeus from Japan and Korea.
Family Cycloctenidae. This family includes troglo¬
bitic species in the genera Cycloctenus and Toxopsiella
from Australia and Tasmania.
Family Dictynidae. The genera Blabomma and
Chorizomma contain troglobitic species in Europe. The
genus Cicurim includes troglobites in the United States,
Mexico, and Japan. The radiation of this genus in the
caves of Texas (U.S.A.) is especially remarkable with
more than 50 eyeless species known.
Family Dysderidae. This family includes a large
number of troglobitic species from caves in the
Mediterranean basin. The most abundant genus is
Dysdera with numerous species recorded. Other signifi¬
cant genera include Folkia, Harpactea, Harpactocrates,
Minotauria, Rhode, Rhodera, Speleoharpactea, and Stalita.
Family Gnaphosidae. Troglobitic species in this
family include two species of Lygromma from the
Galapagos Islands and one of Herpsyllus from Cuba.
Family Hahniidae. This family of small spiders
includes troglobites in the genera Hahnia and Iberina
from Europe.
Family Flolarchaeidae. This family includes only
two troglobitic species, Holarchaea globosa (Hickman)
from Tasmania and H. novaeseelandiae (Forster) from
New Zealand.
Family Leptonetidae. This family is primarily hol-
arctic in distribution with a few species in southern
Mexico. It is well represented in caves, with numerous
troglobitic and troglophilic species in many genera
having been described. About 60 species of troglobites
in several genera, including Falcileptoneta, Leptoneta,
Marisana, and Sataurana, have been described from
caves in China, Korea, and Japan. The fauna of Europe
and northern Africa include many troglobites in the
genera Barusia, Leptoneta, Leptonetela, Paraloptoneta,
Protoleptoneta, Sulcia, and Teloleptoneta. Cave-associated
species in the United States include representatives of
the genera Appaleptoneta, Archoleptoneta, Callileptoneta,
and Neoleptoneta. The genus Darkoneta includes a troglo¬
bitic species in Mexico. Two species of Archeoleptoneta
and eight of Neoleptoneta are known from caves in
Mexico, but only five of the latter genus are considered
troglobites. Other species in the United States await
description.
Family Linyphiidae. This large family contains numer¬
ous troglobitic and troglophilic species in caves throughout
the Northern Hemisphere. The European fauna includes
many troglobites in the genera Caviphantes, Centromerus,
Icariella, Lepthyphantes, Porrhomma, Thyphlonyphia, and
Troglohyphantes. Species of Lepthyphantes have also been
found in caves in Korea and South Africa. Other genera
represented by troglobites include Anthrobia in the United
States and Phanetta in the United States and Mexico,
Erigone and Meioneta in the Hawaiian Islands, Allomengea
and Jacksonella in Korea, Dunedinia in Australia, and
Metopobactrus and Walckenaeria in the Canary Islands.
Family Liocranidae. Genera represented by troglo¬
bites include Brachyanillus from Algeria and Spain,
Liocmnum from Spain, and Agraecina from the Canary
Islands and Romania.
Family Lycosidae. This family of wolf spiders is
represented in caves only by accidentals, with the
exception of two remarkable species, Lycosa howarthi
Gertsch and Adelocosa anops Gertsch from the Hawaiian
Islands.
Family Mimetidae. The only troglobitic species in
this family is Mimetus strinatii Birgnoli from Ceylon.
Family Mysmenidae. This family is represented from
caves by troglophiles in the genus Maymena in Mexico
and troglobites in the genus Trogloneta in Tasmania.
Family Nesticidae. The family Nesticidae is closely
associated with caves, with numerous species of trog¬
lobite and troglophile having been described. The
genus Nesticus includes troglobites and troglophiles in
the United States, Mexico, China, Korea, Ceylon,
Japan, and Europe. Eidmannella includes several troglo¬
bites from Texas (U.S.A.). Other genera containing tro¬
globites include Canarionesticus and Typhlonesticus
from Europe. The genus Nesticiella is abundant in
caves in Asia, central Africa, Hawaii, and Fiji.
Family Ochyroceratidae. This is a tropical family of
small spiders that contains cavernicolous species in
North and South America, the Antilles, Asia, and
Africa. The genus Ochyrocera includes troglobites and
troglophiles in Mexico, Guatemala, Peru, and Cuba.
Troglobitic and troglophilic species of Theotima are
known from Mexico, Belize, Hawaii, Jamaica, and
Cuba. Two other genera, Fageicera and Speocera, include
troglophiles in Cuba. Troglophilic species in the genera
Athepus, Psiloderces, Simonicera, Speocera, and Theotima
have been recorded from caves in Asia. The genus
Speloderces includes a troglophile in South Africa.
Family Oonopidae. This tropical family of small spi¬
ders includes several cave-associated species, including
several troglobites. Troglobitic species include species
of Wanops in Mexico, Oonopsides in Cuba, Gamasomorpha
in Ecuador and Ceylon, and Dysderoides in India.
Family Pholcidae. This cosmopolitan family is well
represented in caves, especially in the New World
tropics. More than 90 troglophiles and troglobites in
the genera Anopsicus, Coryssocnemis, Ixchela, Metagonia,
Modisimus, Physocyclus, and Psilochorus have been
ENCYCLOPEDIA OF CAVES
SPIDERS AND RELATED GROUPS
791
collected from caves in Mexico. Troglobitic species of
Metagonia have been described from caves in Belize,
Cuba, Jamaica, and the Galapagos Islands. Other genera
containing cave-associated species include Artema,
Pholcus, and Spermophora in New Guinea; Aymaria in the
Galapagos Islands; Blancoa, Chebchea, and Mesabolivar in
Venezuela; Priscula in Peru; Spermophora in Tasmania;
and Spermophoroid.es in the Canary Islands.
Family Prodidomidae. This small family includes
only one troglobite, Lygromma gertschi Planick and
Shadab. Lygromma anops Peck and Shear from the
Galapagos Islands was considered troglobitic but it has
since been found on the surface
Family Stiphidiidae. This small African and
Australasian family includes troglobites and troglo-
philes in the genera Baiami, Stiphidion, and Tartarus
from caves in Australia and Tasmania.
Family Synotaxidae. Three species of the genus
Tupua, one troglophile and two troglobites, have been
described from caves in Tasmania.
Family Telemidae. Cave-associated species, includ¬
ing some troglobites, of this family include species of
Telema in Japan, and Guatemala; of Telemofila in New
Caledonia; Usofila in the United States; Cangoderces in
South Africa; and Apneumonella in Tanganyika. Of spe¬
cial interest is the troglobitic Telema tenella Simon from
Europe. This is considered a "living fossil" and is the
only representative of the family from temperate
regions.
Family Tetrablemmidae. This small tropical family
includes cave-associated species in the genus Tetrablemma
from Mexico, Caraimatta from Mexico, and Ablemma from
Okinawa and Sumatra.
Family Tetragnathidae. Genera in this family con¬
taining cave-associated species include Meta in Canada,
the United States, Europe, Russia, and Tasmania;
Metellina in Europe; and Orsinome in Tasmania.
Family Theridiidae. This large family includes
numerous troglophiles in caves in many parts of the
world. Troglobitic species include representatives of
the genera Achaearanea, Coscinidia, and Stemops in New
Guinea; Pholeomma and Steatoda in Austrialia; Icona in
Australia and Tasmania; Robertus in Europe; Theridion
in the Azores and Galapagos Islands; and Thymoites in
the United States.
Family Theridiosomatidae. Cave-associated species
in this family include representatives of the genera
Plato in Venezuela, Wendilgarda in Guadeloupe and
Ceylon, and Andasta in Ceylon.
Order Palpigradi (Micro Whipscorpions)
This is a small order containing 78 species arranged
in two families (Conde, 1998) - note there should be an
accent over the e in Conde. It is remarkable that 27
described species are known only from caves. Several
others have also been reported from caves but also
occur on the surface. Palpigrades range in body length
from about 1.9 to 2.8 mm.
Family Eukoeneniidae. The genus Koenenides is
represented in caves only by K. leclerci Conde from
Thailand. Troglobitic species of Eukoenenia have been
found in Europe (22 species), Cuba (1 species), Thailand
(3 species), and Indonesia (1 species). Additional troglo¬
bitic species in the genera Allokoenenia, Eukoenenia, and
Koenenides have been found in India, Sulawesi, and
Thailand.
Family Prokoeneniidae. The only described troglo¬
bitic species of this family is Prokoenenia celebica Conde
from Sulawesi, Indonesia. Two additional species
of the genus from Southeast Asia and one from
California await description.
Order Pseudoscorpiones (Pseudoscorpions)
This order of about 3000 species in 20 families is
worldwide in distribution and is extremely well repre¬
sented in caves (Fig. 3) (Heurtault, 1994). More than 400
species in 15 families have been described from caves.
Family Atemnidae. Cave-associated species have
been described from the genera Atemnus in the
Philippines, Oratemnus in Australia, and Catatemnus
and Titanatemnus from Africa.
Family Bochicidae. This family is primarily
restricted to the neotropics. The northernmost popula¬
tions are apparent relicts in the caves of Texas (U.S.A.)
and northern Mexico. Most species are troglobitic. Five
genera that contain only troglobites are Antillobisium
with two species in Cuba, Troglobochica with two spe¬
cies in Jamaica, Troglohya with two species in southern
FIGURE 3 Tartarocreagris infernalis from Electro-Mag Cave,
Williamson County, TX, U.S.A. Photo courtesy of Robert and Linda
Mitchell. Used with permission.
ENCYCLOPEDIA OF CAVES
792
s
Mexico, Paravachonium with five species in Mexico,
and Vachonium with seven species in southern Mexico
and one in Belize. The genus Leucohya includes
two troglobites in northern Mexico and one in Texas.
The genus Mexobisium includes one troglobite from
Guatemala, two from Belize, one from the Dominican
Republic, and two from southern Mexico.
Family Cheiridiidae. This family contains few spe¬
cies associated with caves but species of Cheiridium from
Namibia and Cuba; Cryptocheiridium from Australia,
Malaysia, Philippines, and Cuba; and Neocheiridium
from Curasao are probably troglophiles.
Family Cheliferidae Genera containing cave-
associated species in this family include Lissochelifer in
Kenya, Mexichelifer in Mexico, Protochelifer in Australia,
and Stygochelifer in Java.
Family Chernetidae. Many species of this family have
been recorded from caves where they are frequently
present in bat guano in very large numbers. The ecologi¬
cal status of these species is unknown but most are prob¬
ably troglophiles. Genera recorded from caves include
Bitulochemes and Epactiochernes from Cuba; Neoallochernes
from Antigua in the Lesser Antilles; Hesperochernes,
Neoallochernes, Tejachernes, and Dinocheirus from the United
States; Chernes, Lasiochernes, and Pselaphochernes from
Europe; Chelanops, Dinocheirus, Lustrochernes, Parachemes,
and Tejachernes from South America; Dinocheirus and
Megachemes in Asia; Cajfrowithius and Nudochemes from
Africa; and Sundochernes and Troglochemes from Australia.
Family Chthoniidae. This is one of the more impor¬
tant families of pseudoscorpions inhabiting caves
worldwide. An estimated 200 species of the family
have been described from caves, of which a large
number in many genera are troglobites. The European
fauna include about 80 species of Chthonius, as well as
species of the genera Microchthonius, Paraliochthonius,
Spelyngochthonius, Troglochthonius, and Tymnnochthonius.
The Asian fauna include representatives of the genera
Allochthonius, Lagynochthonius, Mundochthonius, and
Pseudotymnnochthonius. Austrochthonius, Lagynochthonius,
Pseudotyrannochthonius, Sathrochthonius, and Tyran-
nochthonius have been recorded from Australia. Chthonius
has been recorded from northern Africa, Selachthonius
from South Africa, and Tymnnochthonius from Kenya.
The South American fauna include species of Lechytia
from Trinidad, Pseudochthonius from Brazil, and
Tyrannochthonius from Peru. Tyrannochthonius and
Kleptochthonius are especially abundant in the caves of
the United States, with other genera including species
of Aphrastochthonius, Apochthonius, Mexichthonius,
Mundochthonius, and Neochthonius. The fauna of Mexico
and Central America include species of Aphrastochthonius,
Lechytia, Mexichthonius, Pseudochthonius, and Tyrannochtho¬
nius. Lagynochthonius and Tyrannochthonius have been
recorded from caves in Jamaica. Species of Tyrannochtho-
nius and Vulcanochthonixis are troglobites in Hawaiian
lava tubes.
Family Garypidae. The family Garypidae includes
species of Archeolarca and Larca in the United States
and of Larca in Europe.
Family Ideoroncidae. The genus Typhloroncus in¬
cludes several species of troglobite in Mexico. Albiorix
includes three troglobites in Mexico and one in Arizona
(U.S.A.). Species of Dhanus have been described from
caves in Malaysia and one of Negroroncus from the
Congo.
Family Neobisiidae. This is a dominant family of
pseudoscorpions in the caves of the Holarctic region,
with particular radiation in the United States and
Europe. Species in the genera Microcreagris, Pamroncus,
and Parobisium are troglobites in Japan. The genera
Neobisium with more than 70 species and Roncus with
more than 40 species are the dominant genera in
Europe and northern Africa. Other genera with troglo¬
bites in Europe include Acanthocreagris, Balkanoroncus,
Insulocreagris, and Roncobisium. The fauna of the United
States include species in the genera Alabamocreagris,
Austrolinocreagris, Lissocreagris, Minicreagris, Novobisium,
Parobisium, Tartarocreagris, and Trisetobisium.
Family Pseudochiridiidae. This small tropical fam¬
ily includes cave-associated species in the genera
Pseudochiridium from Malaysia and Paracheiridium from
Madagascar.
Family Pseudogarypidae. This small family in¬
cludes three troglobitic species of Pseudogarypus in
California and Arizona (U.S.A.).
Family Sternophoridae. This small family includes
a species of Afrosternophorus described from a cave in
Papua New Guinea.
Family Syarinidae. This family contains cave-
associated species in the caves of Europe, North
America, South America, and the West Indies. The
genus Chitrella contains several species, including pos¬
sible troglobites, from caves in the United States.
Chitrellina chiricahuae Muchmore, the only member of
the genus, is an eyeless species from a cave in Arizona
(U.S.A.). Ideoblothrus includes cave-associated species
from caves in Mexico and Australia. The genus
Ideobisium contains species in Ecuador, Venezuela, and
Puerto Rico. The genera Hadoblothrus, Pseudoblothrus,
and Troglobisium include troglobites from Europe.
Family Tridenchthoniidae. The only cave-associated
species of this family is Tridenchthonius juxtlahuaca
Chamberlin and Chamberlin from a cave in Guerrero,
Mexico.
Family Withidae. Cave-associated species of this
family have been recorded for the genera Parawithius
in Venezuela and Pycnowithius in Kenya.
ENCYCLOPEDIA OF CAVES
SPIDERS AND RELATED GROUPS
793
Order Ricinulei (Hooded Tickspiders)
This is the smallest order of arachnids with about 50
extant species in one family and three genera (Fig. 4)
(Juberthie, 1994). One genus, Ricinoides, is exclusively
African and contains no cave-associated species. Three
species from northern Mexico inhabit caves at higher
elevations of isolated mountain ranges surrounded by
desert and are apparent relicts. Once thought to be
extremely rare, populations of both troglophiles and
troglobites may be present in vast numbers.
Family Ricinoididae. Cryptocellus is known from
South and Central America, whereas the genus
Pseudocellus occurs in Texas (U.S.A.), Mexico, Central
America, and Cuba. The only described troglobitic spe¬
cies of Cryptocellus is C. bordoni (Dumetresco and
Juvara-Bals) from Venezuela. Cave-associated species
of Pseudocellus occur in Cuba, Guatemala, and Mexico.
The troglobitic P. silvai (Armas) and the troglophilic
P. paradoxus (Cooke) have been described from Cuba
Pseudocellus krejcae Cokendolpher is a troglobite in
Belize. The Mexican cavernicole fauna include the fol¬
lowing troglobitic species: P. osorioi (Bolivar y Pieltain)
from Tamaulipas and San Luis Potosi, P. reddelli
(Gertsch) from Durango, and P. sbordonii (Brignoli)
FIGURE 4 Pseudocellus osorioi from Sotano del Tigre, San Luis
Potosi, Mexico. Photo courtesy of Robert and Linda Mitchell. Used with
permission.
from Chiapas. Two additional undescribed troglobites
are also known from Mexico.
Order Acari (Mites and Ticks)
This order of about 40,000 described species is prob¬
ably represented in subterranean habitats by thou¬
sands of species, but comparatively few are terrestrial
troglobites (Dusbabek, 1998; Palacios-Vargas, 1998).
Numerous species of mites found in the interstitial
habitat below streams and in cave waters exhibit adap¬
tations to subterranean conditions.
Suborder Notostigmata
Family Opilioacaridae. This rare primitive group
includes two troglophilic species of Opilioacarus from
caves in Cuba. The genus siamacarus includes two spe¬
cies, one of which is troglomorphic, from caves in
Thailand.
Suborder Mesostigmata. Large numbers of species,
both free living and parasitic, in this suborder have
been recorded from caves. Numerous species in the
families Macronyssidae and Spinturnicidae are para¬
sites of bats and have been recorded from caves.
The free-living members of the suborder inhabit a
wide variety of biotopes, including litter and soil, and
they easily colonize caves. Many species have been
recorded from caves in all parts of the world, but only
a few exhibit troglomorphic adaptations.
Family Macrochelidae. This family contains a large
number of cavernicolous species; some are known
only from caves but all are probably troglophiles.
Among the more significant genera that inhabit caves
are Macrocheles in Japan, Europe, Mexico, Venezuela,
and the West Indies; Holostaspis in Europe and Algeria;
and Geholaspis in Europe. Most of these have been
taken from bat guano.
Family Parasitidae. This family includes a large
number of cave-associated species, many abundant in
bat guano. Many species of Eugamasus have been
recorded from caves in Afghanistan, Japan, and
Europe. One European species is considered a proba¬
ble troglobite. Other important cave-associated genera
include Pergamasus and Parasitus in Europe and
Mexico; Paracarpais in Europe and the United States;
and Vulgarogamasus in Europe and the United States.
Family Uropodidae. This family contains a large
number of cave-associated species. They are frequently
present in bat guano with populations in the millions.
Troglobitic species are known in the following genera:
Chiropturopoda (3 species from Romania, South Africa,
and Zaire), Nenteria (5 species from Europe and
Trinidad), Oplitis (2 species from Cuba, Venezuela, and
Trinidad), Trichouropoda (5 species from Cuba, Java,
Afghanistan, and Europe), Uroobovella (12 species from
Europe, Java, Cuba, and Trinidad), Uropoda (2 species
ENCYCLOPEDIA OF CAVES
794
s
from Mexico, the Antilles, and Austria), and Uroseius
(3 species from Japan, Italy, and Bosnia—Herzegovina).
Family Zerconidae. This family of predatory mites
includes cave-associated species of the genera
Dithnozercon from the United States, Prozercon from
Japan, and Zercon from Europe. Paleozercon cavernicolus
Blaszak et al. is a late Pleistocene fossil recovered from
calcite in Hidden Cave, New Mexico (U.S.A.).
Suborder Metastigmata. This suborder includes the
ticks, many species of which have been found as para¬
sites of bats.
Family Argasidae. Many species of this family are
parasites of bats. Species of the genus Carios occur in vast
numbers in bat guano in the caves of Cuba and Mexico
and apparently feed on the guano. A few species of the
genus Argas have unusually long legs and sensory organs
that indicate a degree of adaptation to cave life.
Numerous species of Carios have been taken from caves.
Some are parasites of bats but others have been taken
from a variety of ground-dwelling mammals. Ornithodoros
turicata (Duges) in the caves of the southwestern United
States is a known vector of relapsing fever and many
cases of this disease have been documented in cave
explorers following visits to caves inhabited by it.
Family Ixodidae. Few species of this family have
been recorded from caves except as accidentals.
Several species of the genus Ixodes in Europe are para¬
sites of bats and frequently found in caves. One spe¬
cies, Ixodes conepati Cooley and Kohls, a parasite of
small mammals, is frequently taken from caves in
Texas and New Mexico (U.S.A.).
Suborder Prostigmata (Terrestrial). This suborder
includes the majority of mites considered troglobitic.
The families Myobiidae and Psorergatidae include
numerous bat parasites that have been recorded from
caves. The majority of aquatic mites recorded from
subterranean habitats also belong in the suborder.
Family Cunaxidae. The family includes one troglo¬
bitic species of Bonzia in England. Troglophiles in the
genera Cunaxa and Cunaxoides are known from Spain
and Mexico, respectively.
Family Ereynetidae. This family includes, in addi¬
tion to bat parasites in the genera Neospeleognathopis
and Speleochir, a troglobitic species of Ereynetes from
lava tubes in Japan. Other cave-associated species
include representatives of the genera Ereynetes from
Mexico and Riccardoella from Europe.
Family Leeuwenhoeckiidae. Four species of troglobite
belonging to the genera Heterotectum, Ischnothrombium,
Pentagonotectum, and Tectumpilosum have been described
from bat guano in Cuban caves.
Family Proterorhagiidae. This family was described for
the troglomorphic species Proterorhagia oztotlica Lindquist
and Palacios-Vargas from caves in Colima, Mexico.
Family Rhagidiidae. This is the most important
family of mites inhabiting caves. A large number of
species have been recorded as troglophiles and troglo-
bites. The ecological status of some species is uncer¬
tain, but many are doubtless troglobitic. Troglobites
and troglophiles have been described in the genera
Coccorhagidia from Cuba; Flabellorhagidia from the
United States; Foveacheles from Europe, Mexico, United
States, and Hawaii; Poecilophysis from Europe, Mexico,
and United States; Rhagidia from Europe, Mexico,
United States, South Korea, and Japan; Robustocheles
from Canada, United States, and Mexico; Traegardhia
from Italy; and Troglocheles from Europe.
Family Trombiculidae. This family includes numer¬
ous species of bat parasites that have been recorded
from caves. Species in the genera Cubanothrombium
and Heterothrombium from caves in Cuba appear to be
troglobitic. Two species of the genus Trombicula from
Afghanistan, Java, and Morocco are considered troglo¬
bites. Species of Microtrombicida and Nycterinastes have
been collected from caves in Mexico.
Family Trombidiidae. This family includes parasites
of bats, but others are predators of insects. Many species
have been recorded from caves. Cave-associated species,
including some probable troglobites, are known in the
genera Anomalothrombium from Madagascar; Hannemania
from Mexico; Leptothrombium and Neotrombicida from
Romania; Spelaeothrombium from Europe and Africa;
Speothrombium, Dolichotrombicula, and Discotrombidium
from Cuba; and Typhlothrombium from Belgium.
Suborder Prostigmata (Aquatic). An estimated 4000
species in 100 families and more than 300 genera have
been recorded from aquatic habitats (Schwarz et at.,
1998). A large percentage of these are known only from
subterranean habitats that include caves and phreatic
and interstitial habitats. Subterranean species show
reduction or loss of eyes and pigment and many species
possess elongated bodies and short legs. Only a few
of the more important families are discussed here.
Other families containing subterranean species include
the Anisitsiellidae, Arenohydracaridae, Arrenuridae,
Athienemanniidae, Bogatiidae, Chappuisididae, Feltri-
dae, Halacaridae, Hungarohydracaridae, Hydrovolziidae,
Kantacaridae, Lebertiidae, Mideopsidae, Momoniidae,
Neocaridae, Nipponacaridae, Omartacaridae, Piersigii-
dae, Pionidae, Sperchontidae, Torrenticolidae, Trombidii¬
dae, and Unionicolide.
Family Aturidae. This family contains a large number
of species adapted for subterranean existence. Among
the more important genera are Aturus from Europe,
Mexico, and Japan; Axonopsella from Cuba, South
America, Australia, and Tasmania; Axonopsis from
Europe, North America, Asia, North Africa, and Cuba;
and Frontipodopsis from Europe, Asia, North America,
ENCYCLOPEDIA OF CAVES
SPIDERS AND RELATED GROUPS
795
Central America, and South America. Numerous genera
are known only from subterranean waters.
Family Hydryphantidae. This is an abundant family
in underground waters, with a large number of
described species. Genera containing subterranean spe¬
cies include Clathrosperchon from the southern United
States, Mexico, and South America; Cowichiana from
Canada; Cyclothyas from the United States; Euwandesia
from Chile and New Zealand; Thyasella from Europe;
and Wandesia from Europe, North America, South
America, Cuba, India, Australia, Tasmania, and Siberia.
Family Hygrobatidae. This family contains numer¬
ous species recorded from subterranean waters.
Among the more important genera are Atractides from
Europe, South America, and Asia; Australiobates,
Callumobates, Camposea, and Decussobates from Chile;
Corticacarus from South America; Gondwanobates from
Australia and New Zealand; and Hygrobates from
India.
Family Limnesiidae. This family includes a large
number of subterranean species, including many depig-
mented, eyeless species. Among the more important
genera are Kuwamuraarus from North America, Mexico,
India, Indonesia, and Japan and Neomamersa with many
species in North America, Central America, South
America, and the West Indies.
Family Limnohalacaridae. Two genera, Lobohalacams
and Soldanellonyx, contain species recorded from caves
in Japan. Other genera with species recorded from
caves include Troglohalacarus and Pamsoldanellonyx
from Europe and Homohalacarus from the United States.
Numerous other genera have been recorded from other
subterranean habitats.
Family Stygotrombidiidae. This family is extremely
abundant in subterranean waters of Europe, with
numerous species of Stygotrombidium having been
described. Other genera associated with underground
waters include Cerberotrombidium, Charonotrombium,
Hydrotrombium, and Victatrombium.
Suborder Astigmata. Most of the species in this sub¬
order are parasites of bats and other mammals.
Representatives of the families Acaridae, Chirodiscidae,
Chirorhynchobiidae, Rosensteiniidae, Sarcoptidae,
Teinocoptidae, and Gastronyssidae have been taken
from bats in caves.
Family Acaridae. Species of this family feed on
dead insects and other organic material. The ecological
status of cave-associated species is generally unknown.
Genera recorded from caves include Aellenella from
Kenya, Acotyledon from Puerto Rico, Caloglyphus from
Mexico and Trinidad, Sancassania from Mexico, and
Schwiebea from Europe and Kenya.
Suborder Cryptostigmata. This large group contains
more than 7000 species arranged into 140 families and
700 genera. Epigean species are taken from soil, litter.
and other endogean habitats. They are extremely abun¬
dant in caves but remain poorly studied. A few of the
more important families recorded from caves are dis¬
cussed below.
Family Damaeidae. The genus Belba includes one
troglobite from Europe and one apparent troglophile
from Mexico. Other genera with cave-associated spe¬
cies include Damaeus from Algeria, Europe, and Korea;
Dameosoma from Europe; Epidamaeus from the United
States and Korea; and Hypodamaeus from Scandanavia.
Family Galumnidae. Species of the genus Galumna
occur in the caves of Yucatan, Mexico, and other neo¬
tropical areas. Populations in the millions may be pres¬
ent in some caves.
Family Haplozetidae. The frugivorous species
Rostrozetes foveolatus Sellnick has been taken from
caves in Cuba and Trinidad, where it is extremely
abundant on bat guano.
Family Hypochthoniidae. Genera with cave-
associated species include Eohypochthonius from
Cuba and Trinidad, Hypochthonius from Europe, and
Malacoangelia from Cuba.
Family Microzetidae. The cave-associated species of
this family have been found in the guano of birds
and bats. Genera taken from caves include Acarocems
from Cuba and Mexico, Gymnozetes from Cuba, and
Microzetes from Trinidad.
Family Oppiidae. This is one of the more important
families of cryptostigmatid mites found in caves. Most
species are probably troglophiles. Many genera have
been recorded from caves, including Amerioppia and
Multioppia from Fiji; Bmchioppia from Puerto Rico;
Hypogeoppia, Kunoppia, Lenroppia, Medioppia, Microppia,
Oppia, and Serratoppia from Europe; Lasiobelln from the
Canary Islands; Amerrioppia and Oppia from Mexico;
and Oppia from Cuba.
Family Scheloribatidae. This important family includes
cave-associated species in the genera Monoscheloribates
from Mexico; Poroscheloribates from the Canary Islands;
and Scheloribates from Mexico, Cuba, Puerto Rico, and Fiji.
Family Sphaerochthoniidae. The genus Sphaer-
ochthonius is frequently found in caves, with species
having been recorded from Fiji, Cuba, Puerto Rico,
and Mexico.
Order Opiliones (Harvestmen)
This order contains an estimated 5500 species
divided into four suborders (Fig. 5) (Rambla and
Juberthie, 1994).
Suborder Cyphophthalmi. The Cyphophthalmi is a
small group of about 100 species. These mite-like har¬
vestmen are considered the most primitive members
of the order. Many epigean species are blind, but the
ENCYCLOPEDIA OF CAVES
796
s
FIGURE 5 Texella reyesi from Electro-Mag Cave, Williamson
County, TX, U.S.A. Photo courtesy of Robert ami Linda Mitchell. Used
with permission.
cave species possess longer legs and other adaptations
for a subterranean existence.
Family Neogoveidae. The only troglobitic species in
this small family is Neogovea mexasca Shear from Mexico.
Family Pettalidae. Although all species of this fam¬
ily are blind, the only troglobite is Speleosiro argasifor-
mis from South Africa.
Family Sironidae. Eight troglobitic and one troglo-
philic species of Siro have been described from
caves in Europe. The most highly troglomorphic
species is Tranteeva paradoxa from Bulgaria. The only
species recorded from tropical caves is Marwe coarctata
from Kenya.
Family Stylocellidae. Two species in the genus
Stylocellus have been described from caves in Sarawak.
Family Troglosironidae. This family was created
for the troglobitic species Troglosiro aelleni Juberthie in
New Caledonia.
Suborder Eupnoi. This suborder includes the
delicate long-legged harvestmen commonly found in
cave entrances. Most cavernicolous species belong to
the family Sclerosomatidae with numerous species of
Leiobunum having been recorded as trogloxenes in the
caves of Europe, Canada, Mexico, and the United States.
Suborder Dyspnoi. This small suborder includes
troglobites in three families.
Family Ischyropsalidae. This monotypic family
from Europe includes a large number of troglophiles
and troglobites in the genus Ischryopsalis. Ten species
are troglobitic. Some species are notable for occurring
in caves with a temperature as low as 1.5°C.
Family Nemastomatidae. Troglobitic species in the
following genera have been described: Buresiola (four
species from Europe and Tunisia), Nemaspela (two
species from Europe), Mitostoma (two species from
Europe) and Nemastoma (three species from Europe
and the United States). Ortholasma sbordonii Brignoli is
a possible troglobite in Mexico.
Family Sabaconidae. The only troglobite is Sabacon
picosantrum Martens from caves at high altitude in
Spain. A second species is a troglophile in the caves of
France.
Suborder Laniatores. This suborder occurs world¬
wide but is particularly abundant in the tropics and
subtropics. The taxonomy of the group is still poorly
known with the limitations of some families not yet
settled. Many species from the tropics are large and
heavily spinose.
Family Agoristenidae. Two cavernicole species of
the family Agoristenidae have been described from
caves in Venezuela. Phalangozea bordoni Munoz-Cuevas
is a troglobite.
Family Assamiidae. This family is restricted to trop¬
ical Africa and Asia. Blind, depigmented species have
been described from forest litter and termite nests, but
some from caves appear to be troglobitic and troglo-
plrilic. The following genera include species of troglo¬
bites: Typhlobnnellus from Tanzania, Sijncavernicus and
Metassamia from Assam, and Calloristus from India.
Family Biantidae. The only troglobite in this family
is Decuella cubaorientalis Avram from Cuba. Other spe¬
cies from Cuba and Ceylon are probably troglophiles.
Family Cladonychiidae. This Elolarctic family con¬
tains a few troglophiles in Europe. Speleomaster lexi
Briggs and S. pecki Briggs are troglobites from lava tubes
in the northwestern United States. Erebomaster flavescens
Cope includes two subspecies of troglophile and one of
troglobite from caves in the eastern United States.
Family Gonyleptidae. This neotropical family
includes numerous cavernicolous species, but the only
apparent troglobite is Vima chapmani Rambla from
Venezuela. Genera containing troglophiles or troglox¬
enes include Pachyloides from Uruguay and Venezuela;
Rhopalocranaus, Mendellina, Vima, and Santinezia from
Venezuela; Ancistrotellus from Brazil; and Aucayacuella
from Peru.
Family Phalangodidae. This family is especially
abundant in the caves of North America and Europe.
The genus Banksula occurs in the caves of California
(U.S.A.). The genus Texella includes troglobitic and trog-
lophilic species in California and Texas (U.S.A.). Other
troglobites from caves in the United States belong to the
genera Bishopella, Calicina, Crosbyella, Goodnightiella,
Phalangodes, and Phalangomma. Two troglobites in the
genus Guerrobunus are known from Mexico. The genus
Jimeneziella includes two troglobitic species in Cuba.
Panopiliops inops Goodnight and Goodnight is a troglo¬
bite in Costa Rica. The genus Scotolemon includes sev¬
eral troglobites from caves in Europe. Other troglobitic
species in Europe include representatives of the genera
ENCYCLOPEDIA OF CAVES
SPRINGS
797
Lola and Paralola. The only troglobitic member of the
family from Africa is Conomma troglodytes.
Family Stygnommatidae. The genus Stygnomma
includes troglobites and troglophiles from Mexico,
Belize, and the Galapagos Islands.
Family Stygnopsidae. This family includes numerous
cavernicolous species, many undescribed, in the caves
of Mexico. Troglobites include representatives of the
genera Hoplobunus, Mexotroglimis, and Troglostygnopsis.
Troglophiles are found in the genera Hoplobunus,
Karos, Sbordonia, and Stygnopsis. Two species of
Chinquipellobunus are troglobites in Texas (U.S.A.). South
American troglobites include Pachyspeleus strinatii from
Brazil and Galanomma microphthalma Juberthie from the
Galapagos Islands.
Family Travuniidae. This family includes troglo¬
bites from caves in Europe, the United States, and
Japan. The genus Peltonychia includes numerous spe¬
cies with troglomorphic adaptations ranging from
eyed to totally eyeless from caves in Europe. Other
troglobitic European species occur in the genera
Dinaria and Tmvunia. The remaining troglobitic travu-
niid species are Speleonychia sengeri Briggs from lava
tubes in Washington (U.S.A.) and Yuria pulcra Suzuki
from Japan.
Family Triaenonychidae. This family is largely tropi¬
cal but a few species are known from the United States.
The genus Hendea includes several species of troglobite
and troglophile in New Zealand. The Australian troglo¬
bite fauna include species of the genera Calliuncus and
Holonuncia. The cavernicole fauna of Tasmania include
species of troglobitic Hickmanoxyomma, Lomanella, and
Picunchenops. Other troglobitic genera containing troglo¬
bites include Spelaeomontia in South Africa, Picunchenops
in Argentina, and Cryptobunus in the northwestern
United States.
Family Zalmoxidae. The genus Ethobunus includes
troglobites from Belize and Jamaica.
Bibliography
Conde, B. (1998). Palpigradida. In C. Juberthie & V. Decu (Eds.),
Encyclopaedia biospeologica (Vol. II, pp. 913—920). Moulis, France:
Societe de Biospeologie.
Dusbabek, F. (1998). Acari parasiti. In C. Juberthie & V. Decu (Eds.),
Encyclopaedia biospeologica (Vol. II, pp. 921—928). Moulis, France:
Societe de Biospeologie.
Heurtault, J. (1994). Pseudoscorpions. In C. Juberthie & V. Decu
(Eds.), Encyclopaedia biospeologica (Vol. I, pp. 185—196). Moulis,
France: Societe de Biospeologie.
Juberthie, C. (1994). Ricinulei. In C. Juberthie & V. Decu (Eds.),
Encyclopaedia biospeologica (Vol. I, pp. 231—235). Moulis, France:
Societe de Biospeologie.
Palacios-Vargas, J. G., Decu, V., Iavorski, V., Hutzu, M., & Juberthie,
C. (1998). Acari terrestria. In C. Juberthie & V. Decu (Eds.),
Encyclopaedia biospeologica (Vol. II, pp. 929—952). Moulis, France:
Societe de Biospeologie.
Rambla, M., & Juberthie, C. (1994). Opiliones. In C. Juberthie &
V. Decu (Eds.), Encyclopaedia biospeologica (pp. 215—230). Moulis,
France: Societe de Biospeologie.
Reddell, J. R., & Cokendolpher, J. C. (1995). Catalogue, bibliography,
and generic revision of the order Schizomida (Arachnida). Austin:
Texas Memorial Museum.
Ribera, C., & Juberthie, C. (1994). Araneae. In C. Juberthie & V. Decu
(Eds.), Encyclopaedia biospeologica (Vol. II, pp. 197—214). Moulis,
France: Societe de Biospeologie.
Savory, T. (1977). Arachnida (2nd ed.). New York: Academic Press.
Schwarz, A. E., Schwoerbel, J., & Gruia, M. (1998). Hydracarina.
In C. Juberthie & V. Decu (Eds.), Encyclopaedia biospeologica
(Vol. II, pp. 953—976). Moulis, France: Societe de Biospeologie.
Volschenk, E. S., & Prendini, V. (2008). Aops oncodactylus, gen. et sp.
nov., the first troglobitic urodacid (Urodacidae: Scorpiones), with
a re-assessment of cavernicolous, troglobitic and troglomorphic
scorpions. Invertebrate Systematics, 22, 235—257.
Weygoldt, P. (1994). Amblypygi. In C. Juberthie & V. Decu (Eds.),
Encyclopaedia biospeologica (Vol. I, pp. 241—247). Moulis, France:
Societe de Biospeologie.
SPRINGS
William B. White
The Pennsylvania State University
INTRODUCTION
A spring is a localized orifice where groundwater
returns to the surface. All aquifers must ultimately
drain somewhere. If the drainage occurs at a single loca¬
tion, as distinguished from a line of seeps, a wetland, or
discharge into the bed of a surface stream, the place
where the groundwater appears is called a spring. The
sizes of springs vary greatly from barely perceptible
trickles to full-sized rivers gushing forth from the earth.
Springs have been of immense importance to
humankind throughout history. Most have been used
as water supplies and most farmsteads were sited
next to a spring which provided water for both house¬
hold and livestock. Indeed, in arid regions, springs
may have been the only water supply. The city of
Jerusalem was founded at the site of the sacred Gihon
Spring about 3800 years ago (Amiel et al., 2010). Many
other towns were located at springs so that the water
supply would be safely inside the city walls. Hot
springs and mineral springs have been valued for their
curative properties with many health resorts con¬
structed around springs. Many of the largest springs
come from cave passages or from fractures in carbon¬
ate rock and are known as karst springs. Large
springs are also found in volcanic rocks and springs
of various sizes in nearly all rocks. This article is con¬
cerned only with karst springs. For a comprehensive
discussion of springs including descriptions of
many individual springs, see Kresic and Stevanovic
( 2010 ).
Encyclopedia of Caves.
© 2012 Elsevier Inc. All rights reserved.
798
s
TYPES OF SPRINGS
Springs can be categorized and described in terms of
the nature of the spring orifice and by the characteristics
of the discharge (Fig. 1).
usually remains clear. The temperature and chemical
composition of the water remain constant throughout
the year. Some very large springs, those in Florida, for
example, do drain from conduits but show little storm
response.
Conduit and Diffuse Flow Springs
Springs that drain from cave passages (conduits)
have distinctive characteristics. They tend to be flashy;
that is, flow increases rapidly in response to storms in
the watershed. The springs often become turbid, some¬
times muddy, in response to storms. The chemistry of
the spring water also varies with seasons and in
response to storms. These springs are known as conduit
springs even if the conduit that feeds the spring is not
visible at the spring orifice. Other springs, called diffuse
flow springs, drain from faults, networks of fractures, or
other localized zones of high permeability. These
springs tend to flow at a constant rate with at most
only sluggish response to storms, although the flow
may rise and fall with wet and dry seasons. The water
Gravity Springs
In a sense, all springs are gravity springs because it
is the action of gravity that causes water to drain from
storage in the aquifer. However, the term gravity spring
is reserved for springs that represent the emergence of
an underground stream from within the cave system
that feeds the spring. Some gravity springs are open
cave mouths (Fig. 2). One can enter the cave by simply
walking (or crawling) into the spring mouth. In a
few of these springs, one can follow the cave stream
upstream for considerable distances; in others the
underground stream is found rising from a sump at
varying distances upstream. In other cases, the spring
mouth is obscured by surface debris that has slumped
across the passage. In these cases, the spring water
FIGURE 1 Sketches of various types of springs: (A)
Gravity spring draining from an open cave mouth. (B)
Gravity spring draining through rubble. (C) Alluviated
spring. (D) Shallow artesian spring. Water rises from
flooded cave passages at depths of a few meters to a
few tens of meters. (E) Deep artesian spring. Water
rises along a deep channel in the bedrock. Depths may
be hundreds of meters.
ENCYCLOPEDIA OF CAVES
SPRINGS
799
FIGURE 2 Overholt Blowing Cave Spring, Pocahontas County,
West Virginia, a gravity spring draining from an open cave conduit.
Photograph taken by the author.
emerges from hillside rubble with no obvious cave
passage.
Alluviated Springs
Spring mouths are frequently located on the banks
of rivers or large creeks. Over time, the beds of the riv¬
ers and creeks accumulate thick deposits of sand, silt,
and clay, materials known as alluvium. The sediments
may block the spring mouths, thus forcing the water to
rise through a channel between the hillside and the
alluvial river sediments (Fig. 3). Such blockages have
occurred in the Mammoth Cave (Kentucky) area
caused by changes in the Ohio River and its tributaries
during glacial and postglacial periods. Similar springs
are found in New York where spring mouths have
been choked with glacial drift. The spring orifices of
alluviated springs are rise pools, sometimes called blue
holes, where water wells upward and then flows out
to the surface stream. Divers have explored some allu¬
viated springs and found the flooded cave passage
feeding the spring at depths of 5—10 m.
Offshore Springs
During periods when the glaciers of the Pleistocene
ice ages had their maximum advance, so much water
was stored in the glaciers that sea levels were lowered
by as much as 100 m. Cave systems discharged their
water through springs located near the shoreline.
When the glaciers retreated and sea levels rose again,
these springs were drowned but continued to dis¬
charge freshwater. Freshwater is less dense than sea¬
water. As a result, the location of drowned offshore
springs is marked by a plume of freshwater rising to
FIGURE 3 An alluviated spring in Mammoth Cave National
Park, Kentucky. Photograph taken by the author.
the surface, sometimes with a pronounced boil.
Offshore springs are found in many coastal areas
where carbonate rocks extend below sea level.
Artesian Springs
Artesian springs are those in which the water is
forced upward under pressure from channels in the
bedrock. Some artesian springs are shallow, fed by
cave passages below the water table. Shallow artesian
springs sometimes well up with a pronounced boil,
but with water coming from depths of only a few tens
of meters. Many cave streams end in sumps from
which the connection to the spring is through flooded
passages. Many shallow artesian springs have been
penetrated by divers who re-emerged into air-filled
caves after relatively short distances underwater.
Deep artesian springs are fed by channels that extend
to depths of hundreds of meters. Usually these channels
are guided by geological structures such as major frac¬
ture systems or faults. These springs are sometimes
known as Vauclusian springs, named after the Fontaine
de Vaucluse in southern France. Exploration by divers
and remote sensing equipment has established the
depth of the Fontaine de Vaucluse at 308 m. Deep
artesian springs are found along the Sierra de El Abra
in Tamaulipas, Mexico. The Nacimiento del Rio
Mante was explored to 252 m without reaching bottom.
ENCYCLOPEDIA OF CAVES
800
s
A recent discovery is Cenote de Zacaton, also in
Tamaulipas, Mexico, with an explored depth of 329 m
(Gary, 2010). Exploration of these deep underwater sys¬
tems is at (or beyond) the limits of diving techniques
and as a result little is known about the feeder system
of these springs.
Overflow, Underflow, and Distributary Springs
Springs may have more than one orifice. During dry
season, only the orifice at the lowest elevation, the
underflow spring, may be flowing. Underflow springs
are often low on the banks of surface streams or in the
channel itself and may not be obvious to observers.
When flow through the aquifer increases during wet
seasons or in response to storms, other orifices at high¬
er elevation may become active. Overflow springs may
appear as dry cave entrances during dry seasons and
may discharge a large volume of water during wet sea¬
sons. The Planina Cave in Slovenia is among the larg¬
est springs in Europe, discharging the combined flow
of the Pivka and Rak rivers. It is, however, the over¬
flow spring and the discharge varies by several orders
of magnitude. It is the more constant underflow
spring, the Malenscica, that is used for municipal
water supplies.
Cave systems sometimes reach the surface as a sin¬
gle conduit leading to a single spring. However, it is
common to find downstream distributaries where the
passages split into multiple branches each of which
may reach the surface to produce a spring. Distributary
springs have been observed scattered along several
kilometers of surface stream, all discharging water
from the same karst groundwater basin. Distributary
springs are usually not at exactly the same elevation.
As a result, only one or a few springs may be flowing
during the dry season but many more become active
during storm flow or during the wet season.
Thermal Springs
The water temperature in springs draining shallow
groundwater basins is typically the same as the mean
annual temperature of the region, although water tem¬
perature from some open and fast-flowing conduit sys¬
tems fluctuates with the seasons. The water is warmer
than average in the summer and colder than average
in the winter. The transit of water from inlet to spring
in these very open systems takes place faster than the
water can come into thermal equilibrium with the sur¬
rounding rock. The seasonal average, however, is close
to the mean annual temperature.
There exist limestone springs that discharge water
at temperatures higher than the local average
temperature (Goldscheider et al., 2010). These thermal
springs are called warm springs or hot springs
depending on the temperature. Not all thermal springs
are limestone springs. Hot springs are common in vol¬
canic areas where groundwater is heated by volcanic
activity. In karstic regions, the source of heat is usually
geothermal. The temperature of the earth increases
with depth at rates ranging from 15 to 40 C km If
the groundwater circulation path takes the water to
great depths before reaching the spring, the water tem¬
perature will be increased. Thermal springs draining
from karst aquifers have observed temperatures from
just above ambient to as high as 70°C, although most
springs are somewhat cooler.
Mineral Springs
Springs can also be categorized by the chemical
composition of their waters. Water from a typical karst
spring will contain mainly Ca 2+ , Mg 2+ , and HCO 3
ions with minor concentrations of Na + , K + , S0 4 2 ~,
Cl , NO 3 , and possibly a few other ions. Total ion
concentrations are typically in the range of a few hun¬
dred milligrams per liter. Mineral springs are those
that contain much higher concentrations of dissolved
substances, sufficient to give the water a strong taste
and possibly an odor. Mineral springs have a great
range of both chemical composition and concentrations
of dissolved substances and may be broadly subdi¬
vided according to the dominant chemistry. Saline
springs contain high concentrations of NaCl and other
dissolved salts of magnesium, calcium, and sodium,
with magnesium sulfate often being an important con¬
stituent. Sulfur springs contain dissolved hydrogen
sulfide, H 2 S, which gives the water a rotten egg odor.
Iron springs, also called chalybeate springs, have high
iron concentrations. Mineral springs may or may not
also be thermal springs. Mineral springs were prized
for their medicinal properties, particularly in the 19th
century. Resort hotels and baths were built on the sites
of many of these springs, although only a few remain
in operation.
Seeps and Wetlands
Marshy areas often result from the discharge of
groundwater from any type of aquifer. Karst aquifers
in which the movement of water is through networks of
fractures may return water to the surface through an
extended wetland region rather than through a single
large spring. Other karst regions may be mantled with
glacial drift, volcanic ash, or other recent permeable
cover, thus spreading and diffusing the groundwater
discharge. Seeps and wetlands are found where the
ENCYCLOPEDIA OF CAVES
SPRINGS
karstic carbonate rocks are underlain by impenetrable
layers. Water may emerge from the bottom of the aqui¬
fer along the line of contact with the impermeable layer
as a continuous zone of small seeps and springs rather
than as one large spring.
SPRING DISCHARGES
Magnitudes of Spring Discharges
The volume of water discharged from karst springs
varies by many orders of magnitude. Table 1 lists a
selection of some of the largest karst springs of the
world. Where data were available, maximum, mini¬
mum, and mean discharges are listed. Some spring
discharges vary only slightly about the mean; others
vary by several orders of magnitude between base
flow and peak flow. None of the largest karst springs
are in the United States. The largest reported springs
are in Papua New Guinea where several large rivers
flow directly from caves. For the United States, large
springs occur in Florida, Missouri, and Texas (Table 2).
Although there are hundreds (probably thousands) of
karst springs in the eastern United States, many of the
Appalachian groundwater basins are relatively small
as are their springs. They provide spring discharges
typically an order of magnitude smaller than the large
springs listed in Table 2.
801
A plot of discharge as a function of time is known
as a hydrograph. Hydrographs are routinely measured
for surface streams and provide a graphic picture of
the variations in stream discharge with seasons and
with storm flow. Hydrographs can also be measured
for springs and provide some insight into the character¬
istics of the groundwater basin and the conduit system
that feeds the spring. Figure 4 shows two examples.
The hydrograph of Alley Spring contains many sharp
peaks, which represent the spring response to individual
storms. The hydrograph of Rainbow Spring is smooth.
The contributions of individual storms have been
averaged out and only the rise and fall of discharge
with seasonal wet and dry periods is reflected in
the hydrograph. Aquifers with well-developed conduit
systems drain rapidly. Little water is retained in stor¬
age. Spring discharge has time to fall to base flow
before the next storm arrives. The ratio of peak
flow to base flow gives a measure of the "flashiness"
of storm response. For Alley Spring, this parameter
is 4.7. For Rainbow Spring, it is 1.2. Ratios as high
as 100 have been observed for Davis Spring, West
Virginia.
Ebb-and-Flow Springs
Ebb-and-flow springs are those in which the dis¬
charge varies in a cyclical fashion. Discharge drops to
some minimum value, then rises rapidly to a peak, and
TABLE 1 Some Representative Large Karst Springs Excluding the United States
Spring
Location
Maximum Flow
Mean Flow
Minimum Flow
Tobio
New Guinea
—
85-115
-
Matali
New Guinea
>240
90
20
Dumanli
Turkey
—
50
25
Buna
Bosnia
440
40
2
Ljubljana
Slovenia
132
39
4.2
Ras el Ain
Syria
—
39
-
Chingshui
China
390
33
4
Vaucluse
France
200
29
4.5
Frio
Mexico
515
28
6
Coy
Mexico
200
24
13
Timava
Italy
138
17.4
9
Niangziguan
China
15.8
12.1
8.1
Aachquelle
Germany
24.1
8.2
1.3
Blautopf
Germany
26.1
2.2
0.35
Data from various sources, especially Ford and Williams (2007) and Kresic and Stevanovic (2010). All flow measurements are in m 3 s 1
ENCYCLOPEDIA OF CAVES
802
TABLE 2 A Selection of Large Karst springs in
S
the United States
Spring
Location
Maximum Flow
Mean Flow
Minimum Flow
Silver Spring
Florida
36.3
23.0
15.3
Rainbow Spring
Florida
28.9
19.8
13.8
Wakulla Spring
Florida
54.1
11.1
0.71
Comal Springs
Texas
15.1
9.2
—
San Marcos Spring
Texas
8.5
4.2
2.3
Big Spring
Missouri
36.8
12.1
6.7
Greer Spring
Missouri
25.5
9.4
2.9
Maramec Spring
Missouri
18.4
4.2
1.6
Alley Spring
Missouri
—
3.6
1.5
Davis Spring
West Virginia
28.3
3.1
0.6
Data from various sources. All flow measurements are in m 3 s 1
FIGURE 4 Hydrographs, each drawn for a single water year from
U.S. Geological Survey water records for Alley Spring, Missouri, and
Rainbow Spring, Florida.
then decreases again. Usually this flow-and-flush cycle
occurs on a timescale of minutes to hours. Ebb-and-
flow behavior has been ascribed to a siphon action in
the conduit system that feeds the spring. The deeper
part of the conduit slowly fills with water until the level
reaches a spillover route or the pressure head becomes
sufficient to force water out of the system. Draining
through the spillover route triggers a siphon action so
that the lower, flooded parts of the conduit are also
drained. Once the conduit system has been drained,
siphon action ceases and the system slowly fills to begin
the next cycle. One of the best studied examples is Big
Spring, draining the Lilburn Cave System in Kings
Canyon National Park, California, U.S.A. Detailed anal¬
yses of hydrographs measured at the spring and in the
lowest levels of the cave substantiate a forced siphon
mechanism.
Analysis of Spring Hydrographs
Some information concerning the groundwater
basin that provides the spring discharge can be
obtained from a more careful analysis of the spring
hydrograph. Figure 5 shows a schematic hydrograph
of a single storm to illustrate the various features. If
there has been no rain in the watershed for a long
time, the flow from the spring will decrease to a mini¬
mum value called the base flow, Q B (of course, if the
spring completely dries up during droughts, the base
flow is zero). The storm precipitation is represented by
the bars on the left side of the diagram (drawn upside
down). After a period of time, the storm lag, the flow
of the spring will begin to increase and quickly rise to
a peak value, Q max . The rising limb of the hydrograph
is generally very steep. After the storm passes, the
flow from the spring will begin to decrease, but the
rate of decrease is slower than the rate of rise.
Often the recession limb of the hydrograph can be
represented by an exponential function with two
adjustable parameters, Q 0 and r. The parameter r has
units of time and is a measure of the response time of
the aquifer. Exponential fits to karst spring recession
curves often produce two or more segments. The fast
response segment, with r on the order of 3—10 days,
may represent the draining of the open conduits, while
the slow response segment, with r on the order of
20—50 days, may represent the draining of fractures
and other small tributaries. Sometimes spring hydro¬
graphs have multiple segments or do not fit exponen¬
tial equations at all. These responses do not lend
themselves to straightforward hydrologic interpreta¬
tion. More detailed analysis of spring hydrographs can
provide considerable information about the aquifer
feeding the spring (Geyer et al., 2008).
ENCYCLOPEDIA OF CAVES
SPRINGS
803
FIGURE 5 Hydrograph for Hurricane Francis at Arch Spring, Blair County, Pennsylvania. Solid line is the water level at the spring;
individual points are measured sediment load. Upper figure shows ten-day record; lower figure is expanded three-day record. From Herman
et al. (2008).
Turbidity and Chemical Variability
Flow velocities in the conduit system of karst aqui¬
fers, especially during storm flow, are often sufficient
to take clays, silts, and other small particles into sus¬
pension and carry the suspended particles to the
spring. As a result, some springs become turbid or
muddy following storms with the water gradually
becoming clear as the discharge decreases back toward
base flow conditions (Fig. 5). The presence of muddy
water during storm flow is often an indication that the
groundwater basin is recharged by sinking streams
or open sinkholes that can carry sediment into the
system.
ENCYCLOPEDIA OF CAVES
804
s
FIGURE 6 Sketch of a single storm peak on a spring hydrograph
showing the rapid rising limb and the more drawn out recession limb.
The chemical composition of springs draining from
fracture systems (diffuse flow springs) tends to remain
constant throughout the year regardless of storms
and wet and dry periods. The chemical composition of
springs draining from conduit systems tends to be
much more variable. The variability in such compo¬
nents as Ca 2+ , Mg 2+ , and HCO 3 is due to dilution of
groundwater by storm water, and by changes in
recharge chemistry due to winter/summer changes
in vegetation. The bulk chemistry of the spring water
can be easily monitored by measuring the electrical
conductivity. This produces a curve called a chemograph,
which can be superimposed on the hydrograph as indi¬
cated in Figure 6 . The dip in the chemograph indicates
dilution due to the input of storm water. In Figure 6 ,
the dip in the chemograph is shown offset from the
peak in the hydrograph by a time period called the
chemical lag. If the chemical lag is zero, that is, if the dip
in the chemograph is directly superimposed on the
peak of the hydrograph, it indicates that the storm
water has moved directly to the spring. In this case, the
storm lag is a rough measure of the travel time between
the storm water input points and the spring. If the dip
in the chemograph is offset, it indicates that the conduit
system is at least partially flooded. Rising water levels
at the upstream end of the system force water out of
the flooded conduits to produce the peak in the hydro¬
graph and it is only later that the actual storm water
reaches the spring.
Chemographs can be constructed for individual
chemical constituents if enough samples are collected
and analyzed during the storm flow. These may also
be interpreted in terms of the movement of various
constituents through the system.
Saturation State of Spring Waters
From a chemical analysis of spring water, it is possi¬
ble to calculate how the measured concentrations of
dissolved carbonate minerals deviate from chemical
equilibrium. Springs fed by open conduit systems
receiving a considerable fracture of surface water can
then be undersaturated with respect to carbonate
minerals. The water moved through the aquifer in less
time than the time required for reaction with the bed¬
rock. Springs discharging from fracture aquifer tend to
be close to chemical equilibrium. Water from aquifers
that are recharged at high CO 2 partial pressures can
discharge the excess CO 2 at or just below the spring
mouth, resulting in the precipitation of the form of
calcium carbonate known as tufa.
SPRINGS AS WATER SUPPLIES
Water gushing from the earth has seemed an obvious
source of freshwater for human consumption for millen¬
nia. Villages and towns have grown up around springs
because of the convenient supply of high-quality water.
As a result, there is a popular myth that spring water is
intrinsically pure. Indeed, spring water is often chosen
by water bottling companies (LaMoreaux and Tanner,
2001). In the case of karst springs, the myth of pure
spring water is exceedingly dangerous.
Karst springs frequently receive their water from
cave passages, which in turn receive their water
from surface streams that sink underground or from
storm flow into sinkholes. Sinkholes are often reposito¬
ries for household garbage and farm waste such as
empty pesticide cans, other trash, and dead animals.
Sinking surface streams and sinkholes may admit
spilled petroleum hydrocarbons (gasoline, diesel fuel,
and home heating oil) as well as industrial solvents
such as perchloroethylene (PCE) and trichloroethylene
(TCE) and many others. Springs may be impacted by
discharge from mining activities in the catchment area.
Any spring that becomes turbid after storms or which
shows a high degree of chemical variability should be
immediately suspected. Even fracture flow springs
with clear water and constant chemistry may require
extensive testing as well as filtration before being
approved as public water supplies. Regulations now
often require an elaborate test for surface water influ¬
ence. Minor debris such as bits of plant material or
insect parts is sufficient to require a filter plant.
SPRINGS AS HABITAT
Springs provide a stable habitat for a variety of
aquatic organisms. Spring waters, with their nearly
ENCYCLOPEDIA OF CAVES
STALACTITES AND STALAGMITES
805
constant temperature, remain warm and ice-free dur¬
ing winters. Springs are a transitional environment
between true cave conditions upstream and normal
surface stream conditions downstream. The tempera¬
ture and water chemistry of springs are comparable to
cave stream conditions, but springs are also under the
influence of sunlight. Aquatic plants can grow in
springs and provide sources of nutrient that are not
available in the cave system. However, springs are
generally more sheltered and subject to smaller swings
in microclimate that are completely open surface
streams.
See Also the Following Article
Hydrogeology of Karst Aquifers
Bibliography
Amiel, R. B., Grodek, T., & Frumkin, A. (2010). Characterization of
the hydrogeology of the sacred Gihon Spring, Jerusalem: A dete¬
riorating urban karst spring. Hydrogeology Journal, 18, 1465—1479.
Ford, D., & Williams, P. (2007). Karst hydrogeology and geomorphology.
Chicheser, UK: John Wiley & Sons.
Gary, M. O. (2010). Karst hydrogeology and speleogenesis of Sistema
Zacaton. Association for Mexican Cave Studies Bulletin, 21, 1—114.
Geyer, T., Birk, S., Liedl, R., & Sauter, M. (2008). Quantification of
temporal distribution of recharge in karst systems from spring
hydrographs. Journal of Hydrology, 348, 452—463.
Goldscheider, N., Madl-Szonyi, J., Er5ss, A., & Schill, E. (2010).
Review: Thermal water resources in carbonate rock aquifers.
Hydrogeology Journal, 18, 1303—1318.
Flerman, E. K., Toran, L., & White, W. B. (2008). Threshold events in
spring discharge: Evidence from sediment and continuous water
level measurement. Journal of Hydrology, 351, 98—106.
Kresic, N., & Stevanovic, Z. (Eds.), (2010). Groundwater hydrology of
springs. Amsterdam: Elsevier.
LaMoreaux, P. E., Ik Tanner, J. T. (Eds.), (2001). Springs and bottled
waters of the world. Berlin: Springer-Verlag.
STALACTITES AND STALAGMITES
Silvia Frisia and Jon D. Woodhead'
The University of Newcastle, Australia,
he University of Melbourne, Australia
INTRODUCTION
Stalactites and stalagmites are the most common
speleothems, the morphology of which is basically
controlled by dripping; therefore, both speleothems
can be considered as gravitational forms. Stalactites are
centimeter to meter in scale, hanging from the ceiling
and growing toward the cave floor. Stalagmites grow
from the cave floor upward and are commonly fed by
water dripping from an overhead stalactite (Fig. 1).
Stalagmitic flowstones are a particular type of stalagmite
formed by a thin flowing film of water itself fed by
groups of dripping stalactites, and coat the cave floor
and walls. When a stalagmite and the overhanging
stalactite merge, they form a column (Fig. 1). Most
stalactites and stalagmites are composed of calcite, a
few of aragonite, the rhombohedral and orthorhombic
phases of calcium carbonate (CaC0 3 ), respectively.
Rare stalactites and stalagmites consisting of huntite (a
Mg-carbonate), halite (NaCl), gypsum (CaS0 4 ■ 2H 2 0),
and even opal (amorphous hydrated Si0 2 ) have been
found.
Stalactites and stalagmites likely started to develop
in caves when the first carbonate rocks had been
subaerially exposed and eroded well over 1 billion
years ago. Most speleothems that have been exten¬
sively studied date from the Quaternary, and the gene¬
sis of these is commonly driven by the process of
degassing, which occurs when drip waters having a
high carbon dioxide concentration (pC0 2 ) interact with
the cave atmosphere that has a relatively low pC0 2 . It
is, therefore, believed that occurrence of stalagmites
and stalactites greatly increased since the rise of vascu¬
lar plants in the Devonian, which led to an acceleration
of chemical weathering, greater availability of soil
C0 2 , and a decline in global atmospheric C0 2 concen¬
tration (Alonso-Zarza and Tanner, 2010). Chemical
weathering of Ca-bearing silicate minerals by acidic
waters generated in peat soil, for example, is very
effective in yielding calcite stalactites and stalagmites
in caves cut into granite and gneiss; in such cases
karst dissolution does not play a role in the genesis of
these speleothems, but the presence of vascular plants
does. The focus of the following sections is, therefore,
on the genesis, structure, and chemical properties of
calcium carbonate stalagmites and stalactites since the
Devonian.
MINERALOGY AND PROCESSES
OF FORMATION
Carbonate stalactites and stalagmites are commonly
composed of calcite, the calcium carbonate (CaC0 3 )
phase that is thermodynamically stable at surface
temperature (T) and pressure (P). They may be also
composed of aragonite (Fig. 2A), the high pressure
polymorph of CaC0 3 , thermodynamically unstable at
surface P and T. The formation of aragonite in caves is
most probably related to a combination of factors,
some of which are well understood, others more spec¬
ulative. Aragonite precipitation in stalactites and sta¬
lagmites appears to be favored by high Mg/Ca ratio of
the parent water, high pH, presence of ions and
Encyclopedia of Caves.
© 2012 Elsevier Inc. All rights reserved.
806
s
FIGURE 1 Stalactites, both soda straws and cone stalactites, candle¬
shaped stalagmites, columns (stalactites and stalagmites merged), and
stalagmitic flowstone coating the cave floor. The plastic containers host
glasses onto which in situ calcite precipitation experiments have been
carried out to determine the processes that influence the development
of different crystals.
FIGURE 2 (A) Stalagmite consisting of aragonite layers (A) capped
by calcite layers (C). (B) Backscattered electron diffraction image of cal¬
cite (colored in red) replacing aragonite crystals (colored in blue).
Photo courtesy of I. }. Fairchild.
complexes in solution that may inhibit calcite growth
sites, low drip rate, and forced degassing (which, by
promoting degassing, increases the pH of the solution).
The mineralogical composition of a stalagmite need
not be the same as that of the feeding stalactite. It is
not uncommon, in fact, to observe calcite stalactites
feeding aragonite stalagmites. Aragonite, however,
being thermodynamically unstable, could dissolve and
be replaced by calcite through dissolution and repreci¬
pitation, yielding diagenetically modified stalagmites
and stalactites (Fig. 2B).
The precipitation of calcium carbonate in caves
occurs from solutions that commonly have low ionic
strength and reach supersaturation with respect to cal¬
cite (or aragonite). The capability of a solution to pre¬
cipitate calcite is commonly measured by its Saturation
Index (SI):
SIcalcite = log {(Ca 2+ )(C0 2 ~ )/K ca ldte}
where (Ca 2 XCO 3 ) is the ion activity product and
Kcaicite is the temperature-dependent solubility constant
for calcite. Calcite will precipitate when Sl ca i c i te >0, and
dissolve when SI ca i cite <0.
Monitoring studies have shown that active calcite
growth may occur in stalagmites even if the initial
drip waters are barely at saturation, which is the
case for most high-altitude, high-latitude caves. The
mechanism producing supersaturation of an already
saturated solution and resultant precipitation of calcite
in caves is C 02 -degassing, a set of reactions that
involve the loss of carbon dioxide from the drop to
the cave atmosphere. The precipitation of calcite (or
aragonite) is described by the following equilibrium
reaction:
Ca 2+ + 2HC0 3 o CaC0 3 + C0 2 + H 2 0
If the Ca 2+ ion concentration in the parent solution
does not increase, the loss of C0 2 from the solution
through degassing shifts the equilibrium reaction to the
left, and calcite (or aragonite) precipitates. Surface area
and timing for C0 2 degassing varies for stalactites and
stalagmites, and is dependent on drip rate: the slower
the drip rate, the longer the degassing process may
proceed. Degassing may also be forced by kinetic (irre¬
versible) processes, in particular, in ventilated caves
where the pC0 2 of the cave air may attain values simi¬
lar to those of surface air (slightly less than 400 ppm),
while drips may have a C0 2 concentration at least
ten times higher. Cave ventilation modulated by the
temperature difference between the cave (warm in
winter and cool in summer) and the surface (cold
in winter and warm in summer) may, thus, force
stalagmite and stalactite growth in alpine and high-
latitude cold caves, where calcite precipitation is not
favored given that the solubility of calcite (in this
case, that of the host rock) decreases with increasing
temperature, and soil C0 2 production is restricted to a
short, warm season.
ENCYCLOPEDIA OF CAVES
STALACTITES AND STALAGMITES
807
STALACTITES
The most common variety of stalactite is the tubular
soda straw, consisting of thin translucent, tubular
layers of calcite crystals surrounding a central canal.
Soda straws are speleothems generated by axial feed¬
ing, and characterized by a central canal with constant
diameter and by a wall structure that is controlled by
geometric selection during growth on the meniscus of
a drip. The term soda straw, therefore, is not to be
applied to any stalactite with a central, hollow tube.
Commonly, the outer diameter of a soda straw
varies from 5 to 10 mm, and the diameter of the inner
canal ranges from 2 up to 6 mm. Scanning electron
microscopy (SEM) observations indicate that soda
straws grow by linear, downward accretion of coales-
cent rhombohedra crystallites with their c-axis (the
threefold symmetry axis of the rhombohedral unit cell)
dimension nearly vertically oriented. When a droplet
of water at saturation with respect to calcium carbon¬
ate flows out of pores or fissures at the roof of the cave
it degasses and deposits a small rim of calcite (or, less
commonly, aragonite), which has the diameter of the
droplet. Successive layers of crystals then deposit on
the previous layer, maintaining the original diameter.
Soda straws, thus, commonly show growth bands typi¬
cally 0.05 to 0.5 mm thick, which is the linear annual
extension rate expected for these speleothems. Growth
bands identify a periodic thickening of the inner
channel wall. Field observations and sampling reveal
that the dimension of the central tube is a function of
drip rate. In some cases, air is present within the
channel and degassing occurs within the straw,
causing calcite precipitation and, eventually, its com¬
plete obstruction. The straw then would not be
capable of feeding an underlying stalagmite, unless
rupture followed by development of a new straw at
the same drip point occurs. Aragonite soda straws are
less common and consist of acicular crystals.
Conical stalactites characteristically taper down¬
ward, and may be several tens of centimeters wide.
Their length can reach the meter-scale and, eventually,
the stalactite coalesces with the underlying stalagmite.
Conical stalactites consist of concentric layers of crys¬
tals elongated perpendicular to a central growth axis
(the axis extending from the ceiling toward the floor of
the cave), which may be hollow to allow the passage
of water (Fig. 3). With time, the center of the stalactite
may be clogged by calcite and once the central conduit
is closed, water flows along the outer surface of the
speleothem. A conical stalactite, thus, grows through
deposition of crystals in radial layers developing on
the sides of a remnant soda straw, and spreading out¬
ward. The crystals of the radial layers have their c-axes
FIGURE 3 Thin section perpendicular to the vertical growth axis of
a cone stalactite showing the hollow central canal and concentric layers
of columnar calcite (crossed polarizers). Base of the photo: 8 mm.
perpendicular to the original straw. Growth on a cone
stalactite, therefore, occurs both in vertical and lateral
directions.
STALAGMITES
Stalagmites are convex deposits that grow upward,
commonly fed by an overhanging stalactite, and show
flat, rounded, or slightly hollow tops, but not a central
canal. Calcite and aragonite stalagmites form through
degassing of a saturated thin film of fluid (about
0.1 mm thick), which slowly moves down from the
stalagmite tip to its flanks. The morphology of stalag¬
mites is mostly determined by the drip rate, and in
particular by the interannual variability of the drip
rate and by the distance from the tip of the feeding sta¬
lactite. Given similar SI ca i C i te , drips that show a high
interannual variability in the time elapsed between the
fall of two drops, that is, a high coefficient of variation,
produce cone-shaped stalagmites. By contrast, stalag¬
mites fed by drips that show interannual constant time
elapsed between two drops yield candle-shaped sta¬
lagmites. Development of candle-shaped stalagmites is
favored also by drips with high Sl ca i c i te , whereas, given
the same drip rate variability, a decrease in SI ca i C i te
may result in the passage from candle to cone-shaped
stalagmites (Fig. 4). Relatively high drip rates and
waters with low SI ca i C i te form stalagmitic flowstones,
where calcite precipitation is controlled by degassing
over a large surface area, possibly aided by a flow
regime that is more turbulent than the laminar flow on
top of stalagmites. The relationships between SI ca i cite
and drip rates reported in Figure 4, do not account for
ENCYCLOPEDIA OF CAVES
808
s
FIGURE 4 Relationship between stalagmite morphologies, discharge, and SI ca i cile . Stalagmite ER 77 has become cone-shaped in the last 150
years (top layer), whereas ER76 has become candle-shaped. (Redrawn from Miorandi et al., 2010).
the effect of the height of the drop. When this exceeds
2 m, there is fragmentation of the impinging drop
at the impact point with the stalagmite surface, which
tends to be flat. A great fall height may also give rise
to stalagmites whose morphology resembles stacked
dishes or cylindrical stalagmites characterized by a
sunken splash cup.
The rate of stalagmite growth was theoretically cal¬
culated to be in the range of 0.9 to 2.1 nmol min -1 for a
water film thickness of 50 pm; laboratory experiments
and observations support these theoretical calculations.
The mechanisms of crystal growth, and the presence of
ions or compounds that may poison crystal growth
sites, however, may have the overall effect of promot¬
ing or retarding stalagmite growth rates. A stalagmite
most commonly grows under interannual variable
physical and chemical parameters, which may result in
periodic drying, fluctuations in the supersaturation
state of the solution, or peaks in the numbers of
organic molecules transported by the drip water onto
the speleothem. Stalagmites, therefore, can be consid¬
ered polycrystalline aggregates (Sunagawa, 2005) and
rarely consist of single crystals. Stalagmite growth can
be explained as occurring through the "addition" of
growth units arranged in successive, well-stacked
layers, which resemble miniature strata when the
ENCYCLOPEDIA OF CAVES
STALACTITES AND STALAGMITES
809
stalagmite is cut along the vertical growth axis. The
way each subsequent growth unit attaches to the units
below and adjacent to it is commonly syntaxial, that
is, the new unit keeps the same crystallographic orien¬
tation. The composite crystals, thus, are formed of crys¬
talline calcite in perfect optical continuity, one with
respect to the other, and show uniform extinction.
There are stalagmites where the stacking of successive
growth units is not perfectly syntaxial, due to fluctua¬
tions in drip rate and degree of supersaturation (or in
the amount of impurities), which results in the devel¬
opment of diverse stalagmite fabrics ( e.g ., Frisia and
Borsato, 2010).
Columnar is the most common fabric type observed
in stalagmites consisting of calcite. Macroscopically, it
is composed of crystals >1 mm wide and >2 mm
long, elongated perpendicular to the growing spe-
leothem surface, with a length to width ratio of about
6:1, uniform extinction under crossed polarizers, and
serrated or open crystal boundaries. In the latter case,
intercrystalline pores may host fluid inclusions or
impurities. Each macroscopically visible crystal con¬
sists of the regular stacking of rhombohedra crystal¬
lites typically 50 to 100 pm wide and >100 pm long,
which are deposited during the hydrological cycle
that characterizes each cave (e.g., Miorandi et al.,
2010). Stalagmites composed of columnar fabrics are
commonly translucent unless the crystals do not fully
coalesce and there are intercrystalline voids elongated
perpendicular to the substrate. In this case, the fabric
becomes open columnar and the stalagmite calcite
has an opaque macroscopic appearance. When the
presence of impurities in the drip water or physical
phenomena perturb the regular stacking of the crystal¬
lites, the composite crystals attain a milky, porous
aspect, and show irregular crystal boundaries and pat¬
chy extinction when viewed under crossed polarizers
of a petrographic microscope. This subtype of the
columnar fabric has been termed microcrystalline
(Frisia and Borsato, 2010) and distinguished from the
columnar fabric because it has an environmental sig¬
nificance: it forms under highly variable drips rates
ranging from 30 mL to less than 0.1 mL per minute,
and in the presence of impurities (P ions, fulvic acids,
trace metals). Stacking of crystallites into composite
crystals formed by crystallographic dendrites (Jones
and Renaut, 2010) gives rise to a dendritic fabric,
which is typical of stalagmites formed near cave
entrances, subject to air currents and evaporation, or
in ventilated caves. The presence of dendritic fabric in
stalagmites indicates that they formed under the
influence of kinetic processes, and, most probably,
under highly variable drip rates (Frisia and Borsato,
2010). The role of cave bacteria in the growth of stalag¬
mites, and on their fabrics, is still largely unknown.
but it must not be excluded. Lipids related to cave
microorganisms, in fact, have been found within sta¬
lagmite calcite.
Stalagmites consisting of aragonite are commonly
characterized by acicular crystals, which have a length
to width ratio 5S>6:1, are elongated along the c-axis,
and show needle-like terminations. Acicular aragonite
in stalagmites has been interpreted as being caused by
precipitation from alkaline solutions with high pH and
under low drip rates. Alternating layers of columnar
calcite and acicular aragonite in stalagmites within the
same cave thus clearly indicate that the hydrology of
the cave has varied periodically. In monsoonal settings,
for example, calcite layers form in the wet season, and
aragonite develops in the dry season.
ARCHIVES OF THE DEEP PAST
Changes in the morphology and mineralogy of
stalagmites clearly respond to climate-related phenom¬
ena, such as drip rate variability and temperature-
modulated cave ventilation. The physical and chemical
properties of stalagmites, therefore, provide unparal¬
leled information, often at very high temporal resolu¬
tion (annual to seasonal), on past climate and
environment changes in continental regions. These
records are especially powerful because stalagmites can
be dated precisely with radiometric techniques and by
counting annual growth layers.
Although most of the speleothem-based paleoclimate
research to date has been undertaken on samples from
the Late Quaternary, recent developments have now
extended their use for paleoclimate studies back to the
Paleozoic. For example, a speleothem with the mor¬
phology of a conical stalactite was dated 289 ± 0.68 Ma
BP by U-Pb radiometric determination. The stalactite
shows well-preserved concentric layers of columnar
and mosaic calcite. Due to its old age, diagenetic alter¬
ation such as dissolution and reprecipitation cannot be
excluded, despite the pristine appearance under the
microscope of the calcite crystals. The distribution of
trace elements in the speleothem, however, is coherent
with its concentric structure and, thus, consistent with
a primary geochemical signal in unaltered calcite. The
chemical properties of this ancient stalactite were used
to gain insight into climate on the supercontinent
Pangaea at an important period for early terrestrial ver¬
tebrate evolution (Woodhead et al., 2010).
The preservation of ancient stalagmites is seriously
endangered by erosion, tectonics, and, in the Boreal
hemisphere, the effects of aggressive meltwaters dur¬
ing periodic deglaciations in the Quaternary. There
are, therefore, few karst regions where ancient stalag¬
mites are likely to be extremely well preserved; one of
ENCYCLOPEDIA OF CAVES
810
s
FIGURE 5 Polished, thin slab of a mid-Pliocene stalagmite from
the Nullarbor showing well-preserved layers of columnar calcite
indicative of pristine preservation of its chemical and physical
properties.
these is the Nullarbor Plain of southwestern Australia.
This large, relatively flat karst area is, at present, char¬
acterized by arid conditions, but its caves preserve
dark, layered, candle-shaped stalagmites consisting of
columnar calcite (Fig. 5), most of which formed
between 3 and 4 million years ago, that is, during the
mid-Pliocene climatic warm period when global sur¬
face temperatures are believed to have been several
degrees higher than today, with relatively high atmo¬
spheric CO 2 concentration. The Nullarbor Caves also
preserve stalagmites that are as old as the Late
Miocene, when northward movement of Australia and
New Guinea are believed to have caused the gradual
closure of the Indonesian seaway, precipitating a major
reorganization of oceanic circulation. It is entirely pos¬
sible, then, that the age distribution, geochemistry, and
morphology of the Nullarbor stalagmites preserves a
wealth of information about the influence of plate
tectonics of regional climates, and the timing and
effects of warm periods in Earth's history, which may
provide analogues of the projected climate scenarios
for the next 100 years.
In conclusion, the combination of radiometric dating,
petrography, and geochemical analyses of speleothems
has the potential to become the most powerful tool
to reconstruct climate and environmental changes in
continental settings from at least the Paleozoic through
to the Quaternary.
Bibliography
Alonso-Zarza, A. M., & Tanner, L. H. (2010). Preface. In A. M.
Alonso-Zarza & L. H. Tanner (Eds.), Developments in sedimento-
logy, Vol. 61. Carbonates in continental settings — Facies, environments
and processes (pp. xi—xiii). Amsterdam: Elsevier.
Frisia, S., & Borsato, A. (2010). Karst. In A. M. Alonso-Zarza &
L. H. Tanner (Eds.), Developments in sedimentology, Vol. 61.
Carbonates in continental settings — Facies, environments and pro¬
cesses (pp. 269—318). Amsterdam: Elsevier.
Jones, B., & Renaut, R. B. (2010). Calcareous spring deposits in conti¬
nental settings. In A. M. Alonso-Zarza & L. H. Tanner (Eds.),
Developments in sedimentologij, Vol. 61. Carbonates in continental
settings — Facies, environments and processes (pp. 177—224).
Amsterdam: Elsevier.
Miorandi, R., Borsato, A., Frisia, S., Fairchild, I. ]., & Richter, D. V.
(2010). Epikarst hydrology and implications for stalagmite cap¬
ture of climate changes at Grotta di Ernesto (NE Italy): results
from long-term monitoring. Hydrological Processes, 24, 3101—3114.
Sunagawa, I. (2005). Crystals: Growth, morphology and perfection.
Cambridge, U.K.: Cambridge University Press.
Woodhead, J. D., Reisz, R., Fox, D., Drysdale, R., Hellstrom, ]., Maas,
R., et al. (2010). Speleothem climate records from deep time?
Exploring the potential with an example from the Permian.
Geology, 38, 455—458.
SULFURIC ACID CAVES
Arthur N. Palmer and Carol A. Hill 1
State University of New York, ^University of New Mexico
Most caves owe their origin to carbonic acid gener¬
ated in the soil. In contrast, sulfuric acid caves are
produced by the oxidation of sulfides beneath the
surface. Although sulfuric acid caves are relatively
few, they include some large and well-known exam¬
ples, such as Carlsbad Cavern, New Mexico. They
also provide evidence for a variety of deep-seated
processes that are important to petroleum geology,
ore geology, tectonic history, and the nascent field of
karst geomicrobiology.
GEOLOGIC SETTING
Cave origin by sulfuric acid requires the reaction of
oxygen with either solid or aqueous-phase sulfides.
The most common sulfide source is hydrogen sulfide
(H 2 S) which originates in reducing environments such
as petroleum-rich sedimentary basins. Sulfuric acid
is generated when rising H 2 S-rich water encounters
oxygen-rich water at or near the water table (Fig. 1).
The resulting caves are examples of hypogenic caves, the
origin of which depends on deep-seated processes and
rising groundwater. Oxidation of metallic sulfides (e.g.,
pyrite) accounts for only a few minor caves and gener¬
ally produces only scattered solutional pores. This
Encyclopedia of Caves.
© 2012 Elsevier Inc. All rights reserved.
SULFURIC ACID CAVES
811
FIGURE 1 Typical setting for sulfuric acid cave origin. Alternatively, in many areas the H 2 S is generated within an adjacent basin and driven
in aqueous solution laterally or upward into carbonate rocks by favorable hydraulic gradients or by tectonic or sedimentary overpressuring. The
H 2 S source may also be volcanic.
chapter focuses only on caves produced by H 2 S
oxidation.
The source of cave-forming H 2 S is usually a sedi¬
mentary basin bordering a carbonate upland, or deep
sulfate deposits interbedded within the carbonate rock
itself. Water from basinal sediments can be expelled by
compaction, overpressuring by rapid sediment accu¬
mulation, as well as by flow down a regional hydraulic
gradient. Because of their higher permeability, carbon¬
ate rocks often have lower hydraulic heads than sur¬
rounding materials and serve as the zones of
groundwater discharge to the surface. Cave origin by
these rising fluids requires oxidation of H 2 S beneath
the surface in carbonate rocks. This requires either con¬
vergence with oxygenated groundwater fed by mete¬
oric infiltration, or exchange of air through openings to
the surface. Either way, oxidation is greatest at and
near the water table. Most H 2 S-rich groundwater sim¬
ply discharges at springs without encountering enough
oxygen below the surface to form caves.
DIAGNOSTIC FEATURES OF
SULFURIC ACID CAVES
The typical sulfuric acid cave consists of a central
area of irregular rooms, with local mazes of intersect¬
ing fissures or sponge-like solution pockets and sparse
outlet passages that lead to active or relict springs.
Some caves consist of only a single passage. From
these main zones of cave development, narrow fissures
commonly descend and pinch downward. In active
examples these fissures are the main inlets for H 2 S-rich
water.
Much of the solutional enlargement of a sulfuric
acid cave takes place above the water table. H 2 S
escapes from the rising water into the cave air and is
absorbed by moisture droplets and films, along with
oxygen from the cave atmosphere. The moisture can
be supplied by infiltration and/or condensation.
Production of sulfuric acid within this water dissolves
the adjacent carbonate rock and typically produces a
rind of replacive gypsum. Fractionation during the
redox reactions makes the sulfur isotopes in the sec¬
ondary gypsum relatively "light" compared to those of
primary gypsum. Sulfuric acid can also alter clays,
yielding several by-products including alunite
(KA1 3 (S0 4 ) 2 (0H) 6 ) and hydrated halloysite (or endel-
lite, Al 2 Si 2 05 ( 0 H ) 4 • 2H 2 0), while liberating silica,
which precipitates as intergranular cements and irreg¬
ular chert beds. Native sulfur can accumulate by
incomplete oxidation of H 2 S at low pH, especially
where acids are locally shielded from the carbonate
rock by gypsum or siliciclastic rock.
Cave enlargement can continue even after sulfuric
acid activity has ceased. Through-flowing meteoric water
easily removes the secondary gypsum. Condensation of
moisture in cool upper levels and on cave ceilings is also
prominent, owing to circulation of warm humid air from
lower levels. Because it absorbs C0 2 from the cave air.
ENCYCLOPEDIA OF CAVES
812
s
condensation moisture readily dissolves carbonate rocks.
This moisture consists mainly of capillary films, which
produce smooth, rounded surfaces and cupolas. The dis¬
solved material is carried downward by seepage along
the walls and within pores, and it often reprecipitates as
cave popcorn and needle-like speleothems where the
water descends into evaporative regions. This process is
more noticeable than in most other caves, because sulfu¬
ric acid caves rarely contain through-flowing streams,
and their entrances tend to be small and few, thus
enhancing thermally driven convection within the cave
atmosphere.
Many of the diagnostic features of sulfuric acid cave
origin can be removed by near-surface processes after
the caves cease to be active. This is especially true in
humid climates, where the evidence of former sulfuric
acid cave development is often effaced by shallow
meteoric groundwater. Cave patterns and their relation
to the geologic setting may be the only remaining
genetic clues.
CHEMICAL REACTIONS
Most H 2 S is generated by reduction of sulfates by
organic carbon compounds. Gypsum and anhydrite
are the typical sulfate sources involved in sulfuric acid
speleogenesis. The simplest form of the reduction reac¬
tion is
SO 2- + 2CH z O -► H 2 S + 2HC0 3 (1)
where CH 2 0 represents a generic organic compound.
H 2 S is highly soluble in water, and except at depths of
only a few meters below the water table it rarely reaches
concentrations high enough to produce gaseous bubbles.
Where H 2 S is carried by groundwater into an oxidizing
environment, sulfuric acid is produced:
H 2 S + 20 2 -> 2H + + SO 2 (2)
usually with other sulfur species as intermediate
products. Microbial processes greatly speed the reac¬
tions and help to determine the intermediate reac¬
tions. At low pH, the following equilibrium becomes
important:
2H + + SO 4 <-+ H + + HS0 4 (3)
Within carbonate rocks, sulfuric acid is a potent
cave-former. For limestone, the dissolution reaction
can be stated:
CaC0 3 +2H+ + SO 2 -Ca 2+ + SO 2- + H + +HCO 3 (4)
H + + HC0 3 «-► H 2 C0 3 <-► C0 2 +H 2 0 (5)
Dolomite dissolves in a similar manner. As reaction
(5) proceeds, the partial pressure of carbon dioxide
(PC02) rises, and if C0 2 does not escape, carbonic acid
aids in cave development. Where groundwater flow
rates are small, gypsum can be deposited, often as a
direct replacement of carbonate rock:
Ca 2 + + SO 2- +2H 2 0 «-► CaS0 4 ■ 2H z O ( 6 )
In groundwater, buffering of the acids by dissolution
of carbonate rock generally keeps the pH moderate
(~ 6 —7). Lower pH can develop if the acids are shielded
from the carbonate rock by a nonreactive material.
Gypsum crusts tend to provide this shield so that above
the water table the pH of water films and droplets on
gypsum can decrease considerably. Build-up of acid in
contact with gypsum is limited by reaction (3), which
usually holds pH to values of about 1.5 (depending on
temperature and other dissolved species), while gypsum
dissolution consumes H + to produce HSO 4 ". However,
droplets and films of moisture on relatively nonreactive
materials such as chert, quartz, and microbial filaments
can reach extremely low (sometimes negative) pH
values, as in Cueva de Villa Luz (described below).
MICROBIAL INTERACTIONS
Reduction of sulfate in the presence of organic car¬
bon can take place spontaneously at temperatures
above about 85°C. At lower temperatures the process
depends on bacterial sulfate reduction. Thus in many
cases (perhaps most), H 2 S production requires micro¬
bial mediation. Sulfate reducers include the bacterium
Desulfovibrio, which ingests sulfate and excretes H 2 S.
In the oxidizing zone, where the caves are formed, sul¬
fur-oxidizing bacteria (e.g., Thiothrix ) can speed reactions
by several orders of magnitude. Also, certain species are
able to facilitate the conversion of sulfides directly to
sulfuric acid without intermediate by-products. The
oxidation reactions are strongly exothermic, provid¬
ing an energy source for microbial growth. Bundles
of bacterial filaments are common in active sulfuric
acid caves, and their mineral-coated fossilized forms
have also been observed in relict caves. Microbes are
also active mediators of redox reactions in the deeply
weathered walls of certain sulfuric acid caves.
FIELD EXAMPLES
Some of the best-documented examples of sulfuric
acid caves are described here.
ENCYCLOPEDIA OF CAVES
SULFURIC ACID CAVES
813
Kane Caves, Wyoming
Lower Kane Cave, in north-central Wyoming, is
where sulfuric acid cave development was first
described in detail (Egemeier, 1981). The cave consists
of a single linear stream passage in the Madison
Limestone of Mississippian age, where the limestone is
breached by the Bighorn River along the Sheep
Mountain Anticline. The cave spring is located at the
present river level, and the passage extends at a low
gradient for approximately 350 m. The cave stream is
fed by H 2 S-rich water that discharges from floor fis¬
sures at several places along the passage, including the
upstream terminus (Fig. 2). Bundles of filaments of sul¬
fur-oxidizing bacteria form conspicuous tendrils in
FIGURE 2 Fissure inlet midway in Lower Kane Cave, Wyoming.
Note the mat of white bacterial filaments in foreground. Photo by A.
N. Palmer.
and around the fissures. A gypsum rind coats most of
the walls and ceilings around the inlets as the result of
oxidation of H 2 S within moisture films. The gypsum
flakes off the limestone bedrock and drops to the floor,
where it either falls into the stream or builds mounds
adjacent to it. The stream is undersaturated with gyp¬
sum, and any gypsum that comes in contact with it is
carried away in solution. This process contributes
greatly to the enlargement of the cave.
About 30 m directly above is Upper Kane Cave, a
relict precursor of the active lower cave. The upper
cave is a dry passage similar in pattern and length to
the lower one, but with a larger cross-section and a
breakdown-strewn floor. Breakdown and dissolution
by vadose seepage have obscured or removed much of
the gypsum crust that must have once lined most of its
surfaces. Discharging along both banks of the Bighorn
River are other similar caves that contain H 2 S-rich
water, but they are smaller than the Kane Caves and
present more difficult access.
The water that feeds the active caves in this area
comes from two large regional groundwater systems
with infiltration as much as 150 km away. The appar¬
ent source of the H 2 S is reduction of sulfates inter-
bedded within the Madison in the presence of basinal
hydrocarbons. Sulfates have been removed by shallow
meteoric groundwater in most exposed areas of the
Madison, but it is still abundant within inter-mountain
basins.
Cueva de Villa Luz, Mexico
Of all well-studied sulfuric acid caves, this one has
the most intense H 2 S-H 2 S 04 activity. It is located in the
semitropical jungle of southern Tabasco, in a low
upland of folded and faulted Cretaceous limestone
(Fig. 3; Hose et al., 2000). It lies several tens of kilo¬
meters south of a basinal oilfield and 50 km east of the
active El Chichon volcano. The cave atmosphere is
highly toxic, with fluctuating H 2 S concentrations that
sometimes exceed 200 ppm. (The OSHA safety standard
FIGURE 3 Profile of Cueva de Villa Luz, Mexico. From map by Bob Richards and Louise D. Hose.
ENCYCLOPEDIA OF CAVES
814 s
for H 2 S was once 10 ppm but has recently been
decreased to zero.) Levels of atmospheric carbon monox¬
ide and carbon dioxide are also dangerously high at
times. Gas masks are mandatory for safe entry, although
visitors often unwisely explore the near-entrance regions
without protection.
The cave consists of braided stream passages that
discharge to a single spring, which is the head of the
surface stream Rio Azufre ("Sulfur River"). Total dis¬
charge is a rather uniform 200—300 L s' 1 and shows lit¬
tle response to rainfall variation. The cave water is a
sulfate-chloride-bicarbonate brine compatible with a
basinal origin, presumably from the oilfield to the north.
Light sulfur isotopes in the H 2 S support this hypothesis.
Water flows into the cave through approximately 26
inlets, all too small for human entry. Most are located at
the upstream end of the cave, with a few about midway
in the cave. The cave streams have moderate and rather
uniform gradients, with a few scattered deep pools up
to several meters deep. The ceiling rises and falls
abruptly, with many skylights to the surface.
Each stream inlet delivers one of two types of water
(Table 1). Some are anoxic and high in H 2 S. Others are
oxygenated, with no detectable H 2 S, and have appar¬
ently encountered aerated conditions farther upstream.
Otherwise the two water types are chemically rather
similar. The anoxic water releases H 2 S to the cave
atmosphere, and white filaments of sulfur-fixing bacte¬
ria coat the floors over which it flows. The water is
roughly at equilibrium with dissolved limestone and
dolomite. The oxygen-rich water precipitates iron
hydroxide, which coats filaments of iron-fixing bacte¬
ria that grow on the floors of these streams. This water
is slightly supersaturated with respect to limestone
and dolomite, and in a few places it deposits calcite
travertine. The two waters mix to produce streams that
are white with colloidal sulfur and which have a
nearly neutral pH.
Infiltration moisture on the cave walls and ceilings
absorbs H 2 S and oxygen from the cave atmosphere,
and the resulting sulfuric acid has produced a thick
coating of crystalline gypsum over most limestone sur¬
faces, especially in the vicinity of sulfide-rich inlets.
Sulfur accumulates on the gypsum-coated walls above
these inlets. The gypsum and sulfur are highly
depleted in 34 S, with fc 34 S ranging from —26 to —227™
CDT, showing the influence of microbial mediation.
The cave walls are coated in many places by organic
slime intermixed within the gypsum. Bacterial fila¬
ments hang from many of the gypsum surfaces, and
drops that accumulate on them have an average pH of
1.4 (Fig. 4). However, the drops that linger longest
before falling have measured pH values as low as
zero. They can burn skin and eat holes in clothing.
The filaments form white, gelatinous, elastic bun¬
dles. Their mucus-like texture has inspired a vivid
informal name, but the term biothem is gaining
favor. The bacteria form the base of a complex food
chain, including midges, spiders, gastropods, and fish
(Poecilia mexicana). In an annual spring ceremony, the
local Soque Indians harvest the fish just inside the cave
entrance to provide food for a community feast.
Although it is compelling to think of the entire cave
community as based solely on the energy from sulfide
oxidation, the system is complicated by an influx of
organic material and light from the many openings to
the surface. A more likely candidate for pure chemo-
autotrophy is the less diverse underground commu¬
nity of Movile Cave in southeastern Romania, which
occupies an air-filled H 2 S-rich room that is almost
TABLE 1 Water Chemistry in Lower Kane Cave, July 1970 (Egemeier, 1981) and Cueva de Villa Luz, January 1998 (Hose et al, 2000)
Lower Kane Inlets
Villa Luz H 2 S-Rich Inlet
Villa Luz 0 2 -Rich Inlet
Villa Luz Spring
Temp (°C)
23.5
27.5
28.3
28.0
pH
7.0
6.61
7.23
7.14
Ca (mg L -1 )
97
396
383
393
Mg
31
81
97
88
Na
11
484
477
-
hco 3
218
498
451
477
so 4
189
940
980
910
Cl
12.5
814
792
803
H 2 S (aq)
6 ±1.5
500 ± 50
N.D. (<0.1)
17.5 ±3
0 2 (aq)
0
N.D. (<0.1)
4.3 ±0.7
1.3 ±0.2
Pcc >2 (atm)
0.0005-0.0014
0.11
0.023
0.030
ENCYCLOPEDIA OF CAVES
SULFURIC ACID CAVES
815
FIGURE 4 Bundles of bacterial filaments in Cueva de Villa Luz,
Mexico. The droplets are dilute sulfuric acid with pH as low as zero.
Height of photo is ~15 cm. Photo by A.N. Palmer.
FIGURE 5 Massive gypsum in the main upper level of the
Frasassi Cave System, Italy. Note vertical grooves formed by drip¬
ping meteoric water. The dark stain at the bottom is from late-stage
clay fill. Photo by A.N. Palmer.
totally isolated from external nutrients (Sarbu et al.,
1996). However, Mo vile Cave has an H 2 S activity far
lower than that of Cueva de Villa Luz.
Much of the sediment on the floor of the main
stream of Cueva de Villa Luz is an organic-rich muck
in which sulfate reduction and iron sulfide precipita¬
tion are actively taking place. A large percentage of the
sediment consists of tiny gastropod shells a few milli¬
meters in diameter, as well as fragments of allogenic
metamorphic rock.
The cave enlarges in several ways. The most vigor¬
ous enlargement is by the same gypsum-replacement
mechanism described above for the Kane Caves. It also
accounts for the irregular ceiling profile. The skylights
apparently originated as narrow fissures communicat¬
ing with the surface, which have enlarged as escaping
H 2 S is absorbed by moisture on their walls. Although
the main streams are slightly supersaturated with both
calcite and dolomite during dry periods, they are able
to dissolve the bedrock during periods of high infiltra¬
tion, when the dripping of H 2 SC> 4 -rich water is most
abundant. Beneath skylights, drips from the surface
form cylindrical holes in the gypsum piles. Where
acidic water drains through the porous gypsum of the
cave walls, it forms solutional rills in the limestone
near stream level, where the gypsum is absent.
Frasassi Cave System, Italy
Perhaps the largest known active sulfuric acid cave,
the Frasassi Cave System is one of Europe's best-
known show caves. It is located in folded and faulted
Jurassic limestone along the eastern flank of the
Apennine mountain chain. Its lowest level is a stream
passage with high concentrations of H 2 S and active
sulfuric acid cave enlargement. The source of the sul¬
fides is thought to be reduction of underlying gypsum
beds. Passages and rooms at higher levels are much
larger, and although presently inactive they show evi¬
dence for a sulfuric acid origin, such as massive bodies
of gypsum derived from replacement of the native
limestone (Fig. 5). Passages have highly irregular pat¬
terns, as is typical for caves in highly deformed host
rock, but the overall passage profiles are fairly hori¬
zontal and reflect present and past fluvial base levels.
The processes and passage character at the stream
level are similar to those described above in Cueva de
Villa Luz, although present H 2 S concentrations in
Frasassi are only about 10% as great, and there are
fewer known inlet points for sulfide-rich water.
Infiltrating water dilutes the sulfide-rich water by
30—60%. Measurements by Galdenzi et al. (2008)
show that limestone dissolution by H 2 S oxidation is
ENCYCLOPEDIA OF CAVES
816
s
most intense in the wet season, when the infiltration
of oxygen-rich water is greatest.
Caves of the Guadalupe Mountains,
New Mexico
The Guadalupe Mountains of southeastern New
Mexico contain some of the world's most spectacular
caves, of which Carlsbad Cavern and Lechuguilla
Cave are the largest and best known (Figs. 6 and 7).
Although presently inactive relics, these caves have
been the focus of extensive studies of sulfuric acid
cave origin. Many of the processes recently observed
in Cueva de Villa Luz were anticipated by studies of
now-inactive caves in the Guadalupes.
The Guadalupes are composed of a Permian reef com¬
plex, which today stands at altitudes up to 2600 m as the
result of Cretaceous and Late Cenozoic uplift. The
Capitan Reef, which forms the southeastern escarpment,
grades northwestward into bedded back-reef limestones
and dolomites. An apron of fore-reef talus extends to
the southeast and forms the boundary with the
Delaware Basin, which consists of Permian carbonates
and quartz-rich sandstones capped by sulfates and
halides. Hydrocarbons are abundant at depth within
the Delaware Basin and in smaller concentrations
beneath the Guadalupe Mountains.
The typical Guadalupe cave has a ramifying pattern
consisting of irregular rooms and mazes with passages
branching outward from them (Fig. 6). Overlapping
FIGURE 6 Map and profile of Carlsbad Cavern, New Mexico.
E = entrance. Other letters show correspondence between locations
on map and cross-section. From map by Cave Research Foundation.
FIGURE 7 The Chandelier Maze, Lechuguilla Cave, shows the
sponge-like pattern of dissolution in the massive Capitan reef rock,
vivid wall colors produced by mineral oxidation, and accumulations
of speleogenetic gypsum. Photo by A.N. Palmer.
tiers of rooms and passages are common. Branches do
not converge as tributaries, but instead serve as dis¬
tributary outlets at successively lower elevations. Some
caves involve simple widening of only a few fractures.
Many caves, or parts of caves, have complex network
or sponge-like patterns. Because of the large primary
porosity in the reef, spongework is most abundant in
caves in that rock unit.
As in the Frasassi Caves described above, caves of
the Guadalupe Mountains have considerable vertical
relief. Some of this may be the result of mixing
between H 2 S-rich water and oxygenated water deep
beneath the water table, but successive stages of water-
table lowering appear to be the main cause. Most cave
enlargement takes place above the active passages, as
in Cueva de Villa Luz, so that there is great integration
between levels. H 2 S-rich water rising from depth along
prominent fractures and partings (either from the
Delaware Basin or the Guadalupe block itself), con¬
verged with fresh groundwater that infiltrated at high¬
er elevations to the west and southwest. Although
sulfuric acid production may have begun as much as
100—200 m below the contemporary water table when
the caves were actively forming, the greatest burst of
oxidation took place at and just above the water table,
producing some of the largest cave rooms in North
America. The Big Room of Carlsbad Cavern is up to
80 m high and covers an area of 33,200 m 2 .
Large cave rooms have irregular outlines rimmed
by dead-end galleries and blind alcoves. Some rooms
and adjacent passages have nearly horizontal floors
that disregard stratigraphic boundaries, indicating dis¬
solution at former water tables. The oxygen demand in
producing a large room requires plentiful air exchange
with the surface. Most rooms connect to the surface
ENCYCLOPEDIA OF CAVES
SULFURIC ACID CAVES
817
via older ascending phreatic passages. In the absence
of an open cave passage, air can also be exchanged in
limited quantities through narrow fissures in the over-
lying bedrock. The ceilings of major cave rooms are
typically smooth and arched, and they show evidence
of gypsum replacement. Many are still lined by gyp¬
sum rinds (Fig. 7). Some rooms are floored by thick
gypsum up to 10 m thick, which accumulated in pools
by a combination of subaqueous precipitation and
accumulation of fallen material from the ceiling. The
S 34 S of the cave gypsum is highly negative, with a
mean of —16.8 %> CDT, compared to +10.3%o for the pri¬
mary marine gypsum and anhydrite of the region,
which shows that the H 2 S that formed it is derived
from reduction of primary sulfates (Hill, 1987).
Many passages have steep gradients that connect
different levels or serve as entrance galleries. Although
entrances of this type give the impression that they
were originally surface-water inlets, closer inspection
shows that they may have been initiated as groundwa¬
ter outlets while the caves were forming. The entrance
passages contain no stream entrenchment, and almost
no vertical shafts or coarse sediment. Instead, the pas¬
sage ceilings rise in a series of smooth convex-upward
arcs quite distinct from the abrupt stair-step pattern
typical of vadose passages in humid climates. Local
confinement of rising water by resistant beds forced
some passages to follow an up-dip course. However,
much of the passage enlargement probably took place
by absorption of H 2 S and oxygen in the cave atmo¬
sphere. In places, dripping sulfuric acid produced dis¬
solution rills and potholes in the bedrock. Freshwater
seepage that has entered the cave since the sulfuric
acid phase has formed drip holes that extend through
gypsum blocks.
The caves also contain a wide variety of carbonate
speleothems. Besides the common dripstone, flow-
stone, and pool deposits are several rare forms.
Underwater helictites, produced by the common-ion
effect between dissolved calcite and gypsum, were first
discovered in Lechuguilla Cave. They are among the
very few known in the world. Finger-like pool fingers
and u-loops are calcite-coated bacterial filaments. The
filaments are now fossilized, but the resulting forms
are remarkably like the bundles of bacterial filaments
in Cueva de Villa Luz (Fig. 4).
Other cave deposits include clay alteration products
such as alunite and hydrated halloysite (endellite).
Argon dating can be applied to the potassium-bearing
alunite, which is generated by sulfuric acid attack of
montmorillonite and therefore provides the dates of
active cave enlargement (Polyak et ah, 1998). The highest-
elevation caves in the Guadalupes are about 12 million
years old, and the lowest caves, which apparently repre¬
sent the latest phase of sulfuric acid development, are
about 4 million years old. This was the first application
of this procedure to cave dating.
Because of the close association of major cave devel¬
opment with former water tables, the vertical arrange¬
ment of Guadalupe caves must relate to the erosional
history of nearby rivers. One might expect major pas¬
sages and rooms to correlate with pauses in base-level
lowering, and therefore to cluster at similar elevations.
However, correlations between levels of development
are somewhat obscure from one cave to another, and
even within different sections of the same cave. As the
water table dropped, bursts of cave enlargement
occurred when and where H 2 S happened to rise to the
water table in significant quantity. Major rooms and
passages were produced by episodic influx of H 2 S,
probably during periods of uplift and faulting. When
these releases coincided with periods of rather static
water table, distinct horizontal levels resulted. Alunite
dating suggests that the main episodes lasted at least
10 5 years.
Recent studies suggest that the Guadalupe caves
were probably fed by water from a large recharge area
to the west, which has since dropped in elevation
because of extensional tectonics and no longer serves
as a catchment area for the present Guadalupe
Mountains (DuChene and Cunningham, 2006). If so,
the caves provide a record not simply of local water-
table lowering, but of the Late Tertiary tectonic history
of the entire region.
Caves of the Grand Canyon, Arizona
Grand Canyon National Park, while not primarily a
"cave park ", does contain a widespread karst system.
Most of the caves in this region are located in the
uppermost unit (Mooney Falls Member) of the Late
Mississippian Red wall Limestone. Minor cave devel¬
opment also occurs in the Permian Kaibab Limestone
and Cambrian Muav Limestone. There are two types
of caves in the canyon: (1) older paleocaves in the
Redwall that drained across the Colorado Plateau
before the downcutting of the Grand Canyon, and (2)
younger vadose caves that drained (or are still drain¬
ing) to the Colorado River during or after the down¬
cutting of the Grand Canyon. The older paleocaves
contain a number of deposits that can be dated with
respect to the paleo-watertable. This information can
then be used to help determine the age of downcutting
of the canyon.
Various analyses relate the cave deposits to the evo¬
lution of the Grand Canyon: 40 Ar/ 39 Ar dating of the
minerals alunite and jarosite; U-series and U-Pb dating
of cave spar, cave mammillaries, and other speleo¬
thems; and stable isotope analysis. X-ray diffraction.
ENCYCLOPEDIA OF CAVES
818
s
and chemical analysis of various cave deposits. From
these analyses, the following sequence of events has
been established for the Redwall paleocaves: (1)
Mississippian Redwall Limestone/Stage 1 karst epi¬
sode, which is represented by paleokarst breccias; (2)
Stage 2 (Laramide?) phreatic karst and iron oxide epi¬
sode represented by hematite-goethite cave deposits
formed at redox boundaries; (3) a calcite spar episode
characterized by crystal linings overlying the hematite-
goethite; (4) a shallow phreatic, or water-table, episode
of cave origin, including some sulfuric acid processes,
characterized by solution domes, alunite, mammil-
laries, and replacement gypsum; and (5) subaerial spe-
leothem deposition.
The sulfuric acid episode (3) in Grand Canyon caves
probably represents renewed dissolution at or near
the water table in the Pliocene-Pleistocene, with the
source of H 2 S possibly derived from hydrocarbons
in the underlying Precambrian Chuar Group. In the
Grandview mine, which is at the same level as the Cave
of the Domes on Horseshoe Mesa, an 40 Ar/ 39 Ar date of
~ 700,000 YBP for alunite suggests that water-table cave
development in that part of the canyon occurred around
that time (Fig. 8A). Mammillaries collected from Tse'an
Bida Cave, Mother Cave, and Grand Canyon Caverns
gave U-series dates of >600,000 YBP. One stalactite from
Bat Cave, western Grand Canyon, gave a U-series date
of ~ 460,000 YBP, indicating that the level of the water
table (river level) must have been somewhere below the
level of Bat Cave ~ 0.5 Ma BP and that the rate of
downcutting for that part of the western Grand Canyon
must have been <0.67 mm yr _1 (Fig. 8B). More recent
data suggest that canyon entrenchment may have
extended as far back as 17 Ma BP (Polyak et al., 2008).
RELATION OF SULFURIC ACID CAVES
TO MISSISSIPPI VALLEY-TYPE ORE
DEPOSITS, URANIUM DEPOSITS,
HYDROCARBONS, AND
RESERVOIR POROSITY
Sulfuric acid caves are important not only because
they represent a mode of speleogenesis recognized only
recently, but also because they can be genetically related
to native sulfur deposits, Mississippi Valley— type
(MVT) ore deposits, uranium deposits, and hydrocar¬
bons and reservoir porosity. All of these deposits and
features are interconnected via the generation and
migration of hydrogen sulfide within and around intra-
cratonic basins. The general model is that H 2 S is gener¬
ated at depth by the reaction of hydrocarbons with
evaporite (sulfate) rock according to reaction (1). The
H 2 S can then either stay within the source beds or it can
move into neighboring or overlying carbonates.
Within a basin, aqueous H 2 S can react with oxygen
to form native sulfur deposits. Examples include the
massive economic sulfur deposits at the Culberson sul¬
fur mine, Delaware Basin, southeastern New Mexico,
and west Texas (south of Carlsbad Cavern). Also,
A
A’
FIGURE 8 Relationship of dated cave and
mine deposits in the Grand Canyon to the his¬
tory of downcutting of the canyon. From Hill
et al., 2004.
B'
Distance (km)
ENCYCLOPEDIA OF CAVES
SULFURIC ACID CAVES
819
porosity can be created where dissolution is produced
by the mixing of waters of different H 2 S content or by
the oxidation of H 2 S. Sulfuric acid oil-field karst is a spe¬
cific kind of H 2 S-related porosity where carbonate
reservoirs of cavernous size have been dissolved by a
sulfuric acid mechanism. In an H 2 S system, porosity
can be produced entirely in the deep subsurface and
does not have to represent a paleokarst surface formed
by shallow-phreatic or vadose dissolution.
Mississippi Valley—type (MVT) ore deposits com¬
monly occur in carbonate rock around the edges or
margins of intracratonic basins. This is due to the
migration of H 2 S from hydrocarbon basins into struc¬
tural and stratigraphic traps along the basin margins.
As groundwater moves through these traps, metal
ions such as lead, zinc, and copper that precipitate
readily in the H 2 S-rich environment. Thus the H 2 S
migration mechanism applies not only to the small
MVT deposits around the Delaware Basin, but also
to huge lead-zinc deposits such as those of the
Viburnum Trend, Missouri.
Uranium is another metal that precipitates readily
in the presence of H 2 S. Such precipitation is typical of
roll-front deposits where groundwater encounters and
percolates through a reducing environment, allowing
uranium to precipitate preferentially at the redox inter¬
face. In the case of the Grand Canyon, it is possible not
only that the isotopically light cave gypsum was ulti¬
mately derived from hydrocarbons in the subsurface,
but also that H 2 S from hydrocarbons was the source of
reduced sulfur for the copper and uranium mineraliza¬
tion in the mines of the Grand Canyon. Some of the
highest-grade uranium deposits in North America are
located in breccia pipes of the Grand Canyon, and
these breccia pipes and paleokarst breccias may have
acted as stratigraphic and structural traps.
Sulfuric acid caves are an integral part of this intra¬
cratonic basin model of H 2 S generation and migration.
Where H 2 S migrates from a basin into carbonate rock
surrounding the basin and encounters oxygenated
groundwater, then sulfuric acid forms via reactions
(2) and (3), and cave dissolution occurs via reactions
(4) and (5). As a by-product of these reactions, gypsum
and native sulfur form within these caves, hydrated
halloysite (endellite) and alunite are produced under low
pH sulfuric acid solutions, and the uranium-vanadium
minerals tyuyamunite and metatyuyamunite precipitate
in an H 2 S-rich environment.
CONCLUSIONS
The role of sulfuric acid in cave origin, which was
virtually unrecognized only a few decades ago, is now
recognized as one of the most significant cave-forming
processes. It also provides a view of deep-seated pro¬
cesses that integrates cave studies with other fields,
such as economic geology and microbiology.
See Also the Following Articles
Speleogenesis, Hypogenetic
Lechuguilla Cave, New Mexico
Bibliography
DuChene, H. R., & Hill, C. A. (Eds.), (2000). The caves of the
Guadalupe Mountains. Journal of Cave and Karst Studies (62).
pp. 51-158.
DuChene, H. R., & Cunningham, K. I. (2006). Tectonic influences on
speleogenesis in the Guadalupe Mountains, New Mexico and Texas
(pp. 211—218) (Guidebook to 57th Annual Field Conference).
New Mexico Geological Society.
Egemeier, S. J. (1981). Cavern development by thermal waters.
National Speleological Society Bulletin, 43, 31—51.
Galdenzi, S., Cocchioni, M., Morichetti, L., Amici, V., & Scuri, S.
(2008). Sulfidic ground-water chemistry in the Frasassi Caves,
Italy. Journal of Cave and Karst Studies, 70(2), 94—107.
Hill, C. A. (1987). Geology of Carlsbad Cavern and other caves in the
Guadalupe Mountains, New Mexico and Texas (Bulletin 117). New.
Hill, C. A. (1995). Sulfur redox reactions: Hydrocarbons, native sul¬
fur, Mississippi Valley-type deposits and sulfuric acid karst in the
Delaware Basin, New Mexico and Texas. Environmental Geology,
25, 16-23.
Hill, C. A., Polyak, V. J., McIntosh, W. C., & Provencio, P. P. (2001).
Preliminary evidence from Grand Canyon caves and mines for
the evolution of the Grand Canyon and Colorado River System.
In R. A. Young & E. E. Spamer (Eds.), The Colorado River: Origin
and evolution (pp. 141—145). Grand Canyon, AZ: Grand Canyon
Association (Monograph 12).
Hose, L. D., Palmer, A. N, Palmer, M. V., Northup, D. E., Boston,
P. J., & DuChene, H. R. (2000). Microbiology and geochemistry in
a hydrogen-sulphide-rich karst environment. Chemical Geolog]/,
169, 399-423.
Klimchouk, A. B. (2007). Hypogene speleogenesis: Hydrogeological and
morphogenetic perspective. Carlsbad, NM: National Cave and Karst
Research Institute.
Northup, D. E., Dahm, C. N., Melim, L. A., Spilde, M. N., Crossey,
L. J., Lavoie, K. H., et al. (2000). Evidence for geomicrobiological
interactions in Guadalupe caves. Journal of Cave and Karst Studies,
62(2), 30-40.
Palmer, A. N., & Palmer, M. V. (2000). Hydrochemical interpretation
of cave patterns in the Guadalupe Mountains, New Mexico.
Journal of Cave and Karst Studies, 62(2), 91—108.
Polyak, V. J., McIntosh, W. C., Giiven, N., & Provencio, P. (1998).
Age and origin of Carlsbad Cavern and related caves from
40 Ar/ 39 Ar of alunite. Science, 279(5358), 1919-1922.
Polyak, V. J., Hill, C. A., & Asmerom, Y. (2008). Age and evolution of
the Grand Canyon revealed by U-Pb dating of water table-type
speleothems. Science, 329(5896), 1377—1380.
Queen, J. M., & Melim, L. A. (2006). Biothems: Biologically influenced
speleothems in caves of the Guadalupe Mountains, New Mexico, USA
(Guidebook to 57th Annual Field Conference) (pp. 167—173).
New Mexico Geological Society.
Sarbu, S. M., Kane, T. C., & Kinkle, B. K. (1996). A chemoautotrophi-
cally based groundwater ecosystem. Science, 272(5270), 1953—1955.
ENCYCLOPEDIA OF CAVES
TIANKENG
Xuewen Zhu, Weihai Chen, and Yuanhai Zhang
Institute of Karst Geology, Chinese Academy
of Geological Sciences, Guilin, China
DEFINITION
The word tiankeng is a transliteration from two
Chinese characters, "^$ 1 " (tian keng in pinyin), that
roughly mean sky hole or heaven pit or some similar
variation on that double theme. Tiankeng are very
large dolines formed in carbonate rocks that are more
than 100 m deep and wide, and have volumes of more
than 1 million cubic meters, steep profiles with
vertical cliffs around all or most of the perimeters, and
are connected by underground cave rivers.
TYPES OF TIANKENGS
Tiankengs may be classified as collapsed tiankeng,
erosional tiankeng, and water tiankeng by their origin
and development (Fig. 1) or classified as immature
tiankeng, mature tiankeng, and degraded tiankeng by their
evolution (Fig. 2).
Collapsed tiankengs are dominant in number, distri¬
bution, and dimension as well as difference of varied
ages. The development of a collapsed tiankeng is dis¬
tributed relative to the underground river trace instead
of surface topography of the karst, and is the result of
intensive erosion and dissolution of underground
water. However, the formation of a tiankeng is the
reconstruction of large collapses. The appearance of
tiankengs with different ages may be the symbol
of mature karst development, such as Dashiwei
Tiankengs Group in the Bailang Underground River
System in Guangxi, China (Fig. 3).
Encyclopedia of Caves.
821
© 2012 Elsevier Inc. All rights reserved.
822
T
FIGURE 2 Immature tiankeng, mature tiankeng, and degraded tiankeng.
FIGURE 3 Dashiwei tiankeng.
An erosional tiankeng develops at the site of a sink¬
ing allogenic stream, as distinct from collapse tiankengs
that are independent of surface drainage. The erosional
type therefore has a cave stream draining from it, but
not into it. Qingkou Tiankeng in Chongqing is therefore
an excellent example (Fig. 4), and there may be erosional
tiankengs in the Muller Plateau and in Nakanai, New
Britain, Papua New Guinea.
A water tiankeng has a pool inside, and is generally
of the collapsed type. Some are the skylights of active
underground rivers, such as the Sanmenhai Karst
Windows in Leye-Fengshan Global Geopark. Some are
connected with underground water in a phreatic zone,
such as Grveno Jezero in Croatia and Lago Azul in
Brazil. Some are collapsed tiankengs submerged by
water. The tiankeng of Grveno Jezero is 528 m deep
with a lake of 281 m deep and 16 Mm 3 water in vol¬
ume. Its altitude is —6 m, and there is an underground
stream flowing at the bottom.
FEATURES OF TIANKENGS
Tiankengs are unusually large as karst landforms
and they have clear collapse triangular profiles. They
FIGURE 4 Qingkou tiankeng.
usually destroy any other landforms of the surface
karst, including dry valleys, depressions, dolines, or
cone karst, either individually or collectively, which
indicate that they are young karst landforms. That is the
significant difference between tiankengs and dolines.
There are four main stages of development in a tian¬
keng: (1) a cave with large river passage; (2) a large
cave chamber; (3) a tiankeng that is opened to the sur¬
face; (4) and finally a degraded tiankeng (Fig. 5). A col¬
lapsed tiankeng has the following critical factors:
ENCYCLOPEDIA OF CAVES
TIANKENG
823
FIGURE 5 The four main stages of development in a tiankeng.
1. A great and continuous thickness of carbonate rocks
and a deep vadose zone in the karst
2 . Tropical-subtropical humid climate and abundant
precipitation
3. Slow tectonic uplifting and deep trunk river incision
4. Developed underground water system, particularly
a subterranean river flowing through the tiankeng
bottom
5. Gently dipping beds and networks of joints or faults
in carbonate rocks
6. In contrast to the development of collapse tiankengs
where the surface hole opened abruptly from
beneath, erosional tiankengs are formed gradually
from the surface to the floor; excavation downward
by stream erosion, and fluvial removal of debris
and solutes, complement each other (Fig. 6 )
7. An erosional tiankeng develops at the boundary
between soluble and nonsoluble rocks; it has
formed within underlying limestone that has
propagated to the ground surface through a cover
of insoluble rock
N
FIGURE 6 Erosional tiankeng and underground river.
8 . An erosional tiankeng develops in abundant
thickness of a vadose zone or continuous descent of
an underground water table with tectonic uplifting
and river incision
ENCYCLOPEDIA OF CAVES
824
TABLE 1 Dimensions of the Main Tiankengs
in the World
T
and Some Comparable Features
Tiankeng
Length X Width (m)
Area (m 2 )
Depth (max. m)
Volume (Mm 3 )
Type
Location
Xiaozhai
625 x 535
274,000
662
119.3
C, M
China
Dashiwei
600 x 420
167,000
613
75
C, M
China
Xiaoyanwan
625 x 475
200,000
248
36
C, D
China
Datuo
530 x 380
149,000
290
32.7
C, D
China
Longgang
350 x170
53,000
350
9.2
C, M
China
Huangjing
320 x170
51,700
161
6.3
C, E, M
China
Qingkou
250 x 220
40,700
295
9.2
E, M
China
Yanzi
100 x 60
14,340
250
1.69
C,I
China
Minye
350 x 350
75,000
510
26
C, E, M
Papua New Guinea
Luse
800 x 600
350,000
250
61
C, D
Papua New Guinea
Kavakuna
380 x 300
80,000
480
24
C, D
Papua New Guinea
Nare
150 x120
13,000
310
4.7
M
Papua New Guinea
Garden of Eden
1200 x 800
750,000
300
150
Malaysia
Sendirian
115X90
12,000
240
2
M
Malaysia
Crveno Jezero
200 x 280
56,000
528
30
W
Croatia
Modro Jezero
700 x 400
19,000
290
22
D
Croatia
Verika Dolina
300 x 170
30,000
165
3.5
M
Slovenia
Lisicina
400 x 200
60,000
115
2
D
Slovenia
El Sotano
440 X 210
70,000
445
16
M
Mexico
Golondrinas
300 x130
26,000
400
5
I
Mexico
Peruacu North
450 x 200
85,000
170
10
Brazil
Peruacu South
400 x180
55,000
150
5
D
Brazil
Lago Azul
200 X 140
18,000
280
4.3
W
Brazil
Mangilg
700 x 500
280,000
140
25
Madagascar
Styx2
400 x 300
88,000
140
8
Madagascar
Tres Pueblos Sink
190 x180
25,000
120
2.5
Puerto Rico
C = collapsed; E = erosion; W = water; I = immature; M — mature; D = degraded.
9. An erosional tiankeng develops in regions with a
humid climate and abundant precipitation
Different water tiankengs at different sites may
have different ages, but most of them are young,
formed after the Quaternary.
DISTRIBUTION OF TIANKENGS
Tiankengs are the result of intensive karst evolution
and are mainly distributed in the highly active hydrody¬
namic karst region, which have the essential conditions
including continuous great thick and gently dipping
carbonate rocks, a humid climate, and abundant
precipitation. There are two main tiankeng regions:
South China Karst-North Vietnam Karst and Muller-
Nakanai Karst in Papua New Guinea. Both these regions
are distinguished by unusually large numbers of tian¬
kengs, especially in the fengcong karst area of south
China (Table 1). In addition, tiankengs have been dis¬
covered in Malaysia, Mexico, Brazil, Croatia, Slovenia,
Madagascar, and Puerto Rico.
Bibliography
Garasic, M. (2001). New speleo-hydrogeological research of Crveno
Jezero (Red Lake) near Imotski in Dinaric karst area (Croatia
Europe). Proceedings of the 13th International Congress of Speleology,
Brazil, 2, 168-171."
ENCYCLOPEDIA OF CAVES
TIANKENG
825
James, J. (2005). Tiankengs of the Muller Plateau, Papua New
Guinea. Cave and karst Science, 32(2—3), 85—91.
Waltham, A. C. (2005). Tiankengs of the world, outside China. Cave
and Karst Science, 32(2—3), 67—74.
Waltham, A. C., Bell, F. G., & Culshaw, M. G. (2005). Sinkholes and
subsidence. Chichester, U.K. Praxis Publishing.
Zhu, X. W. (2001). Karst tiankeng and its value for science and
tourism in China. Science & Technology Review, 10(160), 60—63
(in Chinese).
Zhu, X. W., & Chen, W. H. (2005). Tiankengs in the karst of China.
Cave and Karst Science, 32(2—3), 55—66.
Zhu, X. W., & Waltham, A. C. (2005). Tiankeng: Definition and
description. Cave and Karst Science, 32(2—3), 75— 77.
Zhu, X. W., Zhu, D. H., Baojian, H., & Chen, W. H. (2003a). Dashiwei
Tiankeng Group in Leye, Guangxi: Discoveries, exploration, definition
and research. Guangxi Scientific & Technical Publishers.
Zhu, X. W., Zhu, D. H., & Chen, W. H. (2003b). A brief study on
karst tiankeng. Carsologica Sinica, 2 2(1), 51—65 (in Chinese).
ENCYCLOPEDIA OF CAVES
UKRAINE GIANT GYPSUM CAVES
Alexander KLimchouk
Ukrainian Institute of Speleology and Karstology ,
Simferopol, Ukraine
REGIONAL GEOLOGY AND
HYDROGEOLOGY
The host gypsum bed, ranging from a few meters to
more than 40 m in thickness, is the main component
of the Miocene evaporite formation that girdles the
Carpathians to the northeast, from the Nida River basin
in Poland across the Western Ukraine and Moldova to
the Tazleu River basin in Romania. The gypsum occurs
on the southwestern edge of the Eastern European
platform, where it extends along the Carpathian fore¬
deep for over 300 km in a belt ranging from several
kilometers to 40—80 km wide (Fig. 1A,B). It occupies
over 20,000 km 2 , together with some separated areas
that occur to the northeast of the unbroken belt.
Most Miocene rocks along the platform margin rest
on the eroded terrigenous and carbonate Cretaceous
sediments. The Miocene succession comprises deposits
of Badenian (Tortonian) and Sarmatian age. The Lower
Badenian unit, beneath the gypsum, includes mainly car¬
bonaceous, argillaceous, and sandy beds (30—90 m thick)
adjacent to the foredeep, and these grade into rocks of
calcareous biothermal and sandy facies (10—30 m thick)
toward the platform interior. The Miocene gypsum bed
is variable in structure and texture. Most commonly it
grades from microcrystalline massive gypsum in the
lower part through to variably grained bedded gypsum
in the middle, to coarsely crystalline rock in the upper
horizon. A layer of evaporitic and epigenetic limestone,
locally called Ratynsky, commonly overlies the gypsum.
This layer ranges from half a meter to more than 25 m
in thickness. The gypsum and the Ratynsky limestone
comprise the Tyrassky Formation, which is overlain by
the Upper Badenian unit, comprised either by argilla¬
ceous and marly limestones and sandstone or, adjacent
to the foredeep, by marls and clays of the Kosovsky
Formation. In the platform interior parts of the gypsum
belt the Kosovsky Formation is replaced by the Lower
Sarmatian clays, up to 40—50 m thick. The total thickness
of the capping marls and clays ranges from 40—60 m in
the platform interior to 80—100 m or more in the areas
adjacent to the regional faults that separate the platform
edge from the foredeep.
The present distribution of Miocene formations
and the levels of their denudation stripping vary in a
regular manner from the platform interior toward the
foredeep (Fig. 1C). The Tyrassky Formation dips 1—3°
toward the foredeep and is disrupted by block faults
in the transition zone. To the south and southwest of
the major Dniester Valley, large tectonic blocks drop
down as a series of steps, the thickness of the clay
overburden increases, and the depth of erosional
entrenchment decreases. Along the tectonic boundary
with the foredeep the Tyrassky Formation drops to a
depth of 1000 m or more. This variation, the result of
differential neotectonic movements, and incision of the
Dniester Valley and its north subparallel tributaries
during the Plio-Pleistocene, played an important role
in the hydrogeological evolution of the Miocene aqui¬
fer system, and resulted in the differentiation of the
platform edge into four zones (Fig. 1A,C). The gypsum
was entirely removed by denudation within the first
zone, but the other three zones represent distinct
types of intrastratal karsts: entrenched, subjacent, and
deep-seated (for more details on the evolutionary
classification of karst types see this book's "Gypsum
Caves" and "Speleogenesis, Hypogenic" entries). The
gypsum bed is largely drained in the entrenched karst
zone, is partly inundated in the subjacent karst zone,
and remains under artesian confinement in the deep-
seated karst zone.
In hydrogeological terms the region represents the
southwestern portion of the Volyno-Podolsky artesian
Encyclopedia of Caves.
827
© 2012 Elsevier Inc. All rights reserved.
828
FIGURE 1 Distribution, geological conditions, and types of gypsum karsts in the Western Ukraine. (A) Zones of different evolutionary types of
karsts: I = the gypsum entirely denuded; II = entrenched karst; III = subjacent karst; IV = deep-seated karst. (B) Distribution of the gypsum stratum
and large caves: 1 = the fringe of the Eastern European platform; 2—3 = the Carpathian foredeep: 2 = outer zone; 3 = inner zone; 4 = the Carpathian
folded region; 5 = sulfate rocks on the platform. Tectonic boundaries include: 6 = platform/foredeep; 7 = outer/inner zone of the foredeep;
8 = foredeep/folded region; 9 = other major faults; 10 = flexures; 11 = sulfur mineralization; 12 = sulfur deposits; 13 = gas deposits; 14 = oil depos¬
its; 15 = large maze caves in the gypsum. (C) Geological profile and zones of different types of karsts along the line A—B on the panel B.
ENCYCLOPEDIA OF CAVES
UKRAINE GIANT GYPSUM CAVES
829
TABLE 1 Morphometric Parameters of Large Gypsum Caves in the Western Ukraine
No.
Cave
Length,
km
Average Cross-section Area,
m 2
Density of Passages,
km/km 2
Areal Coverage,
%
Cave Porosity,
%
1
Optymistychna
232.0
2.8
147
17.6
2.0
2
Ozerna
131.4
6.0
150
44.6
5.0
3
Zoloushka
92.0
8.0
142
48.4
3.8
4
Mlynki
36.0
3.3
141
37.6
3.4
5
Kristal'na
22.6
5.0
169
29.2
6.0
6
Slavka
9.1
3.7
139
27.6
3.4
7
Verteba
8.5
6.0
118
34.7
12.0
8
Mushkarova
Jama
5.05
n.a.
n.a.
n.a.
n.a.
9
Gostry Govdy
3.58
1.7
270
17.5
4.0
10
Atlantida
2.52
4.5
168
30.0
4.0
11
Bukovinka
2.4
2.5
120
21.5
4.4
12
Ugryn
2.12
3.8
177
33.3
5.7
13
Jubilejna
1.5
2.3
278
37.0
4.0
14
Komsomol'ska
1.24
2.1
177
24.3
3.0
15
Dzhurinska
1.13
2.4
126
17.8
2.0
basin. The Sarmatian and Kosovsky clays and marls
serve as an upper confining sequence. The lower part
of the Kosovsky Formation and the limestone bed of
the Tyrassky Formation form the original upper aquifer
(above the gypsum), and the Lower Badenian sandy
carbonate beds, in places together with Cretaceous sedi¬
ments, form the lower aquifer (below the gypsum),
the major regional one. The hydrogeologic role of the
gypsum unit has changed with time, from initially
being an aquiclude, intervening between two aquifers,
to a karstified aquifer with well-developed conduit
permeability. The regional flow is from the platform
interior, where confining clays and the gypsum are
largely denuded, toward the large and deep Dniester
Valley and the Carpathian foredeep. In the northwest
section of the gypsum belt the confined conditions pre¬
vail across its entire width. In its wide southeast section
the deeply incised valleys of the Dniester and its left-
hand tributaries divide the Miocene sequence into a
number of isolated, deeply drained interfluves capped
with the clays (Podolsky area; Fig. IB). This is the
entrenched karst zone where most of the explored, pres¬
ently relict, maze caves are located. To the south-south-
east of the Dniester (Bukovinsky area) the gypsum
remains largely intact and is partly inundated (the subja¬
cent karst zone). Further in this direction, as the depth
of the gypsum below the clays increases and entrench¬
ment decreases, the Miocene aquifer system becomes
confined (the deep-seated karst zone). In this zone the
groundwater flow pattern includes a lateral component
in the lower aquifer (and in the upper aquifer, but to a
lesser extent) and an upward component through the
gypsum in areas of potentiometric lows, where exten¬
sive cave systems develop, as evidenced by numerous
data from exploratory drilling.
PATTERNS AND MESOMORPHOLOGY
OF CAVES
Fifteen large caves over 1 km in length are known in
the region (Table 1). Most of them are located north of
the Dniester River. Two other large caves, Zoloushka
and Bukovinka, occur in the Bukovinsky region, near
the Prut River and the border with Moldova and
Romania, generally in the area of artesian flow within
the Miocene aquifer system but within local, particu¬
larly uplifted blocks, where entrenchment into the
upper part of the gypsum caused unconfined (water-
table) conditions to be established in the Holocene.
Most of the caves have only one entrance, either
through collapse dolines and swallow holes at the
interfluves or from gypsum outcrops in the slopes of
the major valleys. Some caves and their entrance series
were known to local people since long ago ( e.g .,
Ozerna, Kristal'na, Mlynki, Verteba), but others were
ENCYCLOPEDIA OF CAVES
830
FIGURE 2 Maps of some of the main maze caves in
the gypsum of the Western Ukraine: (A) Ozerna Cave,
(B) Kristal'na Cave, (C) Slavka Cave, (D) Mlynki Cave,
(E) Zoloushka Cave. Courtesy of the speleological clubs of
Ternopil, Chernivtsy, and Kiev.
discovered by cavers via digs (e.g., Optymistychna,
Slavka, Mushkarova Jama, Atlantida). Two caves
(Zoloushka and Bukovinka) became accessible when
opened by gypsum quarries. Systematic cave explora¬
tion and mapping in the region began in the 1960s.
All of the large gypsum caves in the region are mazes
developed along vertical and steeply inclined fissures
arranged into multistory laterally extensive networks.
Aggregating passages form lateral two- to four-story
systems that extend over areas of up to 2 km 2 (Fig. 2).
A notable feature of the mazes is the exceptionally high
passage network density, which is characterized conve¬
niently by using the ratio of a cave length to an area occu¬
pied by a cave system. This parameter varies from 118
(Verteba Cave) to 270 km/km 2 (Gostry Govdy Cave),
with the average value for the region being 164 km/km 2 .
Values of areal coverage and cave porosity (fractions
of the total area and volume of the rock within a cave
field, occupied by passages) vary for individual caves
from 17.5 to 48.4% (average 29.5%) and from 2 to 12%
(average 4.5%), respectively, being roughly an order of
magnitude greater than these characteristics for typical
epigenic caves. Optymistychna Cave (Optimisticheskaya
in Russian spelling) is the longest gypsum cave, and
the second-longest cave of any type known in the
world, with more than 232 km of passages surveyed
(Fig. 3). By area and volume the largest caves are Ozerna
(330,000 m 2 and 665,000 m 3 ) and Zoloushka (305,000 m 2
and 712,000 m 3 ), followed by Optymistychna Cave
(260,000 m 2 and 520,000 m 3 ).
Maze caves in the region are hypogenic in origin,
developed under confined conditions, due to upward
transverse groundwater circulation between aquifers
below and above the gypsum beds (Klimchouk, 2000).
ENCYCLOPEDIA OF CAVES
UKRAINE GIANT GYPSUM CAVES
831
Optimistychna Cave
FIGURE 3 Map of the Optymistychna Cave, the larg¬
est gypsum cave in the world. Courtesy of the speleological
club of Lviv.
They richly demonstrate morphological features charac¬
teristic for hypogene speleogenesis (Klimchouk, 2009).
These features are related in suites indicating rising
flow through caves, with considerable buoyancy disso¬
lution effects (see Fig. 7 in the article "Speleogenesis,
Hypogenic"). According to the morphology, arrange¬
ment, and hydrologic function of cave mesoforms during
the main (artesian) speleogenetic stage, three major com¬
ponents can be distinguished in the cave systems:
1. Feeding channels. These are the lowermost
components in a system: vertical or subvertical
conduits through which water rose from the
subgypsum aquifer to the master passage networks.
Such conduits are commonly separate but
sometimes they form small networks at the
lowermost part of the gypsum. The feeding
channels join master passages located at the next
upper level and scatter uniformly through their
networks (see Fig. 3 in the article "Gypsum
Caves").
2. Master passages. These are horizontal passages that
form laterally extensive networks within certain
horizons in the middle part of the gypsum bed.
They receive dispersed recharge from numerous
feeding channels and conduct flow laterally to the
nearest outlet feature.
3. Outlet features. These include domes, cupolas, and
vertical channels (dome pits) that rise from the
ceiling of the master passages to the bottom of the
overlying bed. They discharge water from cave
systems to the overlying aquifer.
Also common are raising wall channels above
feeders, continued through ceiling half-tubes to nearest
outlets. Where many smaller ceiling channels are
braiding closely, they leave solutionally sculptured
pendants in-between as they curve into the ceiling. All
of these features are the result of buoyancy dissolution
in the natural convection system.
In areas where water table conditions remained
stable within the gypsum bed for prolonged time
ENCYCLOPEDIA OF CAVES
832
u
FIGURE 4 Conceptual model of hypogene
transverse speleogenesis in gypsum in the Western
Ukraine.
during the subjacent karst stage, enhanced dissolution
along the water table caused extensive horizontal
notching in passage walls and considerable widening
of passages.
The modern dissolution in the entrenched karst zone
is limited to points of leakage from the aquifer perched
on the cover clays, which creates vertical solution pipes,
and also to courses of streams swallowed via ponors.
CAVE SEDIMENTS^AND FORMATIONS
The predominant sediments in the maze caves of
the region are successions of fine clays, with minor
beds of silty clays. These fill passages to a variable
extent and can reach 5—7 m in thickness. Breakdown
deposits are also common. They include chip, slab,
and block breakdown material from the gypsum,
as well as more massive breakdown from the overly¬
ing formations. Calcite speleothems (stalactites, stalag¬
mites, flowstones, and helictites) occur locally in zones
of vertical water percolation from the overlying forma¬
tions. Gypsum crystals of different habits and sizes are
the most common cave decorations. They are largely
of subaerial origin. Hydroxides of Fe and Mn occur
as powdery layers within the clay fill of many caves.
indicating repeated transitional cycles from a reducing
to an oxidizing geochemical environment. Massive
deposition of Fe/Mn compounds in the form of pow¬
dery masses, coatings, stalactites, and stalagmites has
occurred in Zoloushka Cave, where a rapid dewater¬
ing caused by groundwater abstraction during the
past 50 years gave rise to a number of transitional geo¬
chemical processes, some of which appear to show
considerable microbial involvement.
SPELEOGENESIS
The Western Ukrainian maze caves are the model
examples of the hypogene transverse speleogenesis
(Fig. 4) (Klimchouk, 2000, 2007). This type of speleogen¬
esis in the Podolsky region took place mainly during
the late Pliocene through to the middle Pleistocene. It
was induced by initial incision of the Dniester valley
and its left-hand tributaries into the confining clays, and
respective activation of the upward cross-formational
groundwater flow in the underlying artesian system.
Breaching of artesian confinement and further incision
of the valleys during the middle Pleistocene caused sub¬
stantial acceleration of groundwater circulation within
the Miocene artesian system. The majority of passage
ENCYCLOPEDIA OF CAVES
UNDERWATER CAVES OF THE YUCATAN PENINSULA
833
growth probably occurred during this transitional
period. Where the water table was established in the
gypsum for a prolonged time, further widening of pas¬
sages occurred due to horizontal notching at the water
table. With the water table dropping below the lower
gypsum contact, cave systems in the entrenched karst
zone became largely relict. Hypogene cave development
under confined and semiconfined conditions continues
today within the zones of deep-seated and subjacent
karst.
See Also the Following Article
Gypsum Caves
Bibliography
Klimchouk, A. B. (2000). Speleogenesis of great gypsum mazes in the
Western Ukraine, in Klimchouk, A. B., Ford, D. C., Palmer, A.
and Dreybrodt, W., Eds., Speleogenesis: Evolution of karst aquifers,
Huntsville, Alabama: National Speleological Society, p. 261—273.
Klimchouk, A. B. (2007). Hypogene Speleogenesis: Hydrogeological
and Morphogenetic Perspective. National Cave and Karst
Research Institute. National Cave & Karst Research Institute
Special Paper 1, Carlsbad, New Mexico.
Klimchouk, A. B. (2009). Morphogenesis of Hypogenic Caves.
Geomorphology 106, 100—117.
UNDERWATER CAVES OF THE
YUCATAN PENINSULA
James Q. Coke, TV
The Woodlands, Texas, U.S.A.
The geologic history of the Yucatan Platform is diffi¬
cult to establish due to the paucity of deep well data,
exposed anatomic strata, and relative inaccessibility of
the inland terrain. In a quest for potential oil reserves,
PEMEX (Petroleos Mexicanos) drilled 10 exploratory
wells into this platform during the 1970s. Most of these
wells were drilled in the state of Yucatan, often in
excess of —2000 m. Data from the test wells summarize
the stratigraphy of the isthmian portion of the Yucatan
Peninsula as that of a great limestone platform (Ward
and Weidie, 1978). Paleozoic metasediments form the
basement strata at varying depths exceeding —2400 m.
The Cretaceous period launched the onset of numer¬
ous marine transgressions, submerging much of the
isthmus under warm shallow seas until the Pleistocene.
Over 2000 m of limestone strata were formed in the
period, including the uppermost (Miocene—Pliocene)
Carrillo Puerto Formation (>300 m thick). Uplift in
the south-central area of the isthmus began in the
Oligocene, encouraging the Carrillo Puerto Formation
to be deposited in stages over the ever-expanding
margins of the peninsula. This slow uplift continued
until the Pleistocene. Pleistocene deposits are the result
of marine transgressions during the Illinois and
Wisconsin ice ages. Most Quaternary deposits on the
isthmus are restricted to the present margins of the pen¬
insula. These deposits are typically thin, extending one
to three kilometers inland from the Caribbean.
Unconsolidated Holocene deposits are restricted to the
present shoreline. As a final consequence, a near-uni-
form deposition of carbonate strata has produced a low
relief plain over much of the peninsula. Both north and
northeastern areas of the isthmus display a steady
increase in elevation from the coastal periphery (3 m
above sea level) toward the interior flatlands (over 30 m
above sea level). This shared region of the isthmus is
generally acknowledged to contain the highest concen¬
tration of known dry and underwater caves.
As a relatively stable karst platform, the peninsula
contains very few surface drainage systems or lakes. It
is also highly vulnerable to the dynamics of speleogen¬
esis. Capped by a scrub tropical forest, the peninsula
receives over 150 cm of precipitation per year
(Tulaczyk et al., 1993). Approximately 10% of this rain¬
water filters rapidly through the topmost calichified
zone and into lower fractured strata; the remainder
is lost to evapotranspiration. Reaching the shallow
freshwater aquifer, it begins a journey toward the
Caribbean and Gulf of Mexico that is guided in part
through bedding fractures and joints. The process of
freshwater aquifer recharge in Yucatan has facilitated
dissolution within the parent limestone since its initial
uplift from the ocean. However, local dissolution is
complicated through interaction with the base salt¬
water intrusion that buoys the lens of freshwater. The
underwater cave environment can embody as many as
3 chemically dissimilar water layers. This includes a
thin upper fresh-cap layer that is independent of the
lower freshwater column during normal meteorologi¬
cal conditions. A middle freshwater layer contains
varying amounts of dissolved chlorides. Depending on
tunnel depths and distance from the coast, this brack¬
ish stratum inundates horizontal underwater caves,
deep underwater pits, and coastal or offshore dis¬
charge vents to different depths within the caves' par¬
ent limestone beds. The base of the water column is
occupied by a near-static anoxic saltwater intrusion.
Today the Mexican Yucatan Peninsula is divided
into three administrative states: Campeche, Yucatan,
and Quintana Roo. Each of these states lays claim to a
variety of dry and underwater caves that display a
multiplicity of environmental and physical conditions.
Underwater cave explorations are not well documented
in Campeche. Exploration teams must be self-sufficient
when providing for compressed gas supplies and
rechargeable electrical devices. Access to the interior of
Encyclopedia of Caves.
© 2012 Elsevier Inc. All rights reserved.
834
u
Campeche is also a challenge due to a limited network of
roads. With few expeditions exploring the interior of
Campeche, it is understandable why limited underwater
cave discoveries are reported to this date.
The state of Yucatan in the north is well respected
for both its dry caves and cenotes (from the Maya
word dzonot). Unlike many of the explored cenotes in
Quintana Roo, the majority of cenotes in Yucatan takes
the shape of flooded classic sinks, or pits. They are rec¬
ognized for their vertical relief rather than horizontal
scope, due to the mechanism of their formation.
Dissolution of deep strata and past sea-level fluctua¬
tions stimulated an eventual collapse of surface strata,
creating near-circular pools of water at the jungle sur¬
face. In profile, most Yucatan cenote entrances crown
an ever-widening bell-shaped pit which is floored by a
central organic debris and breakdown mound. The
size of the collapse is symbolized by the shape, size,
and ultimate depth of the underwater pit. The deepest
sinks in Yucatan offer a direct descent to great depths
from the cenote surface. As an explorer descends
through the freshwater layer, the walls of the cave pit
recede and disappear into a gloomy distance. At great
depths, the explorer's only visual reference within the
void is the dull glow of sunlight from the cenote
entrance above. At —70 m, explorers arrive at the
boundary zone between freshwater and salt water
layers. The halocline is a distinctive density interface.
When disturbed, the mixing of freshwater and salt¬
water layers create temporary, yet dazzling visual dis¬
tortions within the immediate mixing area. A white or
red hydrogen sulfide cloud can also be encountered at
the interface. This corrosive layer is a by-product of
organic digestion by anaerobic bacteria that live in the
lower anoxic saltwater intrusion. A "smell" of rotten
eggs is often detected by divers as they descend
through this thin chemical zone. Absorbed through the
skin, high concentrations of hydrogen sulfide will pro¬
duce headaches and nausea for the diver. Should a
hydrogen sulfide boundary be passed for greater
depths, the chemical cloud will effectively block any
remaining natural light from the cenote entrance. The
deepest areas in Yucatan's classic sinks are attained at
the junction of the bottom of the debris cone and the
surrounding cave wall.
Classic sinks in Yucatan contain small horizontal
freshwater tubes within the first 10 m of the freshwater
column. Substantial horizontal cave development, such
as that found in coastal caves, is not recorded at any
depth within the classic sink environment. Modern
explorers are exploring many of these deep-pit cenotes
for the first time. The deepest explored classic sink cen¬
otes in Yucatan include Cenote Sabak Ha (—147 m),
Cenote Xkolac (—121 m), and Cenote Ucil (—118 m). The
final depth of Cenote Sabak Ha sink ("smoke-water" in
Maya) remains a mystery. Only two underwater caves
in Yucatan display significant horizontal passage devel¬
opment. Further explorations in Cenote Chacsinikche
(670 m) and Cenote Kankirixche (438 m) are serious
undertakings due to the depth and dimensions of their
outlying tunnels.
The state of Quintana Roo has received the greater
share of attention from cave researchers during the
past 27 years. The state contains three of the five lon¬
gest caves in Mexico, along with the longest cave in
Mexico. Sea-level fluctuations during the Sangamon
Interglacial through the Wisconsin postglacial periods
expedited numerous developments within subterra¬
nean drainage networks, as demonstrated by the
underwater caves on the eastern coast. Receding sea
levels during active glaciation resulted in a consider¬
able drop in the local water table and, thus, water in
the subsurface drainage channels. Unsupported cave
ceilings immediate to the surface collapsed, creating
karst windows at irregular intervals along the course
of the conduits. Streambed erosion and dissolution
gradually expanded the existing caves while eroding
new passages into deeper limestone beds (currently
explored to —120m within horizontally developed
caves). About 18 ka BP sea level began a slow rise from
125 m below its present level as the Wisconsin glacia¬
tion period declined. As the freshwater table was ele¬
vated by a rising salt water intrusion, deeper conduits
were rapidly immersed by this intrusion.
Holthuis (1973) coined the term anchialine (also
known as anchihaline) to define specific attributes of
inland, seemingly isolated aqueous pools. Many cave
systems in eastern Quintana Roo contain anchihaline
pools; those that contain salt or brackish waters that
fluctuate with oceanic tides yet lack a surface connec¬
tion to the ocean. Anchihaline caves in Quintana Roo
contain an upper freshwater layer that can flow over a
near-static saltwater intrusion. Unlike the near-static
freshwater layer in classic sinks, a freshwater flow
courses through horizontal cave tunnels as it travels
toward the ocean. A dynamic mixing zone occurs
(enlivened by water current and passage configura¬
tion) at the halocline density interface. The depth of
the halocline and thickness of the freshwater layer
increases with distance from the ocean, while periods
of sustained precipitation, ocean tides, and the per¬
vasiveness of exposed mixing zones in cave pas¬
sages instigate random fluctuations of the halocline.
Investigations conclude that preferential dissolution
of limestone at the mixing zone (Back et al., 1986),
allied with native fracture zone speleogenesis, has pro¬
duced an arena for vast growth and passage complex¬
ity within these cave systems. Volume of recharge,
the availability of coastal discharge centers, and a con¬
stant supply of freshwater for the mixing zone also
ENCYCLOPEDIA OF CAVES
UNDERWATER CAVES OF THE YUCATAN PENINSULA
835
* 5*
Xel-Ha Lagoon to
Sian Ka'an Biosphere
Reserve Region
Quintana Roo, Mexico
Wmag
2010
5 km
Sistema
Ox Bel Ha
Sistema
Sac Actun
Laguna de
Xel-Ha
J. Coke; 2010
Caribbean
Sea
FIGURE 1 Map of underwater caves Xel-Ha to Sian Ka'an Biosphere Reserve region, Quintana Roo. Drawing by ]. Coke, from 2010 data sup¬
plied by the Quintana Roo Speleological Survey.
appears to be decisive for the substantial growth of
Quintana Roo caves (Smart et al., 2006). Coupled with
Pleistocene sea-level fluctuations, a snapshot of histori¬
cal speleogenesis would incorporate both solutional
dissolution mechanisms with preferential dissolution
by halocline migration through the depths of the cave
passage. With the addition of karst windows along the
systems' corridors, injections of organic debris enter
the caves within a wide range of locations to enrich
both chemical and nutrient microenvironments in the
caves.
A persistent assemblage of speleogenic processes is
responsible for creating a multifaceted coastal region
that is riddled by underwater caves and microenviron¬
ments. A section of this cave region surrounds the city
of Tulum on all flanks, while paralleling the coastline
for a total distance of 36 km (Fig. 1). The present extent
of inland cave development is presently limited to a
12-km distance from the coast. Cave passage configura¬
tion within the coastal region of Quintana Roo is ori¬
ented from the northwest to southeast. This orientation
is also representative of the established inland-to-coastal
flow of freshwater drainage. Why the 12-km boundary
exists has perplexed underwater explorers for many
years. Coastal horizontal cave formation appears to sur¬
render to interior classic sink patterns at this frontier.
The first explorers to confront this boundary reported
large, fracture controlled passages just as they reached
the terminal end of the tunnel. Swimming against
strong freshwater currents, they would describe 100-m-
wide tunnels decorated by huge flowstone columns, cal-
cite rafting floors, and slender stalactites hanging from
6-m-high ceilings (Fig. 2). Without warning, this trunk
passage would end abruptly at a vertical wall of lime¬
stone. As exploration efforts continued in other areas, a
collection of unassociated cave systems were surveyed
to an equal distance from the coast. Each of their cave
passages terminated at an all-too-familiar limestone bar¬
ricade. The eastern edge of the Holbox Fracture Zone
has proved to be a formidable opponent for coastal
cave explorers. This fracture zone extends south near
the island of Holbox to Laguna Chunyaxche near the
town of Muyil. We have discovered that this fracture
appears to play a significant role within the local
ENCYCLOPEDIA OF CAVES
836
u
FIGURE 2 Representative cave passage and speleothems in Sistema K'oox baal, Quintana Roo, Mexico. Speleothems deposited during
Pleistocene sea-level low stands. Photographs courtesy of Radoslav Husah. Used with permission.
hydrology. At each intersecting passage with the frac¬
ture zone, today's explorers are quick to note an abrupt
change in the freshwater currents. In each instance, a
strong flow of freshwater is recorded, emanating from
the south. This flow anomaly can be tracked over a
25-km distance through five widely separated cave
systems. It appears that the Tulum Region's aquifer
is partially supplied from a yet-to-be-determined
freshwater source south of Laguna Chunyaxche. The
implications concerning the nuances of the local fresh¬
water aquifer are fascinating. It is a worthy comple¬
ment to the pristine microenvironments that are
supplied by these unknown freshwater supplies.
As the aquifer begins its 12-km traverse to the
Caribbean through limestone, cave passage configura¬
tions and their environmental niches change over the
course of the drainage pattern. Hundreds of natural
openings are encountered along this watercourse. In
earlier times these natural openings were vital centers
of activity for pre-Maya populations. Radiocarbon
( 14 C) analyses date one set of human remains in
Sistema Naranjal (an underwater cave in Quintana
Roo) to over 10 ka BP in age. Three additional pre-
Maya skeletons are documented from the Naranjal
Cave, while another four skeletal sets are under study
in nearby underwater caves. During a period of low¬
ered sea levels, proto-explorers gained entrance to
many caves due to their position above the water table.
Whether they used caves for shelter, a source for
drinking water, or as a spiritual focal point is a matter
of speculation. It was the rise of the Maya civilization
that brought a collective essence to Yucatan's caves.
The Maya were relentless dry cave explorers, at times
traversing hundreds of meters of intricate dry crawl-
ways and galleries with little more than a smoldering
torch. Leaving a legacy of huge stone structures, cave
paintings, rock art, and pottery to mark their explora¬
tions, modern researchers are convinced the Maya
used caves and cave entrances as centers for spiritual
vitality and political advantage into the twentieth cen¬
tury. Sacrificial pots, small structures, and petroglyphs
from past Maya activities continue to be documented
at cenote and dry cave sump entrances to underwater
caves.
Today cenotes play another role within the environ¬
ment of the underwater caves. As a window to the jun¬
gle, terrestrial organic materials are able to invade
limited sections of the watercourse to enrich the caves'
tenuous ecological niche. Each of these openings to the
heart of underwater caves is analogous to bat popula¬
tions' stimulus to the biological web in dry caves.
Cenotes are a point source for organic fuels that sus¬
tain many of the 41 species of cave-adapted stygo-
bionts on the Yucatan isthmus. They also nourish
jungle mammal, reptile, amphibian, and avian popula¬
tions. As underwater passages progress toward the
ENCYCLOPEDIA OF CAVES
UNDERWATER CAVES OF THE YUCATAN PENINSULA
837
ocean, each cenote opening supports a succession of
terrestrial, stygophile, and stygobiont species. The
aquifer's water chemistry is altered during this jour¬
ney. Flowing at an approximate velocity of 250 m per
day, the aquifer is transformed from a substantial
brackish body of water to a thin and near-saline layer
at coastal discharge centers. Not only does this trans¬
formation affect local biological populations, it can
degrade the parent limestone strata of the conduits.
Distinct underwater cave passage begins to disinte¬
grate into a Quaternary mix of breakdown and uncon¬
solidated strata. Erosion by the halocline can also give
the impression of "bedding plane" dissolution. Daily
tidal fluctuations encourage the halocline to focus its
authority on a 45-cm section of strata, creating a
sharply defined horizontal void.
Two extensive caves located in the Tulum and Xel-
Ha regions best illustrate the complex nature of
underwater caves in Quintana Roo. Sistema Ox Bel Ha
(in Maya, "Three Paths of Water") is the longest cave
in Mexico, and the longest underwater cave in the
world. It contains over 231 km of surveyed passage
that integrates four known ocean-discharge vents,
and over 137 inland cenote entrances within 52 km 2 .
Extending nearly 9 km inland from a series of small
offshore vents, the cave traverses beneath a dense
coastal mangrove zone that rises to a scrub jungle
interior that is peppered by karst windows and agna-
das (temporary pools of surface water). However, the
highest density of cave development is found in
a centralized maze along major drainage conduits
within a 38-km 2 area.
As the Ox Bel Ha Cave spreads inland from the
Caribbean Sea, it transforms its overall structure to
reflect two of three general classes of geological and
environmental conditions that typify most underwater
caves of the region. These classes may be subjectively
ordered (progressing inland) as caves found in recent
coastal strata, caves that are developed in older strata
1 to 6 km from the Caribbean, or those underwater
caves that impinge and terminate on the Holbox
Fracture Zone (10—12 km inland). The sections of Ox
Bel Ha within a kilometer of the coast tend to settle
into four independent maze areas. Their passages are
formed in unconsolidated Recent deposits, giving way
to more substantial Quaternary strata farther inland.
Each maze is arranged around one or two dedicated
conduits that are engaged in directing freshwater
drainage toward the ocean. An extensive mixing zone
between the freshwater and saltwater layers creates an
elevated concentration of chlorides with these sections.
Passages are often barren of speleothems, given the
advanced degree of dissolution and limited geological
history in this area. They are likely formed along small
fractures and joints; however, conduit erosion makes
this difficult to discern. Fine clay-like silts cover the
floors and ceilings of the cave, while the parent lime¬
stone is often frangible and unstable. Unlike other
underwater caves proximal to the Caribbean, few if
any conduits attempt to parallel extant coastal topogra¬
phy through linear fractures.
Fracture-based speleogenesis is evident within the
primary discharge conduits as they advance farther
inland. At 1 km from the ocean. Ox Bel Ha assumes
an entirely different character than that of its coastal
temperament. Cave formation in this section is based
on a northwest—southeast pattern of conduit evolu¬
tion that is present in many of the area's underwater
caves. The parent cave strata become more massive
and stable, the mixing zone between water layers
evolves into a more distinct halocline, and spe¬
leothems become more common within the well-
defined passages. This area transitions into older cave
passage and limestone strata. This part of the cave is
also thought to have longer dry periods during
Pleistocene glaciations. Many of the cave tunnels are
situated within 20 m of the land surface; deeper con¬
duit development is not unusual although it is not
widespread.
Lateral branching of fault-controlled tunnels occurs
in this section. Secondary tunnels may rejoin their
original passage or become an independent subsidiary
section. In rare instances, they can connect two trunk
passage sections of Ox Bel Ha that are separated by
over a kilometer. It is not clear if these distant connec¬
tions serve as conduits to pirate water from other
areas of the cave, or if they are a product of simple
fracture evolution. Numerous karst windows become
highly engaged within this lateral passage develop¬
ment. Inline flowing sinks such as Cenote Esmeralda
or Cenote So'sook are focal points in the fracture
patterns.
Sistema Sac Actun (in Maya, "the White Cave") is
the second longest cave in Mexico, and the second lon¬
gest underwater cave in the world. It contains over
215 km of surveyed underwater passage, and an addi¬
tional 2 km of dry passage. The Sac Actun cave inte¬
grates two known discharge vents to the ocean, and
over 162 inland cenote entrances. Unlike the more cen¬
tralized development of Sistema Ox Bel Ha, Sistema
Sac Actun parallels the east coast of Quintana Roo
through a complex set of interconnected high-density
cave centers. Reaching over 8 km inland through frac¬
ture- and dissolution-controlled passages. Sac Actun
extends laterally nearly 12 km between its most north¬
ern and southern survey stations. It incorporates a
highly complex inland drainage system that plunges
toward the coast at established freshwater discharge
points. Given its geographical range, the cave traverses
the first two of the three general classes of geological
ENCYCLOPEDIA OF CAVES
838
u
and environmental conditions that typify most under¬
water caves of the region.
Two coastal sections of Sac Actun tend to settle into
localized maze configurations once the immediate dis¬
charge tunnels are passed. Conduits in the Cenote
Abejas area attempt to parallel coastal topography
through linear fractures in Recent depositional strata.
Similar to the coastal sections of Sistema Ox Bel Ha,
the high concentration of chlorides in the mixing
zone produces low horizontal tunnels ("bedding
plane") and ill-defined conduits that are broken by
periodic fracture-controlled rooms. Also as in Sistema
Ox Bel Ha, these passages are barren of speleothems,
covered in fine clay-like silts, and developed in
unstable limestone. Hydrologically active passages
may be complicated by deformations in a nearby
Pleistocene ridge, or through young Holocene depos¬
its that are positioned between the coastline and the
Pleistocene ridge.
As the cave migrates inland from its discharge
areas, it passes beneath the Pleistocene ridge (~200 m
to 1 km from the coast). The ridge produces a rise in
topographic elevation (~10 m) that accompanies a
change to older more consolidated limestone strata.
Clear of local strata disturbances created by the ridge,
much of Sac Actun begins to adopt an entirely differ¬
ent character from that of the coastal region. Fracture
and fault-based speleogenesis becomes evident within
primary and some secondary discharge conduits as
they progress inland. Cave formation in this section
turns to a northwest—southeast pattern of conduit evo¬
lution. The parent cave strata become more massive
and stable, the cave passage becomes clearly defined,
and lateral branching of fault-controlled tunnels occurs
frequently. Most conduits are situated within 15 m of
the land surface in this area of the cave.
At 3 to 6 km from the coast, both northern (Nohoch
Nah Chich) and southern (Sac Actun-Naval) arms of
Sac Actun enter or impinge on Tertiary (Miocene-
Pliocene) strata. Although some lateral branching of
fault-controlled tunnels occurs in these areas, cave
development is often limited to one or two primary dis¬
charge tunnels. Most secondary tunnels are generally
limited in length and small in size; they often end in
impassable cracks or tubes that discharge freshwater.
Explorations in the northern section of Sistema Sac
Actun are now within 5 m of connecting to Sistema
Dos Ojos (68 km). Due to the wandering nature of the
cave, the possibilities for expanding Sac Actun are far
greater than those in Sistema Ox Bel Ha.
See Also the Following Article
Exploration of Caves—Underwater Exploration Techniques
Bibliography
Back, W., Hanshaw, B. B., Herman, J. S., & Van Driel, J. N. (1986).
Differential dissolution of a Pleistocene reef in the ground-water
mixing zone of coastal Yucatan, Mexico. Geology, 14, 137—140.
Coke, J. G., IV, Perry, E. C., & Long, A. (1991). Sea level curve.
Nature, 353(6339), 25.
Holthuis, L. B. (1973). Caridean shrimps found in land-locked salt¬
water pools at four Indo-West Pacific localities (Sinai Peninsula,
Funafuti Atoll, Maui and Hawaii Islands), with the description of
one new genus and four new species. Zoologische Verhandelingen,
128, 1-48.
Lutz, W., Prieto, L., & Sanderson, W. (Eds.), (2000). Population, devel¬
opment, and environment on the Yucatan Peninsula: From ancient
Maya to 2030. Laxenburg, Austria: International Institute for
Applied Systems Analysis.
Rissolo, D. (2005). Tancah Cave revisited. Association for Mexican Cave
Studies Activities Newsletter, 28, 78—82.
Smart, P. L., Beddows, P. A., Coke, J., Doerr, S., Smith, S., &
Whitaker, F. F. (2006). Cave development on the Caribbean Coast
of the Yucatan Peninsula, Quintana Roo, Mexico. In R. Harmon
& C. Wicks (Eds.), Perspectives in karst geomorphology, hydrology,
and geochemistry/ (pp. 105—128) (Special Paper 404). Boulder, CO:
Geological Society of America.
Tulaczyk, S. M., Perry, E. C., Duller, C. E., & Villasuso, M. (1993).
Influence of the Holbox fracture zone on the karst geomorphol¬
ogy and hydrogeology of northern Quintana Roo, Yucatan
Peninsula, Mexico. In B. F. Beck (Ed.), Applied karst geology
(pp. 181—188). Rotterdam: Balkema.
Ward, W. C. (1985). Quaternary geology of northeastern Yucatan
Peninsula, Part 2. In W. C. Ward, A. E. Weidie & W. Back (Eds.),
Geology and hydrogeology of the Yucatan and quaternary/ geology of
northeastern Yucatan Peninsula (pp. 23—53). New Orleans, LA:
New Orleans Geological Society.
Ward, W. C., & Weidie, A. E. (Eds.), (1978). Geology and hydrogeology
of northeastern Yucatan New Orleans, LA: New Orleans Geological
Society.
Weidie, A. E. (1985). Geology of the Yucatan Platform, Part 1. In W. C.
Ward, A. E. Weidie & W. Back (Eds.), Geology and hydrogeology of
the Yucatan and quaternary geology of northeastern Yucatan Peninsula
(pp. 1—19). New Orleans, LA: New Orleans Geological Society.
URANIUM SERIES DATING
OF SPELEOTHEMS
Christoph Spotl and Ronny Boch
University of Innsbruck
INTRODUCTION
Entering a cave decorated with various sorts of spe¬
leothems commonly prompts questions about the age
of these formations. It is intuitive to suggest that tall
stalagmites or thick flowstone took a long time to
form, but until the middle of the twentieth century no
method existed that allowed the quantitative determi¬
nation of how old speleothems are. First attempts were
made by physicists in the 1950s who applied the radio¬
carbon ( l4 C) method (discovered in the late 1940s) to
Encyclopedia of Caves.
© 2012 Elsevier Inc. All rights reserved.
URANIUM SERIES DATING OF SPELEOTHEMS
839
FIGURE 1 230 Th/ 23S U activity ratio versus age. A high initial
234 U/ 238 U ratio (expressed in delta units) accounts for an extended
dating range as secular equilibrium is reached after a longer time
period (after Richards and Dorale, 2003). Most speleothems show
6 234 U values between ca. 100 and 1200.
dripstones. It was soon realized, however, that these
results must be viewed with great caution, because the
origin of the carbon atom (utilized for age dating) of
dripstone carbonate is different from that in organic
remains, such as wood. In addition, radiocarbon dates
require a calibration to obtain true calendar ages
(which was not available in the early days of radio¬
carbon dating). In short, 14 C dates of speleothems pro¬
vide only a rough estimate of their formation age and
samples older than ca. 50,000 years cannot be dated
using this method.
The dating method of choice for speleothems, known
as the uranium-thorium (U-Th), or thorium-uranium (Th-
U) method, belongs to the family of uranium-series meth¬
ods and was first applied to speleothems in the early
1970s, although its physical principles were known since
the beginning of the twentieth century. The past four
decades have seen dramatic improvements of this
method and today U-series dating of speleothems is
among the most highly respected chronological
techniques, providing invaluable time information for
studies of the environment and climate of the past.
This article provides a concise summary of the
method for cavers, speleologists, and other people
interested in finding out how the age of speleothems
and their growth rate can be measured, and which
practical aspects are relevant when selecting samples
for analysis.
PRINCIPLES OF U-SERIES DATING
From the three naturally occurring U isotopes
(99.28% 238 U, 0.72% 235 U, 0.005% 234 U). 238 U and 235 U
are the starting points for two different radioactive
decay chains comprising numerous unstable isotopes
toward stable Pb isotopes. The rate of radioactive
decay is statistically constant (decay constant) for a par¬
ticular isotope, but different isotopes decay at different
rates (they have different half-lives). As a consequence,
different isotope ratios can be measured and are equiva¬
lent to a specific time elapsed, that is, the isotopic abun¬
dances depend on the decay laws and this accounts
for an absolute chronology. An accurate knowledge of
the decay constants is essential to calculating the age,
and these constants are subject to ongoing reevaluation
and refinement.
U-Th (disequilibrium) dating determines the relative
abundances of unstable U and Th isotopes and is
one of several methods based on the radioactive decay
starting with the parent isotope 238 U. This particular
method, however, is by far the most important one for
determining the age of speleothems, and it lays the
foundation for their prominent role in paleoclimate
research. Depending on the geochemical characteristics
of the sample ( e.g ., the initial 234 U/ 238 U ratio. Fig. 1)
and the laboratory instrumentation, the U-Th method
allows dating of samples as old as 600,000—700,000
years, after which radioactive decays of the isotopes
“ Th and “ U are in a state of secular equilibrium.
An initial state of disequilibrium between the U and
Th isotopes is a precondition for this method, that is,
U and Th are separated from each other prior to spe-
leothem formation due to the vastly different chemi¬
cal properties of these elements: Th is insoluble in
water and tends to be adsorbed on particles (e.g., clay
minerals); in contrast, U is highly soluble, transported
in solution (drip water) and coprecipitated during
speleothem deposition. As soon as the U atoms are
incorporated into the lattice of the carbonate crystals
radiogenic 230 Th starts to build up from U decay, that
is, the U-Th "clock" starts to run. Under ideal condi¬
tions, the 230 Th content originates solely from the
decay of 234 U and its concentration increases system¬
atically with time. No gain or loss of these isotopes,
other than the intrinsic decay, should have occurred
since the formation of the speleothem, that is, the
system must have remained closed. For this reason,
speleothems showing recrystallization or dissolution
should be avoided. Some speleothems contain an
additional, nonradiogenic Th fraction resulting from
the incorporation of mineral or organic particles. In
these cases, the isotope ratios have to be corrected for
the initial 230 Th contribution using the measured con¬
centration of 232 Th and an initial 230 Th/ 232 Th activity
ratio which is either assumed (e.g., the mean Th/U
ratio of the upper continental crust) or determined
using isochron methods. The extent of detrital
ENCYCLOPEDIA OF CAVES
840
u
contamination may vary significantly within a cave
and also within an individual speleothem.
Regarding U-Th technology, the first laboratories
started in the 1970s and the subsequent evolution was
characterized by two major steps prompted by break¬
throughs in analytical science. The first revolution
occurred in 1986 and marked the introduction of
mass spectrometric techniques (instead of week-long
counting of alpha particles from radioactive decay).
Thermal ionization mass spectrometry (TIMS) allowed
measurement of U and Th isotope abundances with
much better precision and smaller sample sizes and
extended the dating range from ca. 350,000 years to ca.
500,000 years. The second improvement involved the
introduction of a new type of mass spectrometer which
was successfully coupled to an ionizing plasma source.
These machines are referred to as inductively coupled
plasma-mass spectrometers (ICP-MS) with the prefix
"single-collector" or "multi-collector" depending on
the number of particle collectors. Introduced in 1997,
these instruments now dominate many fields of geo¬
chemistry and geochronology. Compared to TIMS, the
ICP-MS technology holds the advantages of a reduced
sample size, shorter analysis time (increased sample
throughput), and an increased sensitivity due to higher
ionization efficiency. Using ideal samples and opti¬
mum instrumental conditions, an upper dating limit
of ca. 700,000 years can today be achieved. Modern
developments include the application of in situ laser
ablation ICP-MS which increases the spatial resolution,
decreases the sample size, and facilitates sample prep¬
aration, but currently suffers from poorer precision
and accuracy.
Before measuring the relevant isotopic ratios on
a mass spectrometer of the carbonate samples—
preferentially taken from pristine crystal fabrics
along the central speleothem growth axis and in
powder form—rather time-consuming, wet-chemical
preparation steps have to be conducted in a dedi¬
cated clean laboratory. U and Th are enriched and
isolated from each other and from the carbonate
matrix using an ion-exchange resin. Ultra-pure acids
and other chemicals including well-characterized
radioactive materials (spikes) are used and this work
therefore requires extreme care.
REPORTING AND ASSESSING
U-TH AGES
Individual speleothem U-Th ages are quoted either
in years ( annum ; a) or, in the case of older samples, in
thousands of years (ka), for example, 34,200 a is equiv¬
alent to 34.2 ka. Traditionally, U-Th ages have been
referred to the year 1950 AD (the reference point for
radiocarbon ages), but more recently other reference
years have also been used, for example, the year 2000
AD (abbreviated as b2k for "before 2 ka"). This infor¬
mation has to be stated clearly when reporting spe¬
leothem ages to avoid confusion, in particular when
dealing with young (Holocene) samples.
In general, each measurement has two kinds of
uncertainties ("errors"): accuracy and precision. The
former determines the deviation from the "true" value,
and requires careful evaluation in the laboratory
involving standard reference materials. Using different
decay constants, for example, also results in systemati¬
cally different results (which may still be extremely
precise). Precision is easier to assess and provides sta¬
tistical information on how far replicate measurements
deviate from the mean value. Recent years have seen
an impressive improvement in the precision of U-Th
data, although it is important to bear in mind that
this does not include the uncertainty associated with
accuracy. Dating precision is traditionally reported at
the 2-sigma uncertainty level, for example, an age of
123.3±1.2ka means that 95% of the sample reruns
fall between 122.1 and 124.5 ka. Some other geochrono-
logical methods report their results at the 1-sigma level,
such as 14 C and burial ages. For example, a bone
fragment yielded a (calibrated) 14 C age of 37,140±300 a
BP and is embedded in flowstone whose U-Th age is
36,400 ±300 a BP. Note that (1) the latter date is twice
as precise as the radiocarbon date, and (2) the speleothem
is statistically younger than the bone at the 1-sigma level;
on the 2-sigma level, however, the two dates are indistin¬
guishable from each other. Speleothem ages determined
by alpha-spectrometry were also reported at the 1-sigma
level; U-Th data produced prior to 1987/88 are quoted
in this way. In addition, these dates were calculated using
decay constants that have since been updated, resulting
in an age-dependent deviation.
What precision can be expected for speleothem sam¬
ples? This question depends on a number of factors
which range from the type and geochemical quality
of the sample to the instrumentation and calibration of
the particular laboratory. For typical speleothem sam¬
ples, most laboratories are capable of measuring ages
with relative errors of 1—2% for samples younger than
ca. 200 ka. For example, a flowstone that formed 120 ka
ago (during the last interglacial) can be routinely dated
to ±1.2 to ±2.4 ka. Some of the leading laboratories
report precisions as low as 0.2—0.3%. So, the analytical
uncertainty of the same flowstone can be reduced to
±240 to ±360 a and a very small number of these labs
have recently begun reporting U-Th ages at "epsilon
precision," that is, better than per-mil level. Even
taking further technological advances into account, it
ENCYCLOPEDIA OF CAVES
URANIUM SERIES DATING OF SPELEOTHEMS
841
FIGURE 2 Drilling a small-diameter core near the base of in situ
stalagmites allows the determination of the approximate starting
date of deposition at this drip site, thereby minimizing the impact on
the cave environment.
is highly questionable if such ultra-high precision will
be routinely possible within the next five years. After
all, the most commonly encountered problem is impu¬
rities in the sample. Although various approaches exist
to correct for the nonradiogenic Th fraction, they are
all based on assumptions and therefore result in larger
age uncertainties. Regardless of the instrumentation
used, the overall uncertainty increases with increasing
age—for example ±30—50 a at 10,000 a, ±250—400 a
at 50,000 a, ±700—1500 a at 100,000 a, ±1000—3000 a at
200,000 a, ±3000—15,000 a at 400,000 a—and reaches
infinite values at the upper dating limit.
Finally, it should be noted that the analytical preci¬
sion achievable using state-of-the-art instrumentation
approaches the resolution imposed by sampling a finite
thickness of speleothem material. Given a stalagmite
with a slow vertical growth rate of 25 pm/a, a 2-mm-
thick sample represents some 80 a of growth.
If sampling is not done carefully, for example, if growth
layers are not visible, these 2 mm might encompass sig¬
nificantly more than 80 a.
PRACTICAL ISSUES
Who Is Allowed to Sample Speleothems?
Each cave and each dripstone is a unique part of
nature which took literally thousands of years to form.
In most countries speleothems are therefore protected
by law. As a consequence, sampling of speleothems
requires special permits and should always be done
with great caution and in close collaboration with local
authorities and/or speleologists.
How to Sample Speleothems in Caves
Most paleoenvironmental studies require the entire
stalagmite to be removed from the cave. As significant
parts of the dripstone are expended during laboratory
analyses it is rarely possible to restore the original sta¬
lagmite after the scientific study has been completed.
Although there are ways to replace the stalagmite with
a replica, most scientific studies have not made such
an attempt. Consequently, the removal of a stalagmite
from a cave is an invasive act which should be evalu¬
ated very carefully.
One way to minimize the impact on the cave and
to screen for relevant stalagmites (e.g., covering a time
period of interest) is to take small in situ samples
by drilling horizontally into the base of a stalagmite
(Fig. 2). U-Th dating of such small drill cores (e.g.,
5 mm in diameter) from the inner part of a stalagmite
provides useful information on the starting time of
growth, and the drill hole can be easily sealed, thereby
causing hardly any impact on the appearance of the
stalagmite. Drilling is also an option for sampling
flowstone deposits, but in this case larger diameter
cores (e.g., 2—3 cm) are preferable so that a whole
series of analyses (in addition to dating) can be per¬
formed. Again, such holes can be sealed so that there
is no impact on the appearance of the dripstone forma¬
tion to visitors of the cave.
How Much Sample Material Is Needed
for an Age Determination?
It is possible to obtain a meaningful age from as
little as 10—20 mg of speleothem material. Most labo¬
ratories, however, prefer larger samples (e.g.,
50—200 mg) to improve precision and possibly run
duplicates. As a rule of thumb, the younger the sam¬
ple and the lower its U concentration, the more sam¬
ple is required to obtain a good result. Aragonitic
speleothems have significantly higher U contents than
the more common calcitic ones, so smaller sample
amounts are needed. The subsamples for U-Th dating
are typically drilled from a polished slab cut from the
axial part of a stalagmite (Fig. 3). Some researchers
prefer to use a thin diamond-coated blade- or wire-
saw to cut discrete aliquots from the stalagmite
instead of using powdered drill samples. This, how¬
ever, requires cutting the axial slab in half as it is
advisable to stay close to the axial zone where
the growth layers are rather thick, as opposed to the
flanks of the stalagmite where they are thinner, and
impurities may be present at higher concentrations.
ENCYCLOPEDIA OF CAVES
842
u
FIGURE 3 Polished slab of a small (13.8 cm high) stalagmite with
several elongated holes drilled to sample calcite for U-Th age deter¬
minations. Note that the sampling holes follow individual growth
layers in order to increase the age resolution. The vertical band along
the growth axis is a milling trench—only fractions of a millimeter
deep—along which numerous subsamples for stable isotopic analy¬
ses were taken. Small holes along growth layers are samples to check
for stable isotopic equilibrium.
How Much Does an Age Determination Cost?
In contrast to, for example, radiocarbon analyses
of organic material, only a few laboratories worldwide
offer U-Th analyses on a commercial basis. Prices typi¬
cally range from ca. US$200—400. The majority of the
U-Th facilities, however, belong to universities or other
academic institutions and their pricing may deviate
significantly.
Why Focus on Stalagmites?
Speleothems, despite their uniform mineralogy
(calcite, less commonly aragonite), occur in a great vari¬
ety of forms, dimensions, and colors, and any of these
can be dated by U-series methods provided their age is
within the dating limit of the particular method. Still,
with some exceptions, the majority of scientific studies
have so far relied on stalagmites rather than stalactites.
flowstones, or other formations. The main reason for
this is that regularly shaped (candlestick) stalagmites
are the only speleothem growth variety that shows a
well-defined internal structure. Flowstone also com¬
monly shows a regular internal growth structure, but is
often characterized by growth interruptions due to
switching of the flow paths on the surface. Stalactites
have also been investigated, but given their complex
(chevron-like) internal structure and the presence of an
internal drip-water feeder channel most researchers
tend not to work with them.
What Is a “Good/Bad” Sample for Dating?
Along the lines of the preceding paragraphs, it
is advisable to stay away from samples that macro-
scopically suggest impure calcite (or aragonite), that is,
those showing a brown or gray stain. A clean, white,
outer surface of actively forming (z.e., drip-water fed)
speleothems is a strong indication of low impurities,
but such a diagnosis may fail on older, inactive forma¬
tions, whose outer surface often disguises the interior
composition. Drilling a basal small-diameter core is
a way of screening potentially interesting formations
(see above). Flowstone embedded in fine-grained
clastic sediments should also be viewed with caution
as it likely contains clay particles which compromise
the quality of the dating result.
Many dripstone caves contain abundant broken spe¬
leothems, locally reworked by a cave stream. Sampling
such fragments is clearly easier to justify than removing
in situ stalagmites and still provides valuable insights
into the history of a cave or the regional paleoclimate.
Speleothem fragments reworked by cave streams,
however, are sometimes problematic as the water can
alter the chemical composition of the samples due to
undersaturation with respect to calcite.
GROWTH DYNAMICS OF SPELEOTHEMS
U-Th dating offers the opportunity of using speleo¬
thems as archives of environmental change in the past.
To this end, the internal age structure of a given stalag¬
mite (or flowstone) needs to be known in as much
detail as possible. This typically requires a set of some
10—30 individual U-Th subsamples, depending on the
size and age range of the speleothem. The relationship
between distance (measured from the top or base of a
stalagmite) and the age provides fundamental insights
into the growth history (Fig. 4), including possible
growth interruptions (hiati). Various approaches exist
to mathematically fill the gaps between individual U-Th
data points and to assess the degree of uncertainty
ENCYCLOPEDIA OF CAVES
URANIUM SERIES DATING OF SPELEOTHEMS
843
Distance from top of stalagmite Distance from top of stalagmite
-►
Distance from top of stalagmite
-►
Distance from top of stalagmite
FIGURE 4 Schematic depth-age relationships commonly found in stalagmites. Filled circles and vertical lines represent individual U-Th
samples and their associated analytical uncertainties, respectively. (A) the gradual progression of individual U-Th ages suggests an overall
linear growth model (stippled line). Note, however, that given the relatively large uncertainties associated with some of the ages, more com¬
plex growth histories cannot be ruled out. (B) Stepwise growth model which demonstrates that the rate of vertical extension increased strongly
in the middle part of the speleothem. Such increases in growth rate commonly reflect a change to higher drip rate and hence a shift to more
humid (and possibly warmer) outside climate. (C) This speleothem shows a distinct growth interruption (hiatus) indicated by the jump in the
growth curve, that is, during this time interval the drip above the stalagmite presumably dried up. (D) The older part of this speleothem
shows a large scatter in the U-Th dates (including age inversions), strongly suggesting that this part has not behaved as a closed system.
In contrast, the ages in the younger segment are in correct stratigraphic order and hence provide a robust chronology.
resulting from modeling the growth history. The most
conservative way of linking up individual data points
along the extension axis of a stalagmite is a piecewise
linear interpolation. Alternatively, polynomial regres¬
sions or smooth-spline functions may be used.
Thanks to a large number of stalagmites analyzed
in the past two decades, we have a robust database at
our disposal which shows that stalagmites often show
a near-linear growth pattern, that is, the growth rate
(commonly expressed in pm /a) remained rather con¬
stant over millennia. These growth rates, however, may
vary over 2—3 orders of magnitude in stalagmites and
show a strong relationship with climate: rapidly grow¬
ing stalagmites (fractions of a mm/a) are often found in
temperate and warm regions with sufficient net infiltra¬
tion of meteoric precipitation, for example, regions in
Asia affected by summer monsoon rains, or in humid
circum-Mediterranean regions. At the lower end of the
spectrum are small stalagmites from cold caves located
at high latitudes or high elevation. In these locations,
growth rates do not exceed a few tens of pm/a. Very
slow growth also occurs in caves located in semi-arid
regions, where caves are dry most of the year. By far
the lowest rate was measured in some subaqueous
deposits which precipitated under extremely stablecon-
ditions of low calcite supersaturation. A thick subaque¬
ous calcite deposit present in a warm (34°C) aquifer
underlying the desert of southern Nevada formed at an
astonishingly slow average rate of 0.8 pm/a over the
past half million years.
U-Th dating has also provided insights into the
typical "lifetime" of speleothems. Again, there is a
large variability ranging from stalagmites in semi-arid
regions, where growth phases are confined to humid
climate intervals separated by hiati, to large or very
slowly forming formations where deposition occurred
ENCYCLOPEDIA OF CAVES
844
u
uninterruptedly over tens or even hundreds of thou¬
sands of years. It is worth emphasizing that these
findings are also relevant for hydrological studies and
show that individual drip sites in caves remain
stable over these very long periods.
U'PB DATING OF ANCIENT
SPELEOTHEMS
Even with today's most advanced instrumentation,
it is impossible to date samples beyond ca. 600—700 ka
using the U-Th "clock." Under very special geochemi¬
cal circumstances it may be possible to use a variant
of this method, the 238 U- 234 U decay scheme, to extend
the datable time range to ca. 1.5 million years. There
is also another method within the U-series family
which is capable of dating speleothems, in principal,
as old as the age of the Earth, based on the two radio¬
active decay series from 238 U and 235 U to stable 206 Pb
and 207 Pb, respectively. This is not a routine technique,
however, because it is very time-consuming and
expensive and requires samples of exceptional chemi¬
cal composition. In essence, speleothem samples must
be rich in U (ideally at least ca. 1 ppm) and contain
very little initial (so-called common) Pb. If initial Pb
was present at the time of speleothem deposition, sig¬
nificant corrections have to be applied, increasing the
age uncertainty, in particular, if the speleothem is less
than a few millions of years old. Moreover, elements
such as the noble gas radon, which is part of both
decay series, are volatile and therefore tend to disturb
closed-system conditions. Despite these difficulties,
U-Pb remains a very promising technique to explore,
for example. Tertiary-age cave deposits, provided that
clean, U-rich samples are available.
See Also the Following Article
Paleoclimate Records from Speleothems
Bibliography
Dorale, J. A., Edwards, R. L., Alexander, E. C., Shen, C. C., Richards,
D. A., & Cheng, H. (2007). Uranium-series dating of speleothems:
Current techniques, limits, and applications. In I. D. Sasowsky &
J. Mylroie (Eds.), Studies of cave sediments. Physical and chemical records
of paleoclimate, revised edition (pp. 177—197). Dordrecht: Springer.
Richards, D. A., & Dorale, J. A. (2003). Uranium-series chronology and
environmental applications of speleothems. In B. Bourdon, G. M.
Henderson, C. C. Lundstrom & S. P. Turner (Eds.), Uranium-series
geochemistry, Reviews in Mineralogy and Geochemistry (Vol. 52,
pp. 407—460). Washington, D.C.: Mineralogical Society of America.
Scholz, D., & Hoffmann, D. (20 08). 230 Th/U-dating of fossil corals
and speleothems. Eiszeitalter und Gegemvart, 57, 52—76.
Spotl. C., & Mattey, D. (2012). Scientific drilling of speleothems - a
technical note. International Journal of Speleology, 41, 29—34.
van Calsteren, P., & Thomas, L. (2006). Uranium-series dating
applications in natural environmental science. Earth Science
Reviews, 75, 155—175.
Woodhead, J., Hellstrom, J., Maas, R., Drysdale, R., Zanchetta, G.,
Devine, P., et al. (2006). U-Pb geochronology of speleothems by
MC-ICPMS. Quaternan/ Geochronology, 1, 208—221.
ENCYCLOPEDIA OF CAVES
VERTEBRATE VISITORS—BIRDS
AND MAMMALS
Nikola Tvrtkovic
Croatian Natural History Museum, Croatia
CAVE-DWELLING SPECIES
There are no troglobites, permanent inhabitants of caves,
in the group of warm-blooded animals. Most birds and
mammals documented in connection with caves are ran¬
dom guests, although some species are not accidental
visitors. They prefer to stay in the cave for a shorter or a
longer period of time and have adapted in some ways to
cave habitats, such as in the case of mammals that have
developed special pads for climbing, long vibrissae, par¬
tial heterothermy, and certain types of echolocation.
Like troglobites, these types of cave visitors are referred
to as cave-dwelling animals or cavernicoles (Vandel, 1965),
with different levels of modification for cave life. Most
of them originate from the climbing type of animal,
especially from rock-dwellers (petricola). According to
the ecological classification of cavernicola modified
from the scheme of Schiner—Racovitza, cavernicolous
birds and mammals are habitual trogloxenes after the
Hazelton and Glennie system (Ford and Cullingford,
1976), or troglophiles (eutroglophiles and only partly
subtroglophiles) after the Pavani system (Vandel,
1965)— cavernicola by choice, not by chance.
In the first group of cavernicolous birds and mam¬
mals are facidtative visitors: they only visit caves, but
they are also common in different surface habitats.
Facultative cave visitors by choice are the pack rat
(Neotoma spp.) in America, the edible dormouse (Glis
glis) in Europe, the Egyptian fruit bat ( Rousettus
aegyptiacus) in North Africa and western Asia, and
the bird origma ( Origma spp.) in Australia. Their nests
or roost sites have been found many hundreds of
meters from the entrances, but they are common in
other habitats too, such as in tree holes or shallow
depressions in cliffs.
The second group are regional cavernicoles —they are
permanent cave-dwellers only in some parts of their
range. The cave swallow ( Hirundo = Petrochelidon fulva)
is the first example: it breeds in caves only in southern
parts of its range, in the border area between the
United States and Mexico, and in Cuba. In this group
there are several owls, such as the little owl ( Athene
noctua) and large eagle owls like Bubo bubo in the karst
area of Europe, or Bubo africanus and Bubo ascalaphus in
Saudi Arabia. The monk seal ( Monachus monachus)
rears its young on beaches along the North African
coast, but most of their breeding sites along the north¬
ern Mediterranean coast were (before extinction) in the
sea caves. Two types of crows ( Pyrrhocorax gracidus
and P. pyrrhocorax) are high mountain birds of Eurasia,
nesting in crevices in cliffs. But in the Mediterranean
karst, they nest only in deep cold potholes (pits).
The Geoffroy bat ( Myotis emarginatus), primarily a
Mediterranean species, has large nursery colonies in
sea caves or warmer caves together with the greater
horseshoe bat ( Rhinolophus ferrumequinum). In Central
Europe both species are found in the nonkarst areas.
Their small roost colonies can be found in attics, and
they migrate in the fall to karst areas with their tradi¬
tional hibernacula.
The last group is regular cavernicoles. According to
Brosset (1998) and other sources, the birds among
these regular cave visitors include the guacharo or oil-
bird ( Steatornis caripensis) in Central and South
America, the Guinean cock of the rock ( Rupicola rupico-
la) in South America, the waterfall swift ( Hydrochous
gigas) in Southeast Asia, the swiftlets (. Aerodramus spp.
and Collocalia spp.) in Southeast Asia and the neigh¬
boring islands from the Seychelles to West Oceania,
pallid swifts ( Apus pallidus) on Mediterranean coasts,
and two rockfowls from central Africa ( Picathartes oreas
and Picathartes gymnocephalus). Regular cave visitors
among mammals include the Dinaric or Martino's vole
Encyclopedia of Caves.
845
© 2012 Elsevier Inc. All rights reserved.
846
v
FIGURE 1 The Martino's vole (Dinnromys bogdanovi), subadult
female. Photograph taken by the author.
(Dinaromys bogdanovi ; Fig. 1), restricted only to the west¬
ern Balkans in Europe, and the dawn bat ( Eonycteris
sipelea) from the group of fruit bats of Southeast Asia.
The largest group of regular cavernicola are micro-
chiropteran bats restricted to caves, like the gray bat
(Myotis grisescens ) in North America, the long-fingered
bat ( Myotis capaccinii) from Mediterranean karst habi¬
tats, and Schreiber's bat ( Miniopterus schreibersi) from
Europe, North Africa, Asia, and Australia.
GAPS IN KNOWLEDGE
Lists of cave-dwelling birds and mammals are
incomplete. Most of the data have been documented
accidentally and we have not yet had the ability to
compare the numbers of cave-dwelling species in dif¬
ferent karst regions of the world. Some species have
been documented with only one specimen, like the
shrew Soriculus salenskii in Sichuan (China) inside a
cave with water, or the fruit bat Aproteles bulmerae
from only one cave in Western Papua (Nowak, 1994).
The ecology and behavior of known cavernicolous
birds and mammals are also poorly known. Of more
than 20 species of the Asian bird genus Aerodrama,
until now echolocation is documented in only three
species (del Hoyo et ah, 1999). How successful is the
first flight of nestlings of the Alpine chough from a
70-m-deep nest site in the darkness of a pothole?
Many similar questions about cavernicolous birds and
mammals are without answers even today.
CAVES AS SHELTER FOR RESTING
AND MATING
Caves offer a wide spectrum of microhabitats (Ginet
and Decou, 1977), because they are always a suitable
shelter from rain, snow, sun, and wind and from most
predators. Isolation from surface habitats and from
most of the predators, low level of disturbance, dark¬
ness and a stable temperature, humidity, and airflow
regime are important characteristics for resting and
mating. Caves provide daytime shelter for nighttime
active species such as owls, hyenas, hyraxes, wild
boars, peccaries, porcupines, bats, and various species
of rock-dwelling rodents, or nighttime shelter
suitable for digesting food in the case of owls and bats.
Croatian speleologists observed the karst mouse
(Apodemus epimelas) in many Mediterranean caves and
pits, some specimens were seen at a depth of up to
70 m. In one of the deepest pits in Dinaric karst,
Slovakia pit, bats were observed in flight up to 900 m
deep in summer. Most bats mate in wintering or tran¬
sitional places in caves before and after hibernation.
Males of the lesser mouse-eared bat ( Myotis blythii) in
Europe and Asia position themselves on the ceiling
and walls of caves from spring to fall in isolated soli¬
tary places marked with urine, waiting for females.
Schreiber's bat forms small harems in different parts
of caves early in spring, as well as in August and
September.
CAVES AS DEN SITES
Most cave birds breed colonially, like the oilbird
(Steatornis caripensis), the swifts ( Aptis pallidas), the
swiftlets (Aerodramus spp. and Collocalia spp.), the
crows ( Pyrrhocorax gracnlus and P. pyrrhocorax), and
the rock doves ( Columba livia). Their nests have been
found many hundreds of meters from the cave
entrances and in pits up to 100 m below the entrance.
In limestone caves many traces were documented of
the edible dormouse (Glis glis) such as nests, feces,
footprints, and urine marks. It has been reported that
the nests of the dormouse have been found deep (sev¬
eral hundred meters) in the caves of Italy and
Slovenia.
Desert caves in West Asia are regular den sites for
striped hyenas ( Hyaena hyaena). Hyenas enter caves to
avoid temperature extremes, to eat large scavenged
prey undisturbed, to store large bones for later con¬
sumption, to give birth and to raise their offspring, and
to die. Jackals ( Canis aureus) in the Mediterranean part
of eastern Europe, western Asia, and northern Africa
use caves as den sites too. The famous Ghost Cat of the
Himalayas—snow leopard ( Pardus = Unciauncia) —can
rear cubs for extended periods of time in a cave den in
the mountains in winter. Its cave dens are lined with
the mother's fur.
Warm chambers in caves with temperatures greater
than 12—25°C and high humidity are the summer home
ENCYCLOPEDIA OF CAVES
VERTEBRATE VISITORS—BIRDS AND MAMMALS
847
for nursery colonies of a great number of bats. Relative
high temperatures are important for juveniles, because
they have no possibility to regulate their own tempera¬
ture. Females of some bat species like the long-fingered
bat (Myotis capaccinii) are extremely philopatric to caves
in which they had been born: they usually return to
their natal colony after hibernation (Papandatou et al.,
2009). Different bat species usually form mixed nursery
colonies in caves. The most common were nursery
roosts with up to several thousands of specimens. The
world's largest extant nursing bat colony is that of the
guano bat ( Tadarida brasiliensis) in Bracken Cave in
Texas with about 20 million individuals (Hutson et al.,
2001 ).
CAVES AS WINTERING SITES
In the temperate region or in the upper extremes of
the subtropical regions, caves are suitable wintering
places for many bats. They hibernate alone or in colo¬
nies, in crevices inside caves or on the ceiling of
chambers. Most bat species gather to hibernate even
in larger concentrations in a small number of sites.
From all of the population of gray bat ( Myotis grises-
cens) in the southeastern United States, 90% of bats
hibernate in only three caves. The parti-colored bat
(Vespertilio murinus) migrates more than 1700 km
each year from mountains and steppe regions in
northern Europe to wintering sites in the Alps
(Hutterer et al., 2005). Schreiber's bat forms thick clus¬
ters in several layers, with an average of 2000 speci¬
mens per square meter of roof surface. The lesser
mouse-eared bat prefers sites at only 2—3°C, and the
gray bat prefers sites with temperatures between
10°C and 15°C. The long-fingered bat forms wintering
clusters only above cave waters. In the Pleistocene,
cave bears ( Ursus spaeleus) wintered in caves all over
the Eurasia. Recent species of bears usually use small
caves for wintering, too.
CAVES AS FORAGING HABITATS
Some snakes and owls such as the barn owl ( Tyto
alba) catch bats at cave entrances. Mammals such as
the ringtail ( Bassariscus astustus), the red fox ( Vulpes
vulpes), the fennex ( V. zerda), the coyote ( Cams latrans),
the badger ( Meles meles), the stone marten ( Martes foi-
na), and the striped skunk ( Mephitis mephitis) visit
caves too, but mostly only for eating young bats or
birds that fall to the floor of the roosting site, or to eat
bats during their hibernation (Juberthie et at., 1998).
Throughout the year predators in tropical caves
include carnivores such as Herpestes, Viverra, and Felis,
and shrews such as Crocidura malayana and C. cinerea.
Desmodus rotundus and other vampire bats are a special
case of bat predation, preying on other bats in some
South American caves. In the Late Quaternary the
Eurasian spotted hyena (Crocnta crocuta spelaea) and
the steppe lion ( Panthera leo spelaea) hunted cave bears
especially during hibernation times in caves in low to
medium high mountainous elevations, and in high
alpine regions, leopard ( Panthera pardus) seems to have
used the ecological niche of the absent hyenas
(Diedrich, 2010).
The most dangerous cave predator of cave birds
and bats is man ( Homo sapiens). Native hunters from
Papua New Guinea collect fruit bats in caves for food,
and modern hunters in Asia collect nests made of
swiftlets' saliva to make a special soup. Local people
on the Mediterranean and along other coasts harvest
hatchlings of sea birds from small sea caves: one man
from Socotra Island confessed that they could take 30
nestlings of rare Jounin's petrel ( Bulweria faltax) from a
single cave in a night. Local people of islands of the
Vis archipelago in the Adriatic Sea traditionally har¬
vested eggs and nestlings of Cory's shearwater
(Calonectris diomedea), a social sea bird which breeds in
sea caves and large coastal crevices.
IMPACT OF BIRDS AND MAMMALS
ON CAVE ECOLOGY
Birds and mammals have an important impact on
cave ecology. Juberthie et al. (1998) noted that the
feces of porcupines in two caves in Africa are an
important food for endemic stygobitic isopods. The
input of the droppings of birds and bats is sometimes
extremely high. Invertebrate guano fauna is formed in
guano piles, an important food resource for cave¬
dwelling invertebrates. Rodents such as the Martino's
vole take green plants and mice and dormice take
seeds inside caves for storage. Rodents such as edible
dormice and birds build subterranean nests of plant
particles. Hyenas accumulate remains of its prey and
bones for later consumption in caves. The remains of
nests and food storage, and carcasses of cave visitors
are an energy source for troglobites, although in tropi¬
cal caves an abundance of food input attracts more
surface competitors (Mohr and Poulson, 1969). In
cases of colonial breeding in relatively small cave
chambers, bats have changed habitat factors such as
temperature and humidity. In extreme situations, mil¬
lions of bat specimens decomposing guano produce a
high concentration of ammonia and/or carbon diox¬
ide, which is dangerous for man (Hill and Smith,
1984).
ENCYCLOPEDIA OF CAVES
848
v
CAVES AS IMPORTANT HABITAT FOR
SOME RARE SPECIES
The rare sea bird Jounin's petrel ( Bulweria fallax)
known from the NW Indian Ocean has breeding sites
only in the sea-facing caves of and fissures on Socotra
Island. The Saint Helena petrel (B. bifax) was endemic to
Saint Helena Island, but went extinct in historical times.
In western Balkans the endemic Martino's vole
(Dimromys bogdanovi) is known as a regular cave visitor
in the rough karst landscape. It is a habitat specialist
that generally depends on underground shelters in
limestone karst (Krystufek and Buan, 2008). It is docu¬
mented by several speleologists to be seen deep in caves
and pits. It is also an altitude generalist, known from
sea level to mountain peaks at 2200 m above, in ecolog¬
ically very different plant belts. Paleontological evidence
suggests that the range of the genus Dimromys was his¬
torically small and restricted mainly to the Dinaric karst
in the western Balkans. This small-range species is pre¬
sumably the only surviving member of the Late Tertiary
genus Pliojnys. Caves as shelter were crucial in survival
for this paleoendemic taxon. Mitochondrial DNA
sequencing has linked Dimromys to Prometheomys, a
monospecific subterranean genus endemic to the
Caucasus in Asia Minor with divergence estimated to
have occurred 6.2 Ma. The available evidence suggests
that recent populations are small and frequently isolated
due to high habitat specificity. Last results of molecular
investigation with nuclear markers suggest that
Martino's vole consists of two from phylogeographical
lineages, dating from about 1 Ma, which may represent
distinct cryptic species, and have different threat statuses.
MAN AND CAVERNICOLOUS BIRDS
AND MAMMALS
In the past, ancient man and his relatives in temper¬
ate and northern areas, especially in glacial stages
from Middle to Late Pleistocene, had to compete hard
for caves as dens with large carnivores like the cave
bear and the cave or Eurasian spotted hyena ( Crocuta
crocuta spelaea). Scientists suppose that ancient hyenas
were probably the most dangerous animal for men in
the Late Pleistocene period. It has been estimated that
they weighed up to 102 kg and have hunted in packs
as large as 40 to 50 animals. Some have hypothesized
that the domestication of dogs might finally have
aided humans to successfully defend against maraud¬
ing hyena packs. Cave hyena populations began to
shrink after roughly 20,000 years ago, completely dis¬
appearing from Western Europe between 14,000 and
11,000 years ago with decline of grasslands (Stiner,
2004). Competition and predation of hyenas in Siberia
were probably significant factors in delaying human
colonization of Alaska with crossing the Bering Strait
earlier. The oldest Alaskan human remains coincide
with the same time when cave hyenas became extinct.
In time. Homo sapiens created artificial caves: houses,
mines, and other human-made shelters, which are
suitable habitats for most of cave birds and mammals.
Tin roofs provide a microclimate that is very similar to
that of warm caves, which proves to be a good place
for nursery colonies of bats or the nesting places of
rock doves ( Columba livia). Rock doves nest in lime¬
stone pits in Mediterranean karst and in lava tubes in
Saudi Arabia, for example, in the pit of Dharb al
Najem. Recently, the most numerous populations of
this bird are in cities. The most of European bats
change their former roost sites in caves to lofts of
houses. There has, therefore, been a co-evolution of
man and formerly cave-dwelling animals.
Some cave-dwelling birds and mammals are of
direct economic importance (Nowak, 1994). Local peo¬
ple in Southeast Asia collect bats for food. Swiftlet
nests, mostly of the black-nest swiftlets ( Aerodrama
maximus ) and the edible-nest swiftlets ( A. fucifugns),
are high-priced export goods to some Southeast Asia
countries. Cave guano harvesting was important in the
past for the gunpowder industry (Hutson et al., 2001),
while today it is an important source of nitrates for the
fertilizer industry. Cave-dwelling birds and bats are
consumers of pest insects or insect vectors of different
diseases. Fruit bats are pests in orchards, but the role
of cave bats as pollinators of some fruits and their role
in tree seed dispersion in subtropical areas are very
important.
On the other hand, in some cases, cave bats are
problematic for public health as potential vectors of
rabies. In northern areas we know of some lethal cases
in humans: the death of one bat worker bitten by a
wintering Daubeton's bat ( Myotis danbentonii) in north¬
ern England in 2001. UNEP/EUROBATS recommends
gloves for bat handling. In South America, vampire
bats ( Desmodus rotundus) are vectors of some diseases.
Jackals ( Canis aureus) are accused of being the natural
reservoir hosts of protozoan Leishmania donovani, the
cause of Kala azar disease in areas from the eastern
Mediterranean to west India. Sand flies ( Phlebotomus
spp.) live in jackals' dens in caves and are vectors of
Leishmania to humans. Dry caves inhabited by birds
or bats are dangerous places for humans because
they can act as reservoirs of the fungus Histoplasma
capsulatum. The fungus grows in soil rich with bat or
bird droppings (guano) and men were inoculated by
inhalation of fungal spores from disturbed guano. The
acute phase of this "cave disease" is characterized by
nonspecific respiratory symptoms like cough or flu.
ENCYCLOPEDIA OF CAVES
VICARIANCE AND DISPERSALIST BIOGEOGRAPHY
849
which cause damage to the retinas of the eyes or affect
diverse organ systems, occasionally with fatal results.
The tendency of some species of birds and bats to
aggregate in large numbers only in caves makes them
very vulnerable. The best example is the white-nose
syndrome. This cutaneous fungal infection caused by
the fungus Geomyces destructans was responsible for
massive mortality of hibernating bats in the northeast¬
ern United States (Biehert et ah, 2009). Since the winter
of 2006/2007, over a million bats died: bat declines
exceeding 75% have been observed at surveyed hiber-
nacula. In 2009 and 2010 G. destructans was for the first
time identified in European caves too, but here bats
remain healthy.
There is a long list of threats for cave-dwelling
birds and mammals today. In combination with
threats outside the caves, formerly high regional
populations of cave bats are in drastic decline. In
Eagle Creek Cave in Arizona, the population has
declined from more than 25 million guano bats in
1963 to 30,000 in 1969. Some formerly large colonies of
this species in Mexico are now extinct. On the IUCN
Red Data List of Threatened Animals, there are 11
species of cave birds (Stattersfield and Capper, 2000),
snow leopard, Martino's vole, and many more cave
bats, which make up the vast majority of the 370
threatened species of bats (Hutson et ah, 2001).
See Also the Following Articles
Bats
Paleontology of Caves
Bibliography
Biehert, D. S., Hicks, A. C., Behr, M., Meleyer, C. U., Berlowski-Zier,
B. M., Buckles, E. V., et al. (2009). Bat white-nose syndrome: An
emerging fungal pathogen? Science, 323(5911), 227.
Brosset, A. (1998). Caves. In C. Juberthie & V. Decu (Eds.),
Encyclopaedia biospeleologica (Vol. II, pp. 1249—1256). Moulis,
France: Societe de Biospeologie.
del Hoyo, J. E., Elliot, A., & Sargatal, J. (1999). Handbook of the birds of
the world, Vol. 5, Barn-Owls to Hummingbirds. Barcelona: Lynx
Edicions.
Diedrich, C. G. (2010). Spotted hyena and steppe lion predation
behaviours on cave bears of Europe—Late quarternary cave bear
extinction as result of predator stress? Geophysical Research
Abstracts, 12, EGU2010-2111.
Ford, T. D., & Cullingford, C. H. D. (1976). The science of speleology.
London: Academic Press.
Ginet, R., & Decou, V. (1977). Initiation a la biologie et a I'ecologie sou-
terraines. Paris: Jean-Pierre Delarge.
Hill, J. E., & Smith, J. D. (1984). Bats, a natural history. London: British
Museum of Natural History.
Hutson, A. M., Mickleburgh, S. P., & Racey, P. A. (Eds.), (2001).
Microchiropteran bats: Global status survey and conservation
action plan. Gland, Switzerland: IUCN/SSC Chiroptera Specialist
Group.
Hutterer, R., Ivanova, T., Meyer-Cords, C., & Rodriques, L. (2005).
Bat migrations in Europe. Bonn: Naturschutz und Biologische
Vielfalt, Federal Agency for Nature Conservation.
Juberthie, C., Decu, V., & Radulescu, C. (1998). Mammalia
(Marsupialia, Insectivora, Artiodactyla, Rodentia et Fissipedia).
In C. Juberthie & V. Decu (Eds.), Encyclopaedia biospeleologica (Vol.
II, pp. 1257—1261). Moulis, France: Societe de Biospeologie.
Krystufek, B., & Buzan, E. V. (2008). Rarity and decline in palaeoen-
demic Martino's vole Dinaromys bogdanovi. Mammal Revue, 38,
267-284.
Mohr, C. E., & Poulson, T. L. (1969). The life of the cave. New York:
McGraw-Hill.
Nowak, R. (1994). Walker's bats of the world. Baltimore, MD: The Johns
Hopkins University Press.
Papandatou, E., Butlin, R., Pradel, R., & Altringham, J. D. (2009). Sex-
specific roost and population dynamics of the vulnerable long¬
fingered bat, Myotis capaccinii. Biological Conservation, 112,
280-289.
Stattersfield, A. J., & Capper, D. R. (2000). Threatened birds of the
world. Barcelona and Cambridge: Lynx Edicions and Bird Life
International.
Stiner, M. C. (2004). Comparative ecology and taphonomy of spotted
hyenas, humans, and wolves in Pleistocene Italy. Revue de
Paleobiologie, 23, 771—785.
Vandel, A. (1965). Biospeleology, the biology of cavernicolous animals.
Oxford: Pergamon Press.
VICARIANCE AND DISPERSALIST
BIOGEOGRAPHY
John R. Holsinger
Old Dominion University, Norfolk, Virginia, U.S.A.
OVERVIEW OF BIOGEOGRAPHY
Biogeography is a subdiscipline of biological systemat-
ics briefly defined as the science that documents and
attempts to explain the geographic distribution of
organisms. Two conceptually different approaches
have developed over the years. They include historical
or large-scale biogeography and ecological or small-
scale biogeography. Historical biogeography is the oldest
and most commonly practiced approach and explains
distributional patterns of organisms on the basis of
their dispersal and/or vicariance over evolutionary
time. It is essentially large scale, inasmuch as it exam¬
ines the distribution of biotas from a regional, conti¬
nental, or global perspective. In comparison, ecological
biogeography, sometimes called deterministic biogeogra¬
phy, explains distribution patterns on the basis of rela¬
tively short-term, ecological effects. It is generally
regarded as small-scale biogeography because it focuses
on the effects of environmental differences that have
developed over ecological time. An example of the lat¬
ter would be the Case and Cody (1987) application of
island biogeography theory to explain distributional
patterns on islands in the Sea of Cortez in Mexico.
Encyclopedia of Caves.
© 2012 Elsevier Inc. All rights reserved.
850
v
Historical biogeography had its beginnings in Europe
in the eighteenth century, with the early, largely botani¬
cal, works of Linnaeus, Buffon, Forster, Humboldt, and
de Candolle (Lomolino et ah, 2005). Modern historical
biogeography began in the following century and owes
many of its basic principles to Alfred Russel Wallace, a
colleague of Charles Darwin, who is often called the
"father of zoogeography."
As a science, historical biogeography is very different
now from what it was in the 1800s or, for that matter, in
the first six or seven decades of the twentieth century.
Significant changes came in the 1960s and 1970s with
wide acceptance of the theory of plate tectonics and con¬
tinental drift, coupled with the rapidly developing popu¬
larity and utilization of phylogenetic analysis or
cladistics. With subsequent application of these exciting
new theories and concepts to the interpretation of distri¬
bution patterns, historical biogeography has become
more focused and analytical. Some workers now refer to
it as cladistic biogeography (Humphries and Parenti, 1999).
Prior to the 1960s, biogeographic discussions consisted
largely of lengthy narratives and scenarios, sometimes
highly subjective, to explain distribution patterns in
terms of centers of origin and outward dispersal.
Dispersalist models identified centers of origin and then
speculated on the subsequent peripheral movement of
organisms across potential dispersal barriers, such as
over imaginary land bridges. In retrospect, many of these
barriers appear to have been insurmountable.
The application of cladistics to biogeography has
placed significant emphasis on construction of biologi¬
cal-area cladograms, combined with subsequent exam¬
ination of relationships between terminal taxa and
areas of endemism (Fig. 1). In these analyses, congru¬
ence between area cladograms of different monophy-
letic taxa suggests that these organisms share a
common evolutionary history. This procedure has led
ultimately to the acceptance of vicariance as an alterna¬
tive hypothesis to explain heretofore many otherwise
inexplicable distribution patterns. However, this is not
to say that dispersalist models have diminished in
importance. On the contrary, dispersal is still consid¬
ered very important in explaining distributions. But in
recent studies, centers of origin are rarely evoked, and
dispersal often provides an alternative hypothesis
rather than the sole explanation for a given pattern
of distribution (see Wiley, 1988). Another analytical
method called panbiogeography (Craw et ah, 1999),
which places greater emphasis on congruent distribu¬
tion tracks (z.e., generalized tracts) than area clado¬
grams, has also gained popularity as an analytical
methodology. When used in combination with phylo¬
genetics, panbiogeography has proven effective for
explaining distribution patterns of subterranean ani¬
mals (Christiansen and Culver, 1987).
© © © areas
ABC species
FIGURE 1 Example of biological or taxon area cladograms for a
hypothetical genus composed of species A, B, and C that occur in
discontinuous areas 1, 2, and 3, respectively. These cladograms are
produced by combining the name of the species with the area(s) it
inhabits. In this simple example, each species is endemic (restricted)
to a different geographic area, suggesting a strong relationship
between species distributions and geographic isolation. Open circles
are internal nodes, which indicate splits in lineages leading to termi¬
nal species. Internal nodes can also correspond to vicariant events.
BIOGEOGRAPHY OF CAVE ANIMALS
To date, between 10,000 and 15,000 species of animals
have been described from caves and similar subterra¬
nean habitats throughout the world. However, an exact
count is not available and would be meaningless never¬
theless because new species from subterranean environ¬
ments are being discovered and described on a regular
basis and the number continues to increase. A majority
of these species typically lack eyes and pigmentation
and are said to be troglobites (troglomorphic) if living in
terrestrial environments or stygobites (stygomorphic) if
living in aquatic ones. In either event these organisms
are apparently well adapted for life underground. From
a biogeographic perspective, subterranean environments
typically include cave passages in limestones and other
soluble carbonate rocks, lava tubes in basalt, and inter¬
stitial spaces in a variety of unconsolidated sediments.
However, virtually any habitable space beneath the sur¬
face of the Earth can be a subterranean environment.
Troglomorphic and stygomorphic animals make
excellent candidates for biogeographic studies and are
increasingly studied in this research for five reasons.
First, they are taxonomically diverse and widespread
geographically. Second, they are becoming well known
taxonomically, largely through the diligent collecting
efforts and taxonomic research of the past 50 years,
coupled in the past decade with phylogenetic analyses.
Third, dispersal abilities, although undoubtedly vari¬
able, are generally limited because these organisms are
adapted to the restricted confines of subterranean habi¬
tats. Fourth, ranges are commonly, but not invariably,
narrowly circumscribed, resulting in numerous locally
ENCYCLOPEDIA OF CAVES
VICARIANCE AND DISPERSALIST BIOGEOGRAPHY
851
endemic species. Fifth, many taxa apparently represent
old phylogenetic lineages that have persisted in subter¬
ranean refugia for long periods of time, and both
aquatic and terrestrial taxa are represented by distribu¬
tional and phylogenetic relicts in temperate and tropical
karst regions. What roles have vicariance and dispersal
played in shaping the distribution patterns of subterra¬
nean fauna? Is one of these processes more important
than the other in explaining distributions, or are they
approximately equal in explanatory value? Dispersal is
defined as the movement of an organism from one area
to another, which changes the natural distribution of
that organism, whereas vicariance is defined as the
occurrence of closely related taxa in disjunct areas,
which have been separated by the development of a nat¬
ural barrier (Humphries and Parenti, 1999). The forma¬
tion of a natural barrier that results in splitting a
formerly continuous range is a vicariant event. As
pointed out below, both dispersal and vicariance have
played significant roles in the development of cave fau¬
nal distribution patterns, and constructing both dispers-
alist and vicariance models is necessary for deeper
insight into the complexity of these patterns. Therefore,
because most present-day ranges probably reflect as¬
pects of both dispersal and vicariance, teasing out these
two components and delineating their contribution to
the formation of distribution patterns is one of the prin¬
cipal tasks of large-scale cave biogeography.
Dispersal and vicariance biogeography of terrestrial
troglobites and aquatic stygobites are considered sepa¬
rately below, primarily because of important differ¬
ences between the two with respect to origins,
habitats, and dispersal potentials. Although most of
the examples used here to illustrate various distribu¬
tion patterns are species from North America, many
comparable examples could be cited from Europe and
other parts of the world where subterranean fauna
have also been studied in detail.
TERRESTRIAL TROGLOBITES
Two levels of dispersal have played roles in shaping
the distribution patterns of troglobites. The first level
involves the actual invasion of caves from the surface
by putative preadapted founder populations. This may
occur under constraint when climatic changes precipi¬
tate invasion, which apparently occurred in temperate
karst during climatic vicissitudes of the Pleistocene. It
may also occur when founders invade newly develop¬
ing subterranean food niches, which is apparently com¬
mon and ongoing in tropical karst. These two models
for invasion and colonization are termed climatic relict
and adaptive shift, respectively. They were reviewed
recently in detail (Holsinger, 2000) and both are largely
dispersalist in character because they involve the active
movement of an organism from one environment to
another, effectively changing its natural distribution.
However, because climatic change envisioned under
the climatic relict model would likely preclude reinva¬
sion of surface habitats and isolate founders in caves, it
might arguably be regarded a vicariant event.
The second level of dispersal involves movement of
troglobitic organisms within the subterranean realm,
typically, but not invariably, between caves in karst
areas or lava flows. Both the extent of a given karst
area or lava flow and the dispersal vagility (mobility)
and niche breadth of a given species determine poten¬
tial limits of dispersal. For many troglobites this poten¬
tial is greatly limited and species ranges are often
determined by the extent of interconnected cave pas¬
sages and solution channels within a contiguous karst
area. In a biogeographic sense, contiguous karst areas
are analogous to islands, inasmuch as they are physi¬
cally isolated from each other by noncavernous rocks,
and often contain locally endemic cave species.
Dispersal of terrestrial troglobites includes move¬
ment through larger cave passages (>20 cm) and
numerous smaller mesocaverns (<20 cm) within karstic
limestone or other soluble carbonate rocks or basaltic
lava. It may also include movement through shallow
underground compartments (= milieu souterrain superfi-
ciel or MSS), which occur in some areas in loose rocks
at the base of scree slopes or in cracks and fissures in
mantle rock just beneath the lower layer of soil, both
within and outside karst areas (Juberthie and Decu,
1994). Moreover, it is likely that some of the more
widely distributed troglobites, such as species of liny-
phiid spiders and certain collembolan and dipluran
insects in eastern North America, commonly disperse
through cracks and crevices in epikarstic habitats above
caves and even occasionally through damp leaf litter or
deep soil outside karst areas. At the opposite end of
this spectrum are troglobitic pseudoscorpions of the
genus Kleptochthonius and pselaphid beetles, of which a
majority of species are highly localized endemics
restricted to single caves (Barr and Holsinger, 1985).
Presumably the low vagility and limited niche breadth
of these species prevent them from dispersing far from
very small areas or even a single cave habitat.
Trechine beetle species of the genus Pseudanophthal-
mus (family Carabidae) in eastern North America are a
classic example of well-studied troglobites, whose
ranges are often largely determined by the dispersal
limits of contiguous karst (Barr and Holsinger, 1985).
Many species appear to be limited in their distribution
to areas that typically encompass only interconnected
caves in a contiguous karst area. The effect on dis¬
persal by the extent of contiguous karst is nicely
illustrated by the distribution of species of
ENCYCLOPEDIA OF CAVES
852
v
Pseudanophthalmus in the Appalachian Valley and Ridge
physiographic province, where ranges vary in extent
from a single cave to many caves that occur over a lin¬
ear distance of approximately 95 km. In the Shenan¬
doah Valley of northwestern Virginia, where belts of
cavernous limestones are strongly folded.
FIGURE 2 Distribution of four troglobitic beetle species (numbers
1—4) of the hubbardi group of Pseudanophthalmus in caves of the
northern Shenandoah Valley of Virginia. Cave localities are indicated
by solid circles. Each species inhabits a cave or small cluster of caves
in a physically isolated limestone belt (contiguous karst).
discontinuous, and separated from each other by insol¬
uble sandstones and shales, four closely similar species
of the monophyletic hubbardi group occur in caves
within relatively short distances of each other (Fig. 2).
Extensive searching in caves of this region over the
years suggests that the range of each species is
restricted to an isolated belt of cavernous limestone,
which for all intents and purposes constitutes a small
contiguous karst area. Further evidence of the extreme
isolation of these species is that three of the four species
are recorded from single caves. Elsewhere in the Appa¬
lachian Valley and Ridge, belts of cavernous limestone
are more extensive and continuous over large areas,
and several species of Pseudanophthalmus have much
larger ranges that cover linear distances of 50—95 km
and include many caves (Fig. 3). In the Mississippian
plateaus of Indiana, Kentucky, and Tennessee, where
limestone outcrops are typically broader and less dis¬
turbed by folding than in the Appalachian Valley and
Ridge to the east, ranges of trechine beetles are often
more extensive (Barr and Holsinger, 1985) and provide
further evidence that ranges of these troglobites are
generally defined by the limits of contiguous karst.
The effect of vicariance on the distribution patterns of
troglobites is more difficult to document than that of dis¬
persal, primarily because barriers to subterranean dis¬
persal develop gradually over extended periods of time,
making it difficult to assess their impact. Flowever, vicari-
ant events are probably relatively common in the subter¬
ranean realm and thus are significant factors in isolating
cave populations and promoting speciation. Physical bar¬
riers to dispersal may develop externally from the surface
FIGURE 3 Distributions of the troglobitic beetles Pseudanophthalamus hoffmani (1) and P. delicatus (2) in southwestern Virginia karst areas.
Cave locations are indicated by solid circles. Both species occur in continuous belts of cavernous middle Ordovician limestone. The former is
recorded from caves developed in a long, relatively narrow belt of moderately dipping limestone along the northwestern flank of Big Walker
Mountain. The latter is recorded from caves that are developed in a relatively broad exposure of low-dipping limestone, which extends
throughout much of the Powell Valley in central Lee County, Virginia.
ENCYCLOPEDIA OF CAVES
VICARIANCE AND DISPERSALIST BIOGEOGRAPHY
853
though sinkhole collapse into underlying cave passages
or by deep erosional gorges that eventually completely
bisect subterranean channels. Physical barriers may also
develop internally and destroy or close off subterranean
dispersal corridors by filling them with silt, flowstone
deposits, or rubble from a ceiling collapse. Moreover,
physical barriers may develop vertically as well as lat¬
erally, inasmuch as new generations of caves are actively
forming beneath old ones. Thus, it is assumed that ani¬
mals will move downward to colonize newly developing
living space and available niches, and that the progressive
isolation of populations in newly developing, lower-level
passages may result from this activity.
AQUATIC STYGOBITES (STYGOFAUNA)
The subterranean aquatic environment in general is
more diverse and interconnected than its terrestrial
counterpart, and the extent to which groundwater is
continuous over wide areas plays a significant role in
the dispersal of stygobites. It is generally accepted that
the pervasive nature of subterranean groundwaters
provides stygobites, on average, greater dispersal
potential and therefore wider ranges than their terres¬
trial troglobite counterparts. Available groundwater
habitats occur both within and outside karst regions,
and many areas far beyond the boundaries of karst
terranes are inhabited by an interesting diversity of
stygobites (Ward et al., 2000). However, it is not
uncommon to find the same species, or closely related
species, simultaneously living in caves of a karst area
and adjacent subterranean aquifers in a noncavernous
area. As pointed out below, dispersal between karst
areas is often facilitated by movement through the
interstitial spaces of shallow groundwater aquifers that
are not directly associated with the caves in karst.
Subterranean groundwater habitats in karst caves
typically include streams, drip and seep-fed pools, and
phreatic lakes. They are often more accessible and eas¬
ier to observe than those outside karst areas. In addi¬
tion to typical karst caves with freshwater habitats,
anchihaline caves are a special cave subset that
develop in limestones or basalt in coastal areas near
the sea and are connected to nearby marine waters by
subterranean channels (see Iliffe, 2000). As one might
expect, the water in anchihaline caves varies from
nearly fresh to fully marine and, depending on depth,
may be horizontally partitioned by a halocline. With
some notable exceptions, anchihaline caves are inhab¬
ited by stygofaunas having morphological affinities
with marine faunas living in the nearby sea.
Outside karst areas, the more common subterranean
groundwater habitats include the saturated interstitial
spaces of gravels and coarse sands situated beneath
stream beds (hyporheic) or in coral rubble, coarse sand,
and so on, in littoral and sublittoral zones (macropor-
ous sediments), unconsolidated sediments beneath the
water table ( nappes phreatiques), and the outflow of small
springs and seeps (through fine sediments or leaf litter)
from perched water tables (hypotelminorheic). The
movement of stygobites through the saturated media of
unconsolidated sediments is often termed interstitial dis¬
persal and is apparently the principal reason why many
species that live in the shallow groundwater aquifers of
unconsolidated sediments are typically widely distrib¬
uted. For many species, interstitial dispersal appears to
be a very successful mechanism for range expansion,
and probably accounts for the relatively extensive
ranges of some of the common and widespread stygobi-
tic amphipods in the genus Stygobromns that live in
shallow groundwater aquifers along the outer margin
of the Piedmont and on the Coastal Plain of eastern
North America. These species, along with those in other
crustacean groups (e.g., bathynellids, copepods, iso¬
pods), are excellent examples of stygobites that appar¬
ently rely exclusively on interstitial dispersal to
maintain their distributions.
The distribution patterns of stygofaunas suggest
that dispersal is often not limited to a single type of
groundwater habitat but more frequently involves two
or more such habitats. For example, an important
aspect of cave drip pools is their close relationship
with epikarstic habitats that occur in open spaces and
bedrock fractures between cave passages and the
Farth's surface. Aquatic epikarstic habitats are in the
form of perched aquifers or perched water tables situ¬
ated well above the true groundwater table. A number
of studies have demonstrated passive dispersal of
small (<10 mm in length) stygobitic species (e.g., pla-
narians, amphipods, and isopods) from epikarstic
aquifers via ceiling drips into shallow cave pools on
the floor below. Passive dispersal may also occur
when the water in ceiling drips or wall seeps passes
into a cave stream. There is little doubt that epikarstic
aquifers facilitate dispersal of small stygobitic crusta¬
ceans both within and between caves.
A second example of dispersal involving more than
one cave habitat is illustrated by species that move
actively or passively between pools and small streams.
Studies on Appalachian cave amphipods in the genus
Stygobromus indicate that, in general, the larger, more
widely distributed species are recorded from both pools
and streams, whereas in contrast, smaller species with
greatly delimited ranges are generally found only in
tiny drip pools. Finally, it should be noted that some
species living in cave streams may be passively washed
downstream and into springs at the surface, where a
few can apparently survive for long periods of time
with adequate cover. It is conceivable that the outwash
ENCYCLOPEDIA OF CAVES
854
v
from springs might occasionally facilitate limited pas¬
sive dispersal through surface water between caves.
Larger stygobites, such as crayfish, shrimps, sala¬
manders, and fishes, and amphipods and isopods
exceeding 20 mm in length, are rarely if ever encoun¬
tered in cave drip pools but are commonly found in
cave streams or lakes. In the relatively shallow, wet
caves of the Florida lime sink region, decapods (pri¬
marily crayfish) also commonly inhabit submerged
cave passages. Because of the larger size of cave cray¬
fish, which excludes them from interstitial dispersal,
their distribution patterns are usually defined by the
limits of contiguous karst.
Dispersal over relatively great distances through
deep phreatic water in karst areas is another dispersal-
ist mechanism available to some stygobites, although it
is probably utilized primarily by taxa that inhabit deep
groundwater aquifers. Recent investigations suggest
that deep phreatic water habitats are more common
and interconnected than previously believed, but they
remain relatively poorly known because of limited
accessibility to cave biologists. Most observations have
been made where lower level cave passages allow
access to bodies of phreatic water or through water
wells in limestone bedrock that access deep groundwa¬
ter aquifers. Recent observations on amblyopsid cave-
fishes in the Ozarks and cirolanid isopod crustaceans
in the Appalachians provide strong evidence that deep
phreatic aquifers can serve as important avenues for
dispersal. According to a study in southern Missouri
by Noltie and Wicks (2001), the principal habitats
of the cavefishes Amblyopsis rosae in the Springfield
Plateau and Typhlichthys subterraneus in the Salem
Plateau are numerous small channels, solution tubes,
and vugs beneath the groundwater table in these karst
regions. Interestingly, because a majority of these
organisms live in areas that are inaccessible to cave
biologists, their population size has been underesti¬
mated. Lateral movements in the phreatic zone beneath
the water table facilitate dispersal of the fishes over a
relatively wide range. However, further observations
on A. rosae indicate the existence of four semi-isolated
genetic groupings within its overall range, suggesting a
limited degree of physical impediment to dispersal and
gene flow. Dispersal through deep phreatic water is
probably also responsible for the extensive distribution
of Typhlichthys subterraneus east of the Mississippi River
in caves of the Interior Low Plateaus in southern
Kentucky, central Tennessee, and northern Alabama.
However, there is some evidence in this region for
differentiation into several local "biological species"
separated by extrinsic dispersal barriers (Barr and
Holsinger, 1985).
Biogeographic studies by Holsinger et al. (1994) and
Hutchins et al. (2010) on the stygobitic cirolanid isopod
crustacean Antrolana lira, currently recorded from phre¬
atic water habitats in 17 localities (12 caves and 5 water
wells) in the Shenandoah Valley karst region of north¬
western Virginia and eastern West Virginia (Fig. 4),
suggest that this species, which may reach 21 mm in
FIGURE 4 Distribution of the stygobitic cirolanid isopod
Antrolana lira in the Shenandoah Valley of the Virginias. Solid
circles represent collecting sites (mostly caves except for five
water wells and one deep pit in the northern part of the
range). Note that the range of this species extends beneath a
major surface drainage divide between Staunton and
Lexington, Virginia. The question mark indicates a sight
record for specimens pumped from a water well.
ENCYCLOPEDIA OF CAVES
VICARIANCE AND DISPERSALIST BIOGEOGRAPHY
855
length, has utilized deep phreatic water for dispersal.
But unlike the near flat-lying strata in the Missouri pla¬
teaus, the limestone layers in this region are steeply
folded and separated from each other by nonsoluble
layers of sandstone and shale. However, fracturing of
the sandstones and shales makes it likely that the lime¬
stone beds are not fully isolated from each other at vari¬
able depths below the water table and that minimal
lateral movement of water and isopods has occurred
across the insoluble beds. Therefore, tenuous gene flow
has conceivably occurred throughout the range of this
species in the past and probably accounts for the mor¬
phologically identical populations. However, a recent
genetic study by Hutchins et al. (2010) yielded 14 haplo-
types, and a phylogenetic analysis of the haplotypes
revealed the presence of three distinct clades. The
clades, in turn, correspond to three geographic groups
throughout the presently known range of the species.
Dispersal through deep phreatic water also appears
to be utilized by some of the species that inhabit the
Edwards Aquifer in south-central Texas. This deep,
extensive aquifer is associated with the Balcones Fault
Zone and forms a biological mixing zone where karst
groundwater of the Edwards Plateau comes into con¬
tact with interstitial groundwater of the Coastal Plain.
The intense faulting and fracturing of the bedrock in
this region has also resulted in compartmentalization
and development of phreatic subaquifers within the
main aquifer. Thus, although dispersal obviously
occurs freely in some parts of the aquifer, barriers pre¬
vent it in other parts, and it is probable that a combi¬
nation of these two effects has contributed significantly
to the evolution of the highly diverse stygofauna of the
region. More than 31 stygobitic species have been sam¬
pled from the artesian well in San Marcos, where spe¬
cimens are coming from a depth of approximately 60
m. Sampling has also been carried out from very deep
artesian wells located approximately 90 km southwest
of San Marcos near San Antonio. Deep springs and
caves have also been carefully investigated in this
region. That part of the aquifer extending from San
Marcos southwest to San Antonio appears to be a via¬
ble dispersal corridor for a number of stygobites. Four
of the 10 stygobitic amphipod species recorded from
the artesian well in San Marcos have also been flushed
out of two of the artesian wells near San Antonio, and
a fifth species has been found in phreatic waters of
two nearby caves and possibly additional caves west
of San Marcos. The cirolanid isopod Cirolanides texensis
is also recorded from the San Marcos artesian well.
Like the amphipods, it has been sampled from deep
artesian wells near San Antonio as well as from caves
to the north and west on the Edwards Plateau.
As mentioned above, dispersal between karst areas
or even between caves is often facilitated by movement
through shallow groundwater aquifers that may or may
not be developed in limestone terranes. Thus, dispersal
through these aquifers, which can occur in a variety of
terranes, appears to be a very common means of range
expansion for relatively widely distributed stygobites.
Studies on stenasellid isopod crustaceans in ground-
waters on the Arabian Peninsula by Magniez and
Stock (1999) illustrate how interstitial groundwater in
the gravel sediments of a hyporheic zone and the
"free" water in nearby limestone cavities can be linked
through an ecotone, which is cohabitated by represen¬
tatives of species from the different habitats. Many
species of isopods and amphipods occur in caves and
to a lesser extent in springs and seeps at the surface
and they apparently utilize interstitial dispersal to
broaden their ranges. The amphipod crustaceans
Crangonyx antennatus and C. packardi and the isopod
Caecidotea pricei are three good examples of North
America stygobitic crustaceans that frequently inhabit
drip pools and small streams in caves but utilize
dispersal through epikarstic aquifers and hypotelmi-
norheic habitats (seeps) outside karst terranes to main¬
tain extensive ranges. The geographic distribution of
C. antennatus, which extends for approximately 700 km
through a number of different karst areas in parts of
several states in the Appalachians, is shown on the
map in Figure 5. The range of Crangonyx packardi
extends from southern Indiana and central Kentucky
west to eastern Kansas over a linear distance of
approximately 1000 km, whereas that of Caecidoeta
pricei extends from Rockbridge County, Virginia, north¬
east to Montgomery County, Pennsylvania, covering a
linear distance of approximately 450 km.
Most of the same vicariant events that form physical
barriers to the dispersal of troglobites mentioned
above apply equally to stygobites. For example,
sinkhole collapse into underlying cave passages and
bisection of subterranean channels by deep erosional
surface gorges would both serve to effectively block
potential stygofaunal dispersal corridors. Moreover,
siltation, flowstone deposition, and ceiling collapse can
also destroy or close off aquatic dispersal corridors.
However, three vicariant phenomena are unique to
the geographic distribution of stygofaunas: stranding,
stream capture, and spring failure. In a cladistic biogeo¬
graphic analysis, any one of these events might coincide
with the nodes on the area cladogram of the taxon
being investigated (Fig. 1). The nodes, in turn, reflect
splits in the lineages that gave rise to terminal taxa. The
possibilities for applying this kind of analysis to the
interpretation of biogeographic patterns of various
groups of stygobites is exciting and will increase as
more phytogenies of these organisms become available.
Stranding involves the isolation of marine or brack¬
ish water organisms in gradually freshening
ENCYCLOPEDIA OF CAVES
856
v
FIGURE 5 Distribution of the stygobitic amphipod crustacean Cmngonyx antennatus in the southern Appalachians. The solid circles repre¬
sent one to ten closely proximate localities. Although most of these localities are caves, the species has also been observed to enter cave pools
through ceiling drips from epikarstic zones and occasionally has been found in seeps and small springs at the surface.
subterranean groundwaters following marine regres¬
sions (see Parenti and Ebach, 2009). Founding popula¬
tions adapt to freshwater conditions as the salinity
gradually decreases in concert with regression of
marine waters. This process is gradual, but it is vicari-
ant because ultimately the stranded populations
become isolated from their marine ancestors when eco¬
logical conditions are changed and genetic continuity
is eliminated. The evidence for stranding comes from
the large number of freshwater stygobitic crustaceans
that inhabit subterranean groundwaters in parts of the
world that once bordered on or were covered by the
putative warm, shallow waters of the Tethys Sea. Most
of these taxa have obvious taxonomic affinities with
modern marine species and are apparently relicts of
former marine embayments and island emergence that
occurred in different parts of the Tethyan realm at dif¬
ferent times between the Cretaceous and Middle to
Late Tertiary. Marine relict stygobites are especially
common in Texas, Mexico, and the West Indies in
southern North America, and the Mediterranean
region of southern Europe and northern Africa. They
are also recorded from Eastern Africa, the Arabian
Peninsula, and parts of the western Pacific region
including Australia.
Stream capture involves the piracy or capture of
one stream by another, and in karst areas it includes
diversion of surface streams through sinkholes into
subterranean channels. Initially, preadapted freshwa¬
ter organisms enter subterranean waters passively
when their surface habitat stream is diverted into a
subterranean channel. Founder populations become
permanently established in caves only after they are
isolated in underground streams and contact is cut off
with surface ancestors. The capture of surface streams
by sinkholes is a relatively common phenomenon in
karst areas and is vicariant when remnant surface
streams recede up-gorge or dry up completely. In
accordance with this model, ancestors of cave-adapted
populations of the widely distributed characid fish
Astyanax fasciatus invaded caves in east-central Mexico
when surface streams were captured by sinkholes in
karst areas and were unable to reestablish contact with
epigean forms. Similarly, stygobitic crayfish in north-
central Florida are apparently derived from surface
ancestors that gained entry into subterranean waters
via stream capture in karst areas, presumably becom¬
ing isolated underground by the same processes.
Spring failure involves the retreat of preadapted sur¬
face populations into cave streams from surface
ENCYCLOPEDIA OF CAVES
VICARIANCE AND DISPERSALIST BIOGEOGRAPHY
857
springs during episodic droughts. Subsequent isolation
of founder populations in underground streams is
vicariant when the springs are eliminated by erosional
processes or when stream flow to the surface ceases
due to lowering of the groundwater table. Spring fail¬
ure is a fairly common phenomenon in karst areas
where the climate is becoming progressively more
arid. Apparently, it has been a significant factor in the
evolution of troglomorphic salamanders in central
Texas, which are derived from spring-dwelling ances¬
tors that migrated into cave streams to survive severe
surface droughts since the Pleistocene. Spring failure
may also be implicated in the evolution of cave-
adapted hydrobiid snails in parts of North America.
A final note regarding vicariance includes a brief
discussion of groundwater calcrete aquifers, which
constitute a unique subterranean environment recently
identified from the Western Australian arid zone by
Humphreys (2001). Calcretes are isolated subterranean
carbonate deposits with cavelike, freshwater habitats.
They have developed under arid conditions by evapo¬
ration and became isolated in the upper tributaries of
paleodrainage systems by the progressive upstream
movement of salinity, perhaps beginning in the Eocene.
Calcretes are numerous and inhabited by a stygofauna
consisting predominantly of locally endemic crusta¬
ceans, many of which are relict species. The intervals of
hypersalinity/calcrete discontinuities have acted as bar¬
riers to the dispersal of the stygofauna, and long-term
isolation of crustacean populations in upstream pockets
of freshwater has resulted in the evolution of a remark¬
ably diverse regional stygofauna. In a biogeographic
sense, calcretes are discontinuous groundwater habi¬
tats, which, like isolated belts of contiguous karst, are
analogous to islands.
CONCLUSION
The biogeography of organisms living in subterra¬
nean environments is primarily concerned with expla¬
nation and interpretation of the distribution patterns of
troglomorphic species. Both vicariance and dispersal
have played important roles in shaping the distribu¬
tion patterns of species restricted to a subterranean
environment, which include terrestrial troglobites and
aquatic stygobites. Whereas vicariance delimits ranges
through the development of dispersal barriers, dis¬
persal increases ranges by movement of organisms
into new areas. The effect of these processes has been
different for troglobites and stygobites, largely because
the two groups inhabit ecologically different subterra¬
nean habitats and have different dispersal potentials.
In general, subterranean aquatic environments are
more pervasive and diversified than their terrestrial
counterparts and, as a result, stygobites typically
inhabit a wider variety of habitats and have more
extensive ranges. Stygobites not only exploit karst
groundwaters but also inhabit the extensive interstitial
groundwater media that extends far beyond the
boundaries of karst terranes. Some karst species appar¬
ently also use deep phreatic aquifers for dispersal.
Troglobites may also exploit terrestrial habitats outside
karst, such as "shallow underground compartments,"
but these are generally limited in scope, and the ranges
of a majority of terrestrial species are restricted to
caves in karst or basaltic lava.
Although several forms of vicariance, such as the
destructive geomorphic processes that close subterra¬
nean passages, are similar for both troglobites and
stygobites and often establish the outer limits of distri¬
bution in karst areas, several important vicariant
events are unique to stygobites. They include strand¬
ing, stream capture, and spring failure, and each has
played a significant role in the distribution of stygo¬
bites. The newly discovered groundwater calcretes in
Western Australia are also vicariant in character, inas¬
much as they restrict the distribution of stygobitic
crustaceans in this arid region.
Acknowledgments
I am grateful to David C. Culver, Daniel W. Fong, David A.
Hubbard, Jr., Glenn Longley, and Wil Orndorff for providing useful
information during the preparation of this article; Donald K.
Emminger and Justin Shafer for preparation of the figures; and
Lynnette J.M. Ansell for editorial assistance.
See Also the Following Articles
Adaptive Shifts
Invasion, Active versus Passive
Bibliography
Barr, T. C., Jr., & Holsinger, J. R. (1985). Speciation in cave faunas.
Annual Reviews of Ecology Systematics, 16, 313—337.
Case, T. J., & Cody, M. L. (1987). Testing theories of island biogeog¬
raphy. American Scientist, 75(4), 402—411.
Christiansen, K. A., & Culver, D. C. (1987). Biogeography and the
distribution of cave collembola. Journal of Biogeography, 14,
459-477.
Craw, R. C., Grehan, J. R., & Heads, M. J. (1999). Panbiogeography:
Tracking the history of life. Oxford, U.K.: Oxford University Press.
Holsinger, J. R. (2000). Ecological derivation, colonization, and speci¬
ation. In H. Wilkins, D. C. Culver & W. F. Humphreys (Eds.),
Subterranean Ecosystems (pp. 399—415). Amsterdam: Elsevier.
Holsinger, J. R., Bowman, T. W., & Hubbard, D. A., Jr. (1994).
Biogeographic and ecological implications of newly discovered
populations of the stygobiont isopod crustacean Antrolana lira
Bowman (Cirolanidae). Journal of Natural History, 28, 1047—1058.
Humphreys, W. F. (2001). Groundwater calcrete aquifers in the
Australian arid zone: The context to an unfolding plethora of
ENCYCLOPEDIA OF CAVES
858
v
stygal biodiversity. Records of the Western Australia Museum,
Supplement No. 64, 63—83.
Humphries, C. J., & Parenti, L. R. (1999). Cladistic biogeography.
Oxford, U.K.: Oxford University Press.
Iliffe, T. M. (2000). Anchialine cave ecology. In H. Wilkins, D. C.
Culver & W. F. Humphreys (Eds.), Subterranean ecosystems
(pp. 59—76). Amsterdam: Elsevier.
Juberthie, C., & Decu, V. (1994). Structure et diversite du domaine
souterrain; particularites des habitats et adaptations des especes.
In C. Juberthie & V. Decu (Eds.), Encyclopaedia biospeologica (Vol. I,
pp. 5—22). Moulis, France: Societe de Biospeologie (in French).
Lomolino, M. V., Riddle, B. R., & Brown, J. H. (2005). Biogeography.
Sunderland, MA: Sinauer Associates.
Magniez, G., & Stock, J. H. (1999). Consequence of the discovery of
Stenasellus (Crustacea, Isopoda, Asellota) in the underground
waters of Oman (Arabian Peninsula). Contributions to Zoology, 68
(3), 173-179.
Noltie, D. B., & Wicks, C. M. (2001). How hydrology has shaped the
ecology of Missouri's Ozark cavefish, Amblyopsis rosae, and south¬
ern cavefish, Typhlichthys subterraneus: Insights on the sightless
from understanding the underground. In A. Romero (Ed.),
Environmental biology of fishes (Vol. 62, pp. 171 — 194). Dordrecht:
Kluwer Academic Publishers.
Parenti, L. R., & Ebach, M. C. (2009). Comparative biogeography:
Discovering and classifying biogeographical patterns of a dynamic
Earth. Berkeley, CA: University of California Press.
Ward, J. V., Malard, F., Stanford, J. A., & Gonser, T. (2000).
Interstitial aquatic fauna of shallow unconsolidated sediments,
particularly hyporheic biotopes. In H. Wilkins, D. C. Culver &
W. F. Humphreys (Eds.), Subterranean ecosystems (pp. 41—58).
Amsterdam: Elsevier.
Wiley, E. O. (1988). Vicariance biogeography. Annual Reviews of
Ecology Systematics, 19, 513—542.
VJETRENICA CAVE, BOSNIA
AND HERZEGOVINA
I vo Lucic
Speleological Association Vjetrenica—Popovo Polje,
Bosnia and Herzegovina
For centuries, Vjetrenica was perceived as a magic hole
in the ground, with a freezing breeze coming out even
during the hottest days of summer. The wind always
came with sounds that aptly illustrated its supernatural
force. This marked the colorful history of Vjetrenica
and its rich scientific history—it also gave it the name
( vjetar —wind in South Slavic; vjetrenica —a windy place).
It is the largest cave in Bosnia and Herzegovina, of total
length of 7014 m, with a rich and versatile karst inven¬
tory. It is one of the two leading world hotspots of
underground fauna, a very specific paleontological site, a
prominent shrine, and a protected site of natural heri¬
tage. It is located in the far south of the country, specifi¬
cally at the southwestern edge of Popovo Polje, in the
vicinity of the village of Zavala (at 42°50'45" S, 17°59'02'' I,
268 m above sea level). It is 10.1 km from the
nearest town on the Adriatic Coast, Slano, Republic of
Croatia.
THE KARST AREA OF VJETRENICA
It comprises the end section of the basin of the River
Trebisnjica, with total estimated surface of 2800 km 2 , and
an altitude range from the top of Mt. Lebrsnik (1985 m)
to sea level. The average annual precipitation in the basin
is 1780 mm, of which only 40% can be traced through
surface flows. Trebisnjica runs through several types of
karst poljes (Gatacko p., Cernicko p., Fatnicko p., Popovo
p., Mokro p., Hutovo blato) and receives water from riv¬
ers from other poljes: Brova (Ljubomorsko p.) and Bukov
potok (Ljubinsko p.); it resurfaces several times at power¬
ful springs: Obod (max. 35 m 3 s -1 ), springs of Trebisnjica
(max. >800 m 3 s '), Ombla (max. 112 m 3 s _1 ), and sev¬
eral others; it runs through large capacity ponors: SrQevici
(60 m 3 s' 1 ), Pasmica (25 m 3 s _1 ), Tucevac (>20 m 3 s _1 ),
Doljasnica (55 m 3 s _1 ), and several more; it changes the
name from the first spring of Dobra voda, to Vrba, then
Musnica, Kljucka Rijeka, and Obod, to Trebisnjica,
Ombla, and Krupa; and ends partly in a series of subma¬
rine springs (some 35 have been reported in a narrow
coastal strip off the coast near Vjetrenica, at the site of
Doli). The most prominent part of the network of flows
is the final surface stretch named Trebisnjica, of total nat¬
ural length of 94.5 km, and with a subterranean section
with a total areal distance of almost 34 km from the point
of sinking to the point of resurfacing (Srdevica ponor
and Dejanova pecina Cave). Popovo Polje is the final
stretch of the blind valley of Trebisnjica, which is, at 68
km 2 , some 40 times smaller than the actual basin. In a
matter of days, it would turn into a lake with depths
reaching 40 m and with water staying there for an aver¬
age of 248 days, eventually exiting in several directions,
via a web of subterranean flows of maximum total
capacity of 250 m 3 s _1 , including well-developed exam¬
ples of bifurcations and intersecting subterranean flows.
For that reason, for most of the twentieth century,
Vjetrenica was explored as the key to understanding the
peculiar functioning of the karst, not only in terms of
how its subterranean areas relate to the surface topogra¬
phy, but also as an important drainage canal of the
Trebisnjica basin and Popovo Polje.
SPELEOMORPHOLOGY OF VJETRENICA
The cave is mildly forked, with the Main Channel
(Glavni kanal; see Fig. 1, item 1) as the central level,
angling horizontally toward the southeast and south, of
total length of 2465.9 m. The Main Channel is spacious,
with width varying from 5 m to an extreme width of
25 m (the average width is 9.6 m), with an inclination of
5.6°, and with several pronounced chambers. Exceptions
include the first section of 160 m, with a lower ceiling,
usually of average human height, and several narrow
Encyclopedia of Caves.
© 2012 Elsevier Inc. All rights reserved.
VJETRENICA CAVE, BOSNIA AND HERZEGOVINA
859
VJETRENICA CAVE
Popovo polje, Bosna i Hercegovina
Cave Depth: 141,2 m
0 m 500 m
FIGURE 1 An overview of Vjetrenica with key sites. Map by Darko Baksic.
stretches inside the cave. Those are the Siphon (Sifon,
1007.4 m from the entrance; see Fig. 1, item 11), a low¬
ered area covered in water for most of the year, a sedi¬
ment deposit at the entrance to the Cvijic Chamber
(Cvijiceva dvorana, 1755.4 m from the entrance), as well
as a collapse in the Cvijic Chamber (1919.3 m from the
entrance). The last section is crossed diagonally by the
Main Channel, with a recently discovered narrow pas¬
sage through the blocks named Ass (Prkno—a tradi¬
tional name for narrow passages in estavelles used by
visitors seeking potable water; see Fig. 1, item 26), which
is why the rest of the Main Channel is called the
Flidden Main Channel (Skriveni glavni kanal, see Fig. 1,
item 27). In addition to the Cvijic Chamber (see Fig. 1,
item 22), there are also the Golden Chamber (Zlatna
dvorana, 363.2 m from the entrance; see Fig. 1, item 4,
Fig. 4), the Deep Chamber (Duboka dvorana, 565.4 m
from the entrance; see Fig. 1, item 6), the Tall Chamber
(Visoka dvorana, 1506.5 m from the entrance; item 16 in
Fig. 1), and the Welsh Channel (Velski kanal, 2133.3 m
from the entrance; items 24—25 in Fig. 1).
The floor of the Main Channel is covered with
blocks reaching as far as 650 m from the entrance,
occasionally masked by spacious sequences of calcite
flowstones and rimstone pools, and the rest of the
channel is covered with sediments of clay, loam, and
occasionally sand, with stone blocks fitted over them
in several areas. The actual width and height of the
channels in Vjetrenica reach almost 90 m. The Flidden
Main Channel is almost entirely closed off by a collec¬
tion of breakdown blocks and is the actual end of
Vjetrenica, as it is known today.
Different channels rise from the Main Channel. Most
of them are impassable and insufficiently explored verti¬
cal channels, chimneys, connecting the Main Channel
with the topographic surface above. Their actual connec¬
tion with the surface has not been demonstrated,
although there are several pits and caves on the surface,
and the places where some of them connect with the
Main Channel are marked by large calcite flowstones,
some of them exceeding several tens of meters. The larg¬
est and best explored channels above the level of the
Main Channel are the Welsh Channel and the Leopard's
Channel (Leopardov kanal; item 19 in Fig. 1), the Ravno
Channel (Ravanjski kanal; item 31 on Fig. 1) and the Clay
Figurines Channel (Kanal glinenih figurica; 2410 m from
the entrance; item 30 on Fig. 1), the last three complex
expansions and vistas, attractive beds of permanent
streams, two of which flow toward the Main Channel,
and the Clay Figurines Stream runs southward, not con¬
necting with the Main Channel. The Leopard's Channel
and the Ravno Channel run as occasionally very narrow
meanders, with sections layering out, creating a labyrinth
of passages. These four channels are located deep in the
interior of Vjetrenica, two with orientation toward the
entrance, and one, the Leopard's Channel, runs perpen¬
dicular to the Main Channel. All have stone bases.
Several networks of channels are concentrated around
Lower Vjetrenica (Donja Vjetrenica; item 3 on Fig. 1) and
the Absolon Channel (Absolonov kanal; items 8 and 10
on Fig. 1), farther from the Main Channel. They are youn¬
ger, narrower, with clean stone edges, with active water
flow, though located close to the entrance of the cave
and oriented toward it. This means that almost all the
peripheral channels are inclined due north, that is,
toward the entrance.
In addition to layers of clay and loam there are
numerous examples of calcite sediments, with particu¬
larly prominent rimstone pools called kamenica (tradi¬
tionally also called pjati, Italian for plate. Fig. 2) and
ENCYCLOPEDIA OF CAVES
860
v
FIGURE 2 Rimstone pools—the so-called plates which open the
northern catchment of Vjetrenica, at some 1500 m from the entrance.
Photograph taken by Darko Baksic.
flowstones. A sequence of pjati covers the front part of
the Main Channel, some 150 m. Of the flowstones, the
most prominent are the Yellow Flowstone (Zuti saljev,
in the Golden Chamber, Fig. 4, right side), the White
Flowstone (Bijeli saljev, 2100 m from the entrance), and
several others. There are smaller formations, often ara¬
gonites. Since karstification of the area is well advanced
and has led the cave into its mature stage, there are
numerous examples of broken ceilings and distorted
sediments. This is particularly visible on calcite forma¬
tions, which have been destroyed in some places and
with some sections of the Main Channel with no calcite
sediments at all. There are also different forms of sub¬
terranean corrosion (anastomoses) and erosion (facets),
as well as examples of destruction caused by changing
hydrology.
Vjetrenica has a network of water bodies, which
includes flows (permanent streams) such as the Absolon
Stream (Absolonov potok), the Leopard's Stream
(Leopadov potok), the Ravno Channel Stream (Potok
Ravanjskog kanala), the Flanging Lakes Stream (Potok
visecih jezera), and the Clay Figurines Stream (Potok gli-
nenih figurica), flowing in different directions and at dif¬
ferent heights, along with lakes; in addition to the Great
Lake (Veliko jezero. Fig. 3), 180 m long, there are many
smaller ones. The interior of the cave reacts quickly to
precipitation and water bodies also increase into a series
of occasional lakes. The Main Lake expands and floods
most of the Main Channel for as far as 400 m, thus clos¬
ing of the Siphon. Sources indicate that interior parts of
the cave, inaccessible in winter, also merge into a single
large lake. There are numerous occasional flows as well.
The hydrology of Vjetrenica is organized in two
interior catchment areas, delimited by the Dividing
Flowstone (Vododjelnica, at 1505.5 m from the entrance)
one evacuates the water toward the north, via the Great
FIGURE 3 The Great Lake, view of the exit: to the left is the
Hadezija Cone, the site where Hadesia vasiecki was first found.
Photograph taken by Ivo Lucic. After Lucic (2003).
FIGURE 4 The Golden Chamber, view of the exit: yellow flowstone
to the right, mother rock to the left, with numerous erosion and corro¬
sion formations. Photograph taken by Ana Opalic. After Lucic (2003).
Lake, the Great Lake Stream and the Absolon Stream to
the Lukavac spring, located at the edge of the slope and
the floor of the Popovo Polje near the entrance of the
Vjetrenica cave; the second runs toward the interior of
the cave, due south. There are instances of successive
reduction of subterranean flows: the Great Lake Stream
has four visible ponors in the Main Channel, from 650
to 800 m from the entrance. The last is close to the most
likely local erosion basis, the Absolon Stream.
MICROCLIMATE
In addition to a powerful stream of air at the entrance
area, milder currents are present in some narrow sec¬
tions, particularly at the Siphon and the entrance to the
Leopard's Channel. At higher outdoor temperatures,
the air current is steady as it runs toward the entrance
and maintains a steady climate. At lower outdoor
ENCYCLOPEDIA OF CAVES
VJETRENICA CAVE, BOSNIA AND HERZEGOVINA
861
temperatures, the direction of the air current from inside
the cave changes. Since the cave faces the north, the
strong northerly winds enter the front of the cave and
reduce the temperature. The speed of the air current can
increase to 10 m s 1 or greater. The air temperature
averages 11°C, with seasonal variations. For example, on
February 7, 2008, at 07:20 AM, the recorded temperature
was —3.85°C, relative humidity 38.75% and wind speed
9.83 m s at the same time, an outdoor station at some
300 m from Vjetrenica showed the outdoor temperature
of — 4.33°C, relative humidity of 33.25% and wind speed
of 3.15 m s -1 . The water temperature corresponds to the
air temperature with very little deviation, though there is
a notable decrease in the winter, below major streams.
The soil temperature below the surface may drop to
10.2°C. During average air circulation, CO 2 levels vary
from 400 to 600 ppm. All the values were registered by
monitoring equipment placed close to the entrance.
HABITATS
Vjetrenica has many habitats. Terrestrial habitats
include rocks, calcite sediments, rock, clay and loam sur¬
faces, crevices, and areas under rocks. Aquatic habitats
include fast streams in pure rock beds, slow streams
running over clay surfaces, sand deposits in lakes, as
well as sedentary water: larger lakes and small pools
over clay surfaces, occasional lakes on clay surfaces,
small lakes with sediments of sand, lakes with erosion
depressions on rocks surfaces, and small occasional sur¬
faces on rimstone dams and pools. There are several
hygropetric habitats—flowstones with a thin water flow
enveloping them (Fig. 3, left). There is considerable vari¬
ation in the level of potassium, sometimes as low as
0.1 mg L _ 1 in some lakes, though somewhat higher in
most of the water, and sometimes as high as several
tens of milligrams. Organic particles, often brought by
streams dripping from wide cracks in the ceiling, can be
found all along the cave, including the deep interior
(Sket in Lucic, 2003).
FAUNA
Due to its past geological development and abundant
karst elements that offer versatile subterranean habitats,
Vjetrenica (together with the Bjelusica Cave and the
Lukavac spring) is rich with fauna. It comprises a total of
219 taxa, of which 37 protists and 182 animals with 101
cave-dwelling taxa: 49 troglobiotic (obligate terrestrial
subterranean species) and 52 stygobiotic (obligate
aquatic subterranean species). Sorted by taxonomic
groups, they are: Turbellaria (5), Hydrozoa (1),
Gastropoda (12), Bivalvia (1), Nemertina (1), Polychaeta
(1), Oligochaeta (6), Hirudinea (1), Palpigradi (1), Araneae
(6), Opiliones (2), Pseudoscorpiones (3), Copepoda (6),
Ostracoda (4), Decapoda (4), Isopoda (7), Amphipoda
(12), Chilopoda (2), Diplopoda (5), Collembola (4),
Diplura (1), Thysanura (1), Coleoptera (14), and
Vertebrata (1). The fauna is highly localized with 57 taxa
endemic to Popovo Polje region, 25 of them only in
Vjetrenica and related Bjelusica and Lukavac Spring. The
distribution is still unknown for 37 protists and 24 ani¬
mals, determined only up to family or genus levels, so
levels of endemism may be even higher.
Vjetrenica is the type locality for 38 taxa: Scutariella
stammeri Matjasic, 1958,* Stygodyticola hadzii Matjasic,
1958 (Temnocephalida, Scutariellidae), Lanzaia vjetrenicae
Kuscer, 1933,* Zavalia vjetrenicae Radoman,* 1973
(Gastropoda, Flydrobiidae), Vitrea kiliasi Pinter, 1972
(Gastropoda, Pulmonata, Zonitidae), Eukoenenia remy
Conde, 1974* (Palpigradi, Eukoeniidae), Stalagtia
(Stalagtia) hercegovinensis (Nosek, 1905), Stalitella noseki
Absolon & Kratochvil, 1933 (Araneae, Dysderidae),
Tegenaria conveniens Kulczynski, 1914 (Araneae,
Agelenidae), Lephtyphantes spelaeonim Kulczynski, 1914*
(Araneae, Linyphiidae), Chthonius (Chthonins) occultus
Beier, 1939 (Pseudoscorpiones, Chthoniidae), Neobisium
vjetrenicae Hadzi,* 1933 (Pseudoscorpiones, Neobisiidae),
Dinaria vjetrenicae (Hadzi, 1932)* (Laniatores, Travuniidae),
Diacyclops karamani (Kiefer, 1932), Eua/clops imrmatus
Kiefer, 1932, Acanthocyclops troglophilus (Kiefer, 1932)
(Crustacea, Cyclopidae), Pseudoa/pridopsis hartmanni
Petkovski et al, 2009 (Ostracoda, Cyprididae), Troglocaris
hercegovinensis (Babic, 1922) (Crustacea, Decapoda, Atyidae),
Troglomysis vjetrenicensis Stammer, 1936* (Crustacea,
Mysidacea, Mysidae), Monolistra (Psendomonolistra) hercegovi-
niensis Absolon, 1916 (Crustacea, Isopoda, Sphaeromatidae),
Proasellns hercegovinensis (S. Karaman, 1933) (Crustacea,
Isopoda, Asellidae), Annadillidium absoloni Strouhal, 1939*
(Crustacea, Isopoda, Armadillidiidae), Niphargus balcanicus
(Absolon, 1927), Niphargus boskovici S. Karaman, 1952,
Niphargus vjetemicensis S. Karaman, 1932, Niphargus tndlipes
Sket, 1958, Niphargus factor G. Karaman & Sket, 1991,*
Niphargus cvijici S. Karaman, 1950, Niphargus zavalanus S.
Karaman, 1950* (Crustacea, Amphipoda, Niphargidae),
Hadzia fragilis S. Karaman, 1932 (Crustacea, Amphipoda,
Hadziidae), Lithobius (Troglolithobius) sketi Matic &
Darabantu, 1968 (Chilopoda, Lithobiidae), Typhloiulus
(Attemsotyphlus) edentidus Attems, 1951* (Diplopoda,
Julidae), Plusiocampa (Stygiocampa) remyi Conde, 1947
(Diplura, Campodeidae), Hadesia vasiceki (J. Muller, 1911),
Nauticiella stygivaga Moravec & Mlejnek, 2002,*
Speonesiotes (Speonesiotes) schweitzeri Jeannel, 1941
(Coleoptera, Cholevidae, Leptodirini), Aphaenopsis
(Adriaphaenops) pretneri Scheibel, 1935*, Aphaenopsis
(Scotoplanetes) arenstorffianus Absolon, 1913 (Coleoptera,
Carabidae, Trechini).
An asterisk indicates taxa found in the Vjetrenica
system only.
ENCYCLOPEDIA OF CAVES
862
v
Some animals belong to monotypic genera:
Spelaeoconcha paganettii (Gastropoda), Marifugia cavatica
(Polychaeta), Velkovrhia enigmatica (Hydrozoa), Zavalia vje-
trenicae (Gastropoda), Stalitella noseki (Araneae), Dinaria vje-
trenicae (Opiliones), Typhlogammarus mrazeki, Spelaeocaris
pretneri, Troglomysis vjetrenicensis (Crustacea), Nauticiella sty-
givaga (Coleoptera), and Proteus anguinus (Vertebrata). The
cave-dwelling fauna of Vjetrenica includes some fauna of
special interest. The hydrozoans Velkovrhia enigmatica is a
unique species of a single genus, in a unique freshwater
species of the family BougainviUiidae. The clam Congeria
kusceri is a relict, a living fossil, the only survivor of a line¬
age of 100 Congeria species present during the Upper
Miocene. Marifugia cavatica is the only freshwater serpulid
worm and the only stygobiotic polychaete. Pholeoteras
euthrix (Gatropoda, Prosobranchia) is the only known
cyclophorid in Europe, a relic of old, tropical fauna. Proteus
anguinus is the only European stygobiotic salamander.
With ten species of genus Niphargus (Amphipoda)
Vjetrenica holds a world record as a single cave; seven of
which also have their type locality there.
PALEONTOLOGY
Vjetrenica is a paleontological site which acted as a natu¬
ral trap in the Pleistocene for the leopard ( Panthera pardus).
Skeletons of four leopards have been found in different
parts of the cave, one of them very well preserved. All the
findings of the species, now extinct in Europe, were almost
2 km of the modern-era entrance, either found on the sur¬
face or under shallow sediment. The level of preservation
of the skeletons, the number of samples, and the fact that
they were found deep in the interior make Vjetrenica a
unique Pleistocene paleontological site.
CULTURAL HISTORY
Legends, myths, and material cultural remains tes¬
tify that throughout recorded history, Vjetrenica has
been considered a prominent spiritual site. This fact
gave a primary determination to its place and function
in the cultural communities it served. Myths say that
Vjetrenica was the place where fairies gathered, danced,
and sang, and that the "song" coming from inside the
cave was, in fact, a lullaby for the fairies' children, and
so on. Sounds from the Vjetrenica cave were often com¬
pared with bagpipes. A written source from between
1580 and 1584 claims that visible tracks of "cracked"
footprints were visible on the floor (according to leg¬
end, fairies had cloven hooves instead of feet). Until as
late as early twentieth century, peasants from Zavala
swore they had seen the fairies with their own eyes.
Vjetrenica also had a system of sounds used for telling
the future. They appeared in several places in the ceil¬
ing of one of the chambers, called the Millstone, the
Drum, and the Mill. Particular sounds, varying from
one to three, could be heard at the entrance to the cave.
In short, its own mythology always rested on the inter¬
pretation of phenomena related to air currents.
Only traces of material artifacts remain, as they
were gradually removed by new cultures: the prehis¬
toric human habitat was at the very entrance, which
was also the site of discovery of Roman coins as well
as coins minted in Dubrovnik, as well as a medieval
tomb, destroyed in modern times. The only reminder
of the tomb is an interesting ornament, typical of medi¬
eval tombstones in Bosnia, called stecak. It is unusual
that instead of the commonly used freestanding mono¬
lith, the stecak used a rock at the entrance, with two
shallow carvings. The visual expression of the carvings
is also original and authentic, with integrity that fits
perfectly with the cave itself. Just like elsewhere, it
depicts old motifs of hunting and tournaments, sym¬
bolizing mystic rituals striving toward immortality,
present since ancient Trace. Here, however, two
opposing horsemen are not separated by the com¬
monly found image of a woman or a symbol represent¬
ing female deity (a circle, a rosette, a crescent moon,
etc.) (Wenzel, 1961). The reason is simple: whoever
stood in front of the relief could confirm the presence
of a female deity (a fairy), since her voice spoke
through the wind from Vjetrenica.
According to some sources, the sacred status of the
cave evident in pagan times continued also in the
Christian era, though this remains only marginally
explored. Namely, there are traces of an edifice at the
entrance to Vjetrenica. Some interpret it as the pres¬
ence of a shrine of the medieval Bosnian Church,
and some as the castle of an unknown nobleman.
Historical sources indicate that the interior of the edi¬
fice used the cave wind to keep cool, which is an inter¬
esting and rare example of early air-conditioning.
Vjetrenica seems to have been abandoned as human or
animal shelter for a long time, though remains of
domestic animals indicate that this period may not
have been all that long. Vjetrenica has also been the
generator of sanctity of the surrounding area, as certi¬
fied by material remains of different cultures. There
are the foundations of an early Croatian Church of
St. Peter (with cultural elements specific to the ninth to
twelfth centuries), subsequently surrounded by a
necropolis of late medieval tombstones, as well as a
medieval Eastern Orthodox Monastery of Presentation
of the Blessed Virgin Mary, with the church built on
an inferior cave, partly connected to live rock. Its inte¬
rior is richly decorated with frescos dating back to
1619. The construction of this church disturbed the cul¬
tural sediments where a Mesolithic tool was found.
ENCYCLOPEDIA OF CAVES
VJETRENICA CAVE, BOSNIA AND HERZEGOVINA
863
EXPLORATION OF VJETRENICA
The known history of Vjetrenica is a web of mythol¬
ogy and science conditioned by the ever-present fasci¬
nation with the "wind." The oldest written sources are
those from Historia Naturalis by Pliny the Elder, from
the year 77 BC. When speaking about the sources of
wind (Plinio Secondo, 1884:245) the source speaks of an
unnamed cave on the Dalmatian coast, which
the Renaissance tradition of Dubrovnik names as
Vjetrenica. The Dubrovnik writer Benko Kotruljevic
wrote in 1464 that the wind at its entrance was colder in
the summer than any winter wind in Italy. The duke of
the nearby town of Ston (in the medieval Dubrovnik
Republic), Jakov Sorkocevic, wrote (1580— 1584) to the
natural scientist Ulise Aldrovandi and gave its first
description. In his Sopra le Metheore d' Aristotile (1584),
Nikola Gucetic Gozze, a leading intellectual, scientist,
and politician of the then Dubrovnik Republic, ana¬
lyzed the wind and how it differed from the Sipun cave
in Cavtat. He explained it via Aristotle's natural philos¬
ophy, on the basis of the four elements—earth, water,
wind, and fire—and four features—warm, cold, dry,
and wet. Inside the cave, the sun triggered the evapora¬
tion of the element of earth, which was warm, dry, and
thick. Once it reached the cold interior surface of the
cave, it stopped and as it found an exit at the entrance,
which is where it created the great wind. In Sipun, how¬
ever, there was no evaporation, since the interior was
damp, hence no wind, according to Gucetic. In 1830, a
monk from the Eastern Orthodox Monastery of Zavala
named Joanikije Pamucina wrote that, in his opinion,
the phenomenon of air currents and the creation of
sediments should be interpreted as a phenomenon of
physics. In 1850, a Catholic missionary named Antonio
Ayala compared the wind with the sound of a thousand
steam engines on the move.
The first research exploration was organized in the
late nineteenth century, following the establishment
of the Austrian rule. In July 1887, Joseph Riedel con¬
ducted the first modern temperature measurements.
The hegumen (head) of the Zavala Monastery, Hristifor
Mihajlovic, described Vjetrenica in 1889 and 1890 and
this description was accepted all across the monarchic
realm. He also produced the first sketch of the cave,
and after that (in 1893) Josip Vavrovic, a railway
engineer, produced an even more detailed description
and a map of the first 600 m. During his explorations in
the period 1912—1914, Karel Absolon, a Czech explorer
who conducted a total of 27 visits to Vjetrenica, discov¬
ered most of the segments known today, but he never
produced a map. He described his expedition as the
most difficult in the history of speleology, and he
insisted that Vjetrenica surpassed anything he had ever
seen inside any cave. He also suggested that it was there
that he established the measurements to assess limestone
formations in the Balkan region. Under instruction by
his mentor Jovan Cvijic, Mihajlo Radovanovic conducted
detailed explorations throughout the 1920s, including
detailed mapping. He presented his explorations in a
dissertation, which he then published as a monograph in
1929. In 1958, Bosnian geographers Ratimir Gasparovic
and Orhan Zubcevic produced a detailed geodetic map
of the Main Channel (up to 2084 m). In 1968, members of
the South Wales Cave Club found two previously
unknown channels of a total length of some 500 m.
New speleological explorations were launched in
2000, eventually producing a monograph entitled
Vjetrenica—A View into the Soul of the Earth (2003) and a
totally new and modern topographic map, published in
2002 (Lucic et al, 2002). Regular summer speleology
camps were organized from 2002 to 2009 by the
newly established Speleological Association Vjetrenica—
Popovo Polje. The camps led by Darko Baksic (Velebit
Speleology Section from Zagreb) brought together a
host of clubs and individuals from the region. The
camps uncovered new channels of total length of
1668 m. In 2004 the camp also established a biospeleol¬
ogy program, led by Roman Ozimec from the Croatian
Biospeleological Society from Zagreb. In an unexplored
channel, the SWCC speleologists found a complete
leopard skeleton; in the late 1990s Mirko Malez found
remains of a leopard and a brown bear; the Velebit
Speleology Unit explorers found remains of a leopard.
During the 2007 paleontological excavations led by
Kazimir Miculinic from the Croatian Biospeleological
Society from Zagreb, skeletal remains of three leopards
and one bear were excavated and handed over to the
National Museum in Sarajevo.
The most intensive exploration of Vjetrenica has been
by biospeleologists, though their findings were available
to a limited scientific community only. Vjetrenica
became attractive for biospeleologists after L. Vasicek
and Lucijan Matulic had found cave-dwelling insects,
described by Giuseppe (Josef) Muller as Antroherpon
apfelbeckia and Hadezia vasiceki. Exploration was boosted
by K. Absolon (research from 1908 to 1922) who unfor¬
tunately never published his list of the taxa. It was then
explored by members of the Slovenian Society for Cave
Research (1931), by Stanko Karaman in the 1950s, and
by Slovene biologists Janez Matjasic, Joze Bole, and
Boris Sket throughout the second half of the twentieth
century, as part of the visits organized by the Slovene
Academy of Arts and Science or the School of
Biotechnology of the Ljubljana University. More than 30
researchers have explored Vjetrenica and described its
fauna or submitted information for description. Most of
the new species were described by S. Karaman (9 taxa),
followed by K. Absolon (et al.) and K. Verhoeff (8 taxa).
ENCYCLOPEDIA OF CAVES
864
v
J. Matjasic (7), F. Kiefer and L. Kulczynski (5), and
others. They thus created a database that allowed Boris
Sket to prepare an integral description of the fauna of
Vjetrenica (2003). It was also of help in synthesizing con¬
clusions on the international prominence of Dinaric sub¬
terranean fauna. Vjetrenica was known for 75 cave¬
dwelling taxa at the onset of the biospeleological camp
program, and by 2009 101 taxa were described. The
recent exploration, which includes the first year-round
climate monitoring and new environment-sensitive
lighting for tourist access, has been followed by a host
of events aimed at promoting the cave, such as the pub¬
lication of the monograph, a web page, documentary
films, and numerous texts. This is also a new presenta¬
tion of karst, based on an environmental outlook, as
well as a new view of contemporary organization and
management of Vjetrenica.
VISITOR ACCESS
The first facilities to allow access to visitors date
back to the late nineteenth century, when the entrance
to Vjetrenica was expanded. The old entrance was low
and narrow, barely allowing a crouching adult to
enter, though often blocked due to strong wind from
the inside. A railway track was laid immediately in
front of Vjetrenica in 1901 and the entrance was fitted
with a metal frame and a gate with vertical bars. The
first visitor trail of 1250 m was built between the two
world wars. Construction included flattening of the
footsteps over small rock, placement of a fine stone
layer over clay beds, placement of gravel over rim-
stone pools to create paths. In the early 1960s, the path
was expanded to 110 cm and extended to 1700 m; elec¬
tric lighting was introduced on 1045 m of the path in
1964, along with a full-time guide service and a motel
in the vicinity. Official reports of 7000—10,000 visitors
per year seem conservative. All the equipment and
facilities for visitors were destroyed during the
1991—1995 war, and real interest in the revival of tour¬
ism was triggered by the new speleological research at
the onset of the twenty-first century.
ECOLOGICAL CHANGES
Following the massive hydropower construction on
the River Trebisnjica (1965—1979), which reduced this
karst river to just 67 km and gradually converted
almost all of its flow into a concrete canal, thus reduc¬
ing its environment of some 4 billion cubic meters of
water, Vjetrenica was at the very edge of the endan¬
gered area, though not physically included. The villages
above it have no industrial facilities, but there is also
no garbage disposal system. Also, the 1991—1995 war
left behind damaged military technology based on
PCB compounds, which no one has yet investigated.
Conversion of Vjetrenica to accommodate visitors
included several expansions of the entrance—which
terminated the magical sounds—excavation to set
paths, introduction of construction materials and metal,
tiling and laying of concrete, defragmentation of rock
and sediments, gravelling and placement of gravel over
plates and flowstones, disintegration of several lakes,
introduction of powerful lighting, which led to the gen¬
eration of lampenflora, as well as collection of garbage.
And sources identify a disposed camera battery as the
principal cause of the disappearance of several crusta¬
ceans from some of the ponds of the Main Channel. The
extinction of the entire fauna in the Third Calcite Lake
was noted in the early autumn of 1996, but no causes
were identified. Vjetrenica is threatened by the post¬
socialist transition of a war-torn country, where a cave
is used for commercial exploitation with absolutely no
systematic state-backed protection of the natural sys¬
tem. The struggle to protect Vjetrenica culminated in
early 2009, with a 14-day hunger strike of an activist
from a local NGO. Although the struggle to protect it
enjoys the support of many prominent international
professional speleologists, there has been no response
from Bosnian authorities.
PROTECTION
Vjetrenica was declared a natural monument (1950)
and was subsequently declared a special geological
reservation (1965 and 1985). In 2004, the BiH Academy
of Arts and Science nominated it for the UNESCO List
of Natural Heritage Sites and in 2007 the National
Commission for Monuments expanded the nomination
to include the architectural ensemble of the village
Zavala. However, at the time of this writing, in late
autumn of 2010, it remains only on the tentative list.
Translated into English by Amira Sadikovic.
See Also the Following Article
Diversity Patterns in the Dinaric Karst
Bibliography
Lucic, I. (2003). Vjetrenica—pogled u dusu zemlje. Ravno, Zagreb,
Croatia.
Lucic, I., Baksic, D., Mulaomerovic, J., & Ozimec, R. (2005). Recent
research into Vjetrenica cave (Bosnia—Herzegovina) and the cur¬
rent view of the cave regarding its candidature for the World
Heritage List. Proceedings of the 14th International Congress of
Speleology, Athens, August 21—28, 2005 (Vol. 2, pp. 430—435)
Athens: Hellenic Speleological Society.
ENCYCLOPEDIA OF CAVES
VOLCANIC ROCK CAVES
865
Ozimec, R., & Lucic., I. (2010). The Vjetrenica cave (Bosnia &
Herzegovina)—One of the world's most prominent biodiversity
hotspots for cave-dwelling fauna. Subterranean Biology, 7, 17—23.
Wenzel, M. (1961). A medieval mystery cult in Bosnia and Herzegovina.
Journal of the Warburg and Courtauld Institutes, 24, 89—107.
VOLCANIC ROCK CAVES
Stephan Kempe
Institute of Applied Geosciences, University of Technology
Darmstadt, Germany
INTRODUCTION
Exploration of volcanic rock caves, termed vulcanospe¬
leology by William R. Halliday, the Nestor of the field,
did not receive much attention until the 1970s. Before,
it remained a topic of local caving interest, for example,
in Australia, in the western United States on Tenerife,
Sicily, and Iceland, or a topic of ecologists looking at
biological diversity. One of the reasons for this general
lack of attention may have been the question of accessi¬
bility but even more so the term lava tube may have
suggested "tubular caves" of little morphological vari¬
ety and interest ("you have seen one, you have seen
them all"). Even in volcanological textbooks, the impor¬
tant role of lava tubes for the lateral transport of lava
and the shape and functioning of shield volcanoes has
rarely been acknowledged. These textbooks would
publish at most one picture of a lava tube (very often
of Thurston Lava Tube, Hawai'i), state that they form
by "crusting over of channels" and then move on.
The beginning of lava cave research on Hawai'i in the
1970s and the consecutive acknowledgment of the enor¬
mous Hawai'ian cave potential spurred activity there
and worldwide, leading to the foundation of the Hawaii
Speleological Survey in 1989 during the International
Congress of Speleology in Budapest, Hungary, and the
International Commission of Vulcanospeleology in 1993.
Furthermore, the Vulcanospeleological Symposia, initi¬
ated by W.R. Halliday in 1972 (Halliday, 1976) and held
every other year (the 14th was in Australia, August
2010), is attracting many speleologists and other scien¬
tists. The proceedings of these symposia form the largest
body of professional papers available in this field.
They are now accessible at www.vulcanospeleology.
org/symposia.html. At the International Congresses of
Speleology, lava cave symposia are also held; the papers
of the latest one in Kerrville, Texas, August 2009, are
available at http://www.karstportal.org/FileStorage/
IntlCongressofSpeleology/15thproceedings-v002.pdf.
Other sources are the Newsletters of the Hawaii
Speleological Survey and the Hawaiian Grotto of the
National Speleological Society. U.S.-centered termi¬
nology of vulcanospeleology was summarized by
Larson and Larson (1993).
In general lava caves, or better, caves in lava, can be
formed by many different processes. Most of them are
primary caves, but secondary caves occur as well. Kant
(or one of the editors of his work in 1803) was probably
the first to differentiate between primary and second¬
ary caves, albeit on the basis of quite abstruse surmises.
SECONDARY VOLCANIC ROCK CAVES
Secondary caves in volcanic rocks, that is, created
long after the lava was deposited, form tectonically, by
collapse or by erosion.
Tectonic caves are still poorly documented. Along
the Great Crack, the southwest rift zone of the Kilauea,
several essentially tectonic fissure caves have been
explored, including Pit H (183 m deep) and the Wood
Valley Pit Crater (90 m deep).
Pit craters form by cold collapse of the roof of small
sections of a magma chamber (collapse of the entire
roof of a magma chamber leads to the formation of
large-scale depressions, called calderas). One recent
example is Devil's Hole on the Island of Hawai'i
(Hawai'i National Park, Chain of Craters Road), that
started as a small hole in the ground leading into a
conical chamber about 50 m deep. Further collapse of
ceiling and walls led to an open pit over the course of
a few decades. Many other older pits occur in the
vicinity. Pit craters have recently obtained more atten¬
tion because they may be the first cave-like objects
detected on Mars. Collapse of hypogene karst caves
and the consecutive stoping upward of the cavity can
also result in lava caves. An example is the Basalthohle
near Ortenberg, Germany, a chamber in columnar
basalts that was intercepted by quarrying for basalt.
Erosional caves in lava may either be formed by waves
along coasts or by running water. Wave-cut caves in vol¬
canic rocks are quite common and may be of substantial
size. The most famous is Fingal's Cave of the Island of
Staffa, Scotland. Some of the most spectacular examples
occur along the cliffs of the Hawai'ian Islands, specifically
on the Na-Pali Coast of Kauai. Sometimes basaltic
dykes are eroded by waves leading to long and narrow
fissure-like sea caves such as the >40-m Sorte Gryde on
Bornholm, Denmark. Quite common are also small caves
that are produced by the lateral erosion along river val¬
leys that cut down into stacked series of lava flows. The
flowing water can then preferentially remove loose a'a
rubble from underneath the solid cores of the a'a flows
that serve as cave roofs. For example, such caves occur
along the sides of Wadi Rajil in Jordan near the famous
Bronze Age city of Jawa. Water entering primary lava
Encyclopedia of Caves.
© 2012 Elsevier Inc. All rights reserved.
866 v
caves can cause internal water erosion such as documen¬
ted for the Pa'auhau Civil Defense Cave on Mauna Kea,
Hawai'i (Kempe et al., 2003). In this cave, several spectac¬
ular waterfalls eroded the bottom of the cave, polishing
the rocks and gravel to a surprising extent. It is difficult
to understand how water can erode a cave in lava with¬
out following an existing lava cave, such as the
kilometer-long and 100-m-deep Kuka'iau Cave (Kempe
and Werner, 2003) which is formed in a series of
pahoehoe flows and diamict layers in weathered Mauna
Kea lavas. It not only pirates the tributaries of a parallel
valley but it also forms a series of sumps and chutes
where the water moves upward, similar to karstic caves
under phreatic conditions. At flood, this cave transports
large amounts of water, gravel, rounded blocks, and trees
and may pond up to 60 m above its sump.
PRIMARY VOLCANIC ROCK CAVES OF
LARGE EXTENT (PYRODUCTS)
Tunnels formed by the lateral transport of lava
through interior conduits of pahoehoe lava flows com¬
prise the most important group of primary caves world¬
wide. Within the tunnel, heat is lost only conductively
while surface lava loses heat convectively at the upper
and conductively at the lower interface. Tunnel trans¬
port is thus the reason why basaltic shield volcanoes
have low flank-slopes, often <2° (compare Table 1) and
why multiple eruptions can form extensive intraconti¬
nental lava plateaus. What appears later on the geologi¬
cal map as an individual pahoehoe flow is actually a low
ridge that grew only at the downhill end being supplied
by the internal duct that in most cases is not visible to
the outside. In fact, one can walk across an active tunnel
without even noticing it. A'a flows, on the other hand,
move like glaciers, gliding downhill with their entire
mass, moving like a caterpillar on a bed of rubble that
falls down at the front and is overridden. On Hawai'i
a'a flows do not develop tunnels.
The first to describe a lava tunnel having inspected
it himself was probably Eggert Olafsen (1774—1775,
§ 358, p. 130) who visited Iceland from 1752 to 1757.
He wrote about Surtshellir: "... the running lava
flowed through this channel like a river...." Uno von
Troil (1779, p. 225), who visited Iceland with Joseph
Banks and Daniel Solander in 1772, wrote (transl. from
German by author):
The upper crust sometimes cools and solidifies, even
though the molten matter keeps running underneath; in this
way large caves form, the walls, floors and ceiling of which are
composed of lava and where a lot of dripstones of lava occur.
TABLE 1 Comparison of Some Morphological Indices of Some of the Hawaiian Lava Tunnels
Cave
Total
Length, km
Main Trunk
Length, km
End-to-End
Bee-Line, km
Sinuosity
Vertical
Distance, m
Slope
Volcano
Kazumura Cave
65.50
41.86
32.1
1.30
1101.8
1.51°
K, A
Ke'ala Cave
8.60
7.07
5.59
1.25
186
1.51°
K, A
J. Martin/Pukalani System
6.26
K, A
Epperson's Cave
1.93
1.13
0.80
1.41
-
-
K, A
Thurston Lava Tube
0.490
0.490
432
1.13
20.1
2.4°
K, A
Ainahou Ranch System
7.11
4.82*
4.27
1.13
323
3.83°
K, A
Ke'auhou Trail System
3.00
2.27
1.99
1.13
213.3
5.36°
K, A
Charcoal System
1.5
1.4
60
2.6°
K
Earthquake System
0.34
33
4.7°
K
Huehue Tube
10.8
6.17
5.13
1.2
494.6
4.58°
H, HH
Clague's Cave"
2.73
1.39
1.18
1.15
157.1
6.49°
H, HH
Pa'auhau Civil Defense C.
1.00
0.58
0.50
1.14
49
4.87°
MK
Whitney's Cave
0.651
0.509
0.438
1.15
17.3
1.97°
ML
Manjang Gul, Jeju, Korea
4.304
3.197
1.32
32.4
O
►L*
0
Jeju
*horizontal
**upper part of Huehue
For sources of data see Kempe, 2002.
ENCYCLOPEDIA OF CAVES
VOLCANIC ROCK CAVES
867
The first to report having seen an active tunnel was
Coan (1844) who, in 1843, scaled Mauna Loa:
But we soon had ocular demonstration of what was the
state beneath us; for in passing along we came to an opening
in the superincumbent stratum, of twenty yards long and ten
wide, through which we looked, and at the depth of fifty feet,
we saw a vast tunnel or subterranean canal, lined with
smooth vitrified matter, and forming the channel of a river of
fire, which swept down the steep side of the mountain with
amazing velocity. The sight of this covered aqueduct—or, if I
may be allowed to coin a word, this pyroduct —tiled with min¬
eral fusion, and flowing under our feet at the rate of twenty
miles an hour, was truly startling.
Even though Coan used the more general term
tunnel to start with, he coined a new one: pyroduct. This
highly specific term should take precedence over youn¬
ger terms describing the same phenomenon. Early geol¬
ogists such as James Dana continued to use the word
"tunnel"; J.W. Powell introduced the term volcanic pipes
and Tom Jaggar used "tunnel" as well as "tube"; only
after 1940 did the term lava tube become standard
(Loowood & Hazlett, 2010). For reasons of scientific pri¬
ority, the term pyroduct should be used. Avoiding the
word "tube" is also advisable, because it has invoked
the picture of pipes in which lava can flow up and
down under pressure like in plumbing and it implies a
circular cross-section that is only rarely found.
The discovery of extensive lava flows on Venus,
Mars, and the Moon and even active volcanoes on Io
has increased the interest in pyroducts in the past few
decades. On Earth, the longest surveyed lava tunnel is
Kazumura Cave (main trunk length 41 km) (Hawai'i,
Kilauea Volcano) (Allred et al., 1997), and the longest
Quaternary pyroduct-fed flow on Earth is that of
Undara/Australia (Atkinson and Atkinson, 1995). The
author's group explored and surveyed many other caves
on Hawai'i {e.g., Kempe, 2002) and in Jordan that give the
opportunity to study formation and evolution of pyro¬
ducts from the inside. Many other areas are under investi¬
gation as well: The islands of Galapagos, Rapanui, Jeju,
Mauritius, Comores, Sicily, Iceland, Canaries, Azores, and
in the intracontinental lava fields of Syria, Saudi Arabia,
East Africa, the western United States, Mexico, and the
Andes. Petrographically, almost all pyroducts documen¬
ted (yet) occur in tholeiitic and alkali basalts, only the
caves in Mount Susua, Kenya, are formed in phonolites
(pers. com. C. Wood).
Formation of Pyroducts
Even though in many text books lava tunnels are
described as having formed by "crusting over of chan¬
nels" {e.g., Francis, 1993), there is a totally different
mechanism by which they can form, and that is by
inflation (Hon et al., 1994). This is an incremental pro¬
cess starting at the distal tips of pahoehoe flows where
hot lava rapidly covers the ground in a thin sheet. This
sheet cools quickly, causing the dissolved gases to
form vesicles that diminish the overall density of the
lava. This sheet will float on top of the next pulse of
advancing melt (the lava flow is "inflated" from below
because of buoyancy) before forming the next distal
surface sheet. Multiple advances can occur, forming a
primary roof composed of several sheets, separated by
sheer interfaces (only the first or top sheet will display
the subaerially formed typical ropy pahoehoe struc¬
ture) (Fig. 1). The "oldest" lava sheet is therefore on
top of the stack, contradicting normal stratigraphic
rules. Below the primary roof the lava can stay hot and
can keep flowing. This is the initial conduit. Inflation
caves are characterized by roofs built of one or several
sheets, sometimes more than ten continuous sheets of
lava. This roof structure can be studied at roof col¬
lapses, called pukas, in Hawai'i. Many of the long
Hawai'ian caves are inflationary in origin (Kazumura,
Keala, Huehue, Ainahou, Keauhou Trail; just to name
a few). However, the "crusting-over" of channels can
also lead to long caves. Inspection of cross sections of
the roof of the second longest cave on Hawai'i, the
Kipuka Kanohina System (37 km of interconnected
passages), has shown that it consists of welded, irregu¬
lar fragments of pahoehoe plates, stabilized by inserted
lava fingers {squeeze balls) and a 10—50-cm-thick lining
welding the roof from below. Such structure suggests
that the roof is formed by agglomeration of floating
lithoclasts that are wedged into each other similar to a
logjam on a river. Sometimes roofing-over may also
occur by accretion of vertical to subvertical thin lava
layers growing from the sides inward and having a
central vertical parting where the growing lateral
shelves meet. The levees of open-surfaced lava chan¬
nels grow by overbank events of thin lava sheets or by
stranded lava floats. Sometimes large sections of the
bank break loose and float downstream. These can
then jam and form short roofs that are stabilized
by spatter. Examples have been documented for the
channels of the Puhia Pele eruption of 1801 on Hualalai.
Internal Development
If the area to be covered by the first advance is
rather flat, many small, parallel conduits can develop
(Fig. 2). Each of them can start to erode down soon
after. One of the threads will, however, erode fastest
and attract most of the lava. This passage will then
drain the other parallel ducts, one by one, often leav¬
ing them as small-scale labyrinths high above the final
floor. Since these mazes are drained when they are still
ENCYCLOPEDIA OF CAVES
868
v
Longitudinal Section
Plan View
1. Advance
top sheet with ropy structure
1. Inflation
FIGURE I Sketch illustrating pyroduct (lava
tunnel) formation. At the tip of a pahoehoe flow
lava advances quickly in the form of a delta of
thin, ropy lava. The next pulse of lava lifts the
first sheet up (inflation). This process is
repeated until a stack of lava sheets (the pri¬
mary roof) is formed, below which the hottest
flow thread later becomes the main conduit.
2. Advance
2. Inflation
very hot, their floors are mostly smoother than that of
the later main tunnel where the terminal flow can con¬
vert to a'a rubble.
As the erosion continues, the lava flows with an
open surface in a self-generated underground canyon
cut into older rocks, not associated with the current
eruption. This fact can be studied at places, where the
thin lining of the side walls has fallen away (e.g.,
Greeley et al., 1998; Kempe, 2002). Often we find a'a
blocks behind the lining or even ash horizons,
both clearly not integral parts of pahoehoe flows.
Downcutting is facilitated by a variety of processes;
one of the more spectacular is backcutting of lavafalls
(e.g., Allred and Allred, 1997) (Fig. 3). These are quite
common in long Hawai'ian caves, but none has been
found yet in Jordan. The falling lava acts like a sledge
hammer, forcing loose rocks from the floor. These are
less dense than the lava itself and float up and are
transported on the surface of the river. Thus, in con¬
trast with a water river, the bed is not protected by
bedload and is therefore prone to continued erosion.
The mobilized blocks are cold and receive a coating of
lava, forming lavaballs. Some of the lavafalls seem to
be stationary forming large plunge-pools and cham¬
bers (Allred and Allred, 1997). Due to these erosive
mechanisms the cave grows in depth and width in an
uphill direction. The passage above the lavafalls is
quite small in contrast. As one proceeds uphill, the
canyon becomes larger and larger until one enters the
next plunge pool chamber. It is not quite understood
how much mechanical erosion and how much
melting of the river bed occurs ( e.g., Greeley et al.,
1998; and citations in Kempe, 2002). Other enlarging
processes may occur, such as small phreatic explo¬
sions, blowing out sections of the wall or floor as
intersected groundwater is vaporized.
As the downcutting continues the lava river mean¬
ders, undermining walls and destabilizing the roof.
Breakdown falling into the flowing lava is also carried
away. If the primary ceiling collapses entirely, a sky¬
light or puka opens up. If the flow is still active, the
rubble can be carried away and we speak of a hot puka.
If the collapse occurs after the flow terminated, break¬
down will remain, sometimes giving easy access to the
cave below, sometimes sealing it completely; this is
termed a cold puka. Through a hot puka gases can
escape from the tunnel, triggering convective cooling.
Heat loss is specifically efficient if two pukas open up:
cold air will be drawn into the lower one and hot gases
will escape from the upper one, freezing out a second¬
ary ceiling above the flowing lava in the canyon in
between the pukas, splitting the passage into two
levels on top of each other (Fig. 3). This process can be
repeated, forming even more internal ceilings. Later
spills from below through breakdown holes or from
upstream can reinforce these ceilings. In Ke'ala Cave,
ENCYCLOPEDIA OF CAVES
VOLCANIC ROCK CAVES
869
FIGURE 2 Detail of the ground plan of the Huehue Cave illus¬
trating how a primary maze of parallel lava threads (A) was drained
until the master trunk that cut down fastest remained (E). (A) Initial
pattern of lava tunnels; flow was from right to left in up to five par¬
allel conduits. (B) The southern-most conduits were drained first. (C)
Further downcutting reduced the number of active conduits to three.
At the same time, lava from a parallel flow (Mystery Flow) covered
the area. (D) Only one tunnel remained (following the thick line in B
and C), its bed 2 m below the original surface. The added overbur¬
den caused collapse of pukas (labeled 8, 6, and 4), the breakdown of
which was removed by the lava river. Now external air began flow¬
ing uphill from puka 8 to 4. As a consequence a secondary ceiling
froze out uphill and downhill of puka 8. (E) Various spill-events
reinforced the secondary roof, even spilling into the already drained
upper conduits, closing the northern most one. These spilled lavas
were oxidized by the passing surface air and attained various shades
of red caused by the crystallization of very fine-grained hematite.
The last event was the collapse of puka 5 while the tube cooled. Its
material is still in place.
Hawai'i, one section of secondary ceiling is over 1 km
long. Very often the upstream end of the secondary
ceiling is sealed. This is caused by lavaballs floating on
the lava river that are too buoyant to be dragged below
the secondary ceiling. Instead they strand on the upper
edge of the secondary ceiling. The accumulated blocks
are then welded together by splashed-up lava. Floating
blocks can be very large: in Waipouli Makai Cave there
is a block about 12 m wide, 8 m long, and 5 m thick
welded into the ceiling of the cave. The cold, oxygen-
containing external air that is drawn into the cave can
oxidize lava surfaces that are still hot. The iron, con¬
tained in the volcanic glass, is oxidized to fine-grained
hematite, tinting the surfaces of secondary ceilings in
various hues of red. If steam is present, goethite or limo-
nite can also be formed, introducing ocher or yellow
colors.
Hot pukas can also serve as temporary rootless vents
when the tunnel below is obstructed or even closed
entirely. Lava erupting from these pukas can form rap¬
idly cooling, thin, ropy pahoehoe, reinforcing the pri¬
mary roofs from above. The Puka 17 Flow out of the
lower part of the Huehue tunnel (Fig. 4) is an example.
Pukas, cold or hot, can also serve as entrances for lava
of later flows, such as at the upper and lower ends of
Ke'ala Cave which are plugged by later invasive lava.
All these processes act to form caves of complex
patterns, morphologically not representing "tubes" at
all. Also the total length of the caves is usually much
larger than the simple distance along the main tunnel.
Table 1 gives some basic morphometric data, such as
sinuosity and slope for some pyroducts.
General Types of Pyroducts
Overall, we can differentiate several general types
of lava tunnels. These are (1) single-trunked systems;
(2) double (or multiple)-trunked systems; and (3)
superimposed-trunked systems
Most of the lava tunnels so far documented in
enough detail appear to belong to the single-trunked
category. They are fed by one eruption vent and the
mesomorphological internal structure can be explained
by the processes discussed above. The tunnel size
depends on the lava discharge rate and on the length
of activity (days, weeks, months, and possibly even
years), that is, on the time available for erosion. If the
eruption stops or the cave collapses or is blocked, the
pyroduct will cool. The next or even—in case of a
blocked tunnel—the same eruption will then create a
new pyroduct. Normally, it will be situated to either
side of the previous flow because that now forms a
topographic ridge (flow lobe). Typical examples of sin¬
gle-trunked systems are Kazumura, Ainahou Ranch,
Ke'ala, and others of the long Hawai'ian caves. If lava
from the new tunnel should spill through a puka, or
break (because of its overburden) into any older,
underlying tunnel, then the older tunnel will be filled
by lava cooling in the same pattern as at the surface.
Double-trunked systems are comprised of two lava tun¬
nels, active side by side at the same time and fed by two
separate eruption points. Such tunnels can interact and
cause more complex morphologies than described above.
ENCYCLOPEDIA OF CAVES
870
v
Formation of Canyons by Back-cutting of Lavafalls
layered primary roof
spills reinforce the
secondary ceiling^frotn above
back-cutting continues below
the secondary ceiling
roof collapse opens
‘puka”
convective heat loss
leads to the formation of a
secondary roof
FIGURE 3 Longitudinal section of an evolving lava tunnel. Top: Erosional enlargement of the underground canyon by backcutting lavafalls.
Bottom: On static failure and partial collapse of the primary roof, a skylight (puka) opens up, allowing cold air to enter the tunnel, freezing an
internal secondary roof below which erosional enlargement can continue. Spills from uphill or through holes reinforce the secondary roof.
One example is the interaction between the Huehue
How and the secondary Mystery Flow (Fig. 4; Kempe,
2002). The Huehue Flow established its tunnel first, then
a second vent (the very inconspicuous, low Mysten/
Shield) erupted lava, forming a small tunnel in parallel.
Part of the Mystery lava quickly cooled, forming a'a
flows. These superseded the upper part of the Huehue
tunnel. Once thick enough, the primary, sheeted roof of
Huehue collapsed and left a roof composed of Mystery
a'a lava. The resulting breakdown was removed with the
active Huehue lava river. Due to the large, hall-like cav¬
ity that formed, a secondary roof froze out over the
active flow of Huehue. Later rockfall covering the newly
formed false floor gives the upper passage the appearance
as if the tunnel was formed in a'a, an impossibility near
to a vent issuing very hot basaltic lava.
The least understood and documented category is
the superimposed-trunked system. It is defined as a set of
lava tunnels superimposing and crossing each other,
all being active at the same time. The upper tunnels
stop their activity first, so that the lower ones carry on
for some time before they also stop operating and
become evacuated. There may even be connecting
openings between the levels exchanging lava between
crossovers. In such systems at least the lower ducts
must have been filled to the ceiling until very late in
their development. Such systems could arise when a
volcanic vent increases its output volume during an
ongoing eruption. Then the already established pyro-
ducts cannot accommodate the increased flow volume
and a new level of independently operating tunnels is
built on top of the already active ones. The Kipuka
Kanohina System on Hawai'i (Coons, 2009) is the larg¬
est example of such a superimposed-trunked system.
Sistima Tlacotenco (16 km long), a segmented system
of superimposed passages, most probably also belongs
to this category of conduits.
PRIMARY VOLCANIC ROCK CAVES
OF LIMITED EXTENT
Primary volcanic rock caves also occur in many differ¬
ent types: hollow imprints of trees and animals, partings
along the central plane of lava sheets, hollow tumuli.
ENCYCLOPEDIA OF CAVES
VOLCANIC ROCK CAVES
871
geological map
j km Puhia Pele and Huehue Flows t k
• = GPS Positions of Pukas f J
cave after D. & H. Medville,
S. Kempe, M. Oberwinder,
Hawai'i Speleological Survey
A
Puhia Pele Spattercones 1637f
Vent of Mystery Flow
Vent of Huehue Flow
to Kona 9.0 mi
FIGURE 4 Geological map of the Huehue Flow (Hualalai Volcano, Hawai'i) of 1801 according to surveys of the author's group. Flow was
from right to left, vertical distance 500 m. Oldest are the lavas (blue) of the Puhia Pele vent (a series of spectacular spatter cones) that was gas-
rich and formed an open channel system (with a few roofed-over sections). After its termination the Huehue Flow erupted (numbers label
pukas on the Huehue Cave, thick line; lava light yellow-green at left and at puka 1) accompanied by the Mystery Shield eruption that was
active for only a short time (upper right) and covered much of the upper part of the Huehue Flow (brown). Zoe's Puka is a tunnel belonging
to the Mystery Flow. From puka 17 a surface flow issued (light green) and the terminal lava from the mystery shield formed several short a'a
flows (dark red-brown at the right).
drained lava tongues, pressure ridge cavities, and volca¬
nic vents, and there may be more to be discovered.
Imprints
One of the most astounding caves in any respect is
the hollow imprint of a diceratherium in Miocene pil¬
low lavas (Rhino Cave, U.S. Quadrangle Park Lake,
Grant Country, Washington; C. Holler, pers. comm.).
Not only is it extremely rare that an animal gets
encased in lava, even more unlikely is the fact that the
cave is just now opened by erosion, so that it can be
entered through "the rear." Imprints of trees are, com¬
pared to those of animals, more common. A Mikado¬
like jumble of trees was encased in lava of Mount St.
Helens and is now publicly accessible to the joy of chil¬
dren who can easily crawl through the hollow imprints
from one tree to the other. A large, accessible tree
trunk is also encountered in Pa'auhau Civil Defense
Cave on Mauna Kea, Hawai'i (Kempe et al., 2003).
Often tree trunks are still standing, encased with lava
that cooled around them. When the lava flow is sub¬
siding around them they may be left standing, such as
in Lava Tree State Park, Hawai'i, which includes an
accessible pit-like imprint of the former tree.
Partings
In the process of cooling, gas exsolves from the
lava, forming small vesicles. The more time is avail¬
able, the larger they become. Cooling is fastest from
the surface downward and slower from the bottom of
the sheet upward; therefore the largest vesicles are
mostly found in the lower third of the sheet.
Sometimes they become dense enough to cause a
parting along which the upper section of the sheet
can be separated and bent upward by lateral pres¬
sure. These caves are low, but can be quite wide; they
are closed on all sides and only accessible if opened
in a road cut or by erosion.
Hollow Tumuli, Peripheral Lava Rise Caves,
and Drained Lobes
Tumulus is a morphological term describing a vari¬
ety of hummocks or small hills rising above the general
lava surface (Walker, 1991). Some appear to be pressed
up by lateral forces; others may result from lava being
injected from below under pressure, resulting in the
extrusion of lava from the tumulus. A few of these
tumuli are hollow, forming dome-like cavities. Some of
the largest occur in Kilauea Caldera in the Postal Rift
ENCYCLOPEDIA OF CAVES
872
v
Flow of 1919 ( e.g ., Tumulus El; Walker, 1991). Other
caves follow the perimeter of larger lava rises (e.g., lava
rise E5) that deflated in their centers once the lava
drained from them. These can be rich in rock spe-
leothems, specifically cylindrical stalactites. Other caves
in the same flow appear to be drained lava flow lobes
and lava tongues. All in all about 250 mostly shallow
caves have been recorded of various genetic origins in
the 1919 Caldera Flow (W.R. Halliday, pers. comm.).
Pressure Ridge Caves
This class of caves is much wider and longer than
tumuli caves. In Jordan we know of ten caves of this
type, all occurring in the lava field of the Qais/Makais
eruption. The longest of these caves is Al-Ameed, with
a total horizontal extent of 150 m. It actually consists of
two low, 30-m-wide (now centrally collapsed) and
15-m-wide chambers connected by a low passage.
These caves are not related to tumuli nor do they
show flow features suggestive of drainage. Rather they
seem to be associated with low ridges that are thought
to be created by the lateral compression of the upper
lava layers, already solidified, caused by the general
movement of the lower, still plastic layers, thereby
pressed upward and forming low, arched domes. On
Hawai'i, Eclipse Cave is of similar type, forming a 70-
m-long, up to 2.5-m-high hall, perpendicular to the
direction of the flow.
Volcanic Vents
Volcanic vents form pit-like or slanted caves, which
are potentially very deep. However, solidified lava and
wall collapse normally limit the accessible depth of such
caves. Kaukako Crater on Molokai, probably over 350 m
deep (100 m above and >250 m below water), is one of
the deepest open vents on record. Its diameter narrows
down to about 15 to 20 m, circa 30 m below the water
level. The lake is anaerobic below 4 m of depth and was
dived by M. and S. Garman to a depth of 140 m, possibly
one of the deepest cave dives in anoxic waters. Of similar
depth is the Na-One pit on Hualalai, a vent explored to a
depth of over 268 m. The current eruption on Kilauea, at
the wall of the Halemaumau Crater, has opened up a
40-m-wide and 160-m-deep pit, at the bottom of which
the top of the magma chamber is seen boiling and
through which gases, clasts, and ash are ejected. Many
more vents exist on Hualalai and Kilauea that have not
yet been explored due to the high risk of rockfall. On
Iceland, the brilinukagigur is 120 m deep (200 m total),
funneling out to a width of 49— 70 m below its orifice
(Stefansson, pers. comm.). On Terceira, Azores, the Algar
do Carvao, a 90-m-deep vent, has been made accessible
to the public. The vent leads into a large chamber hol¬
lowed out by convecting basalt magma in a body of tra¬
chyte. Into this class of caves we also must count caves in
hollow dikes that have been reported from several places.
ROCK SPELEOTHEMS
The term speleothem is composed of the Latin word
spelaeum (poetical for "cave" after the Greek root to
(nrrjXaiov; to spaelaion = "the cave") and the Greek
word 6 Grpdjv (ho thaemon = "the pile, deposit").
According to Hill and Forti (1997, p. 13) the term "refers
to the mode of occurrence of a mineral and not to its
composition." Thus, lava formations are excluded
because they are not composed of "a mineral" but of
"rock." Hill and Forti (1997, p. 217) admit that "there
are deposits in caves or in the outside world which,
while not speleothems in the strictest sense, neverthe¬
less mimic the forms taken by speleothems." However,
since the term speleothem is just indicating a "deposit
in a cave" it is logical to use the term rock speleothem
for those deposits in lava caves that are strikingly
homomorphous in appearance to their cousins, the min¬
eral speleothems (Kempe, 2011).
Among the most common forms of rock speleothems
in lava caves are stalactites and stalagmites, of which
several types can be differentiated. Allred and Allred
(1998) have investigated cylindrical stalactites. They
appear to be extrusions of the ceiling of residual melt. It
is not clear if they are growing like soda straws at their
tips or if they are extruded at the ceiling interface.
Below these stalactites often stalagmitic driplet spires
occur, composed of discrete drips of lava melt. In other
cases, where lava cascades into pukas, large curtains,
stalagmites, and columns can form. Rarely, spattering
inside a cave is observed; in Manu Nui (A. and P.
Bosted, pers. comm.) apparently water-bearing strata
were eroded into and steam explosions threw up spat¬
ter that was oxidized surficially, showing all colors
from black to tan.
CONCLUSIONS
In spite of the tremendous progress made in lava cave
exploration, we still are far from understanding all the
features and processes that interact to form caves in vol¬
canic rocks, specifically the large and extensive pyroduct
systems. It is clear that the concept of a "tube," simply
piping lava downhill, is far too simple to explain the
observed morphologies. Furthermore, many published
lava cave maps are of limited value because they are not
linked to a geological map; many of them lack morpho¬
logical details and cross-sections often do not show
ENCYCLOPEDIA OF CAVES
VOLCANIC ROCK CAVES
873
structural information of the lava flow itself. Thus, much
more process-oriented analysis is needed in order to
advance lava cave research.
See Also the Following Articles
Kazumura Cave, Haivaii
Root Communities in Lava Tubes
Bibliography
Allred, K., & Allred, C. (1997). Development and morphology of
Kazumura Cave, Hawai'i. Journal of Cave Karst Studies, 59(2),
67-80.
Allred, K., & Allred, C. (1998). Tubular lava stalactites and other
related segregations. Journal of Cave and Karst Studies, 59(2),
131-140.
Allred, K., Allred, C., & Richards, R. (1997). Kazumura Cave Atlas,
Island of Hawai'i. Hilo, HI: Hawaii Speleological Survey.
Atkinson, V., & Atkinson, A. (1995). Undara Volcano and its lava tubes:
A geological wonder of Australia in Undara Volcanic National Park,
North Queensland. Brisbane, Queensland, Australia (privately
published).
Coan, T. (1844). Letter of March 15, 1843, describing the Mauna Loa
eruption of 1843. Missionary Herald, 1844.
Coons, D. (2009). The Kipuka Kanohina Cave System, southwestern
Hawai'i. In A. Palmer & M. Palmer (Eds.), Caves and karst of the
USA (pp. 320—321). Huntsville, AL: National Speleological
Society.
Francis, P. (1993). Volcanoes, a planetary perspective. New York: Oxford
University Press.
Greeley, R., Fagents, S. A., Harris, R. S., Kadel, S. D., & Williams,
D. A. (1998). Erosion by flowing lava, field evidence. Journal of
Geophysical Research, 103{ Bll), 27325—27345.
Halliday, W. R. (Ed.), (1976). Proceedings 1st International Symposium
of Vidcanospeleology and its Extraterrestrial Applications, White
Salmon, WA., 16 Aug. 1972 Vancouver, WA: Western Speleological
Survey, ABC Publishing.
Hill, C., & Forti, P. (1997). Cave minerals of the world (2nd ed.).
Huntsville, AL: National Speleological Society.
Hon, K., Kauahikaua, J., Denlinger, R., & Mackay, K. (1994).
Emplacement and inflation of pahoehoe sheet flows: Observa¬
tions and measurements of active lava flows on Kilauea Volcano,
Hawai'i. Geological Society of America Bulletin, 106, 351—370.
Kempe, S. (2002). Lavarohren (Pyroducts) auf Hawai'i und ihre
Genese. In W. Rosendahl & A. Hoppe (Eds.), Angewandte
Geowissenschaften in Darmstadt, Schriftenreihe der deutschen
Geologischen Gesellschaft, 15, 109—127 (in German).
Kempe, S. (2011). Morphology of speleothems in primary (lava) and
secondary caves. In A. Frumkin (Ed.), Karst geomorphology.
Amsterdam: Elsevier.
Kempe, S., Bauer, I., & Henschel, H. V. (2003). The Pa'auhau Civil
Defense Cave on Mauna Kea, Hawai'i, a lava tube modified by
water erosion. Journal of Cave and Karst Studies, 65(1), 76—85.
Kempe, S., & Werner, M. S. (2003). The Kuka'iau Cave, Mauna Kea,
Hawai'i, created by water erosion, a new Hawaiian cave type.
Journal of Cave and Karst Studies, 65(1), 53—67.
Lockwood, J. P., & Hazlett, R. W. (2010). Volcanoes, a global perspective.
New York: John Wiley.
Olafsen, E. (1774—1775). Des Vice-Lavmands Eggert Olafsens und des
Landphysici Bianre Povelsens Reise durch Island, veranstaltet von der
Koniglichen Societal der Wissenschaften in Kopenliagen (from the
Danish 1st ed., 1772). Kopenhagen und Leipzig: Heinecke und
Faber (in German).
von Troil, U. (1779). Briefe welche eine von Herrn Dr. Uno von Troil im
Jalir 1772 nach Island angestellte Reise betreffen. Upsala und Leipzig:
Magnus Swederus (in German).
Walker, G. P. L. (1991). Structure, and origin by injection of lava
under surface crust, of tumuli, "lava rises," "lava rise pits," and
"lava-inflation clefts" in Hawaii. Bulletin of Volcanology, 53
(546-558).
ENCYCLOPEDIA OF CAVES
WAKULLA SPRING UNDERWATER
CAVE SYSTEM, FLORIDA
Barbara Anne am Ende
Deep Caves Consulting, South Riding, VA, L7SA
The source of the Waculla forms a large circular basin of great
depth, in which the water appears to be boiling up from a fathomless
abyss, as colorless as the air itself.
Charles Latrobe, 1833
INTRODUCTION
Wakulla Spring, located 30 km south of Tallahassee,
Florida, is a compelling source of freshwater. An
average of nearly 1 million cubic meters of water
discharges every day—enough to fill a thousand
Olympic-sized swimming pools! A multikilometer-
long cave system, with passages up to 30 m in diame¬
ter at an average depth of nearly 90 m, provides water
to the spring vent. However, during the Pleistocene,
sea level was lower and mastodons and other terres¬
trial animals found their way deep into the passage
that currently feeds the spring.
The earliest evidence of human visitation to
Wakulla Spring comes from Paleo-Indian artifacts
dated from 8500 to 12,000 years BP (Revel, 2002). The
first European explorer in the area was Panfilo de
Navarez, who in 1528 was searching for the province
of Apalachen with its purported gold and riches. He
and his Spanish crew ran out of provisions, diverting
to the village of Aute for resupply. The exact location
of Aute is not certain, though it was probably at or
near Wakulla Spring. A recent archaeological investi¬
gation at the spring found red, unglazed earthenware
olive jar fragments, suggesting Hernando de Soto
(retracing Navarez's route) visited Wakulla Spring in
1539. The beauty of this spring has caused much spec¬
ulation that it might have been the Fountain of Youth
sought by Ponce de Leon, though there is no evidence
that he visited Wakulla.
The spring's name is a European language bas¬
tardization of the Muskhogean Wacara (pronounced
"wakala") where cam meant "spring of water/' which
is similar to the Creek word zvahkola meaning "loon."
Over the years, the name has been published variously
as Wackhulla, Wakully, Waculla, Wahkula, and Wakulila. In
1843, the name Wakulla (pronounced "Wah-cuh-lah")
was officially given to the county. This large spring, and
the nearby smaller one, Sally Ward, are grouped
together and termed the Wakulla Springs.
Beginning in the early nineteenth century, American
settlers started moving into the Tallahassee area. The
Wakulla River, which suddenly appears in the midst
of a pine, live oak, and cypress forest, was a draw for
local tourists. Its remote location with poor roads
helped protect the natural environment. The trip from
Tallahassee through the pines was long and monoto¬
nous, prompting the rumor that "young people never
went to Wakulla Springs without coming back
engaged." Enterprising inhabitants provided boat
tours for visitors. Several published reports comment¬
ed on the vertigo induced in visitors who boated over
the spring basin because of the clarity of the water.
The county held an annual picnic on the site on the
first Saturday in May for many years. Several plans
and attempts were devised to commercialize the
spring given the success of other Florida springs such
as Weeki Wachee and Silver Springs, but none of these
plans materialized.
In 1934, Edward Ball, while working for his brother-
in-law, Alfred I. du Pont, began purchasing land
around Wakulla Springs. A year later, he began con¬
structing a hotel next to the spring basin. Though
rumored it was to become a private lodge, the hotel
was always a commercial operation. However, local
hunters were excluded from the land. A fence built
across the river denied access to the spring and caused
much contention. The county could no longer hold its
Encyclopedia of Caves.
875
© 2012 Elsevier Inc. All rights reserved.
876
w
annual picnic. Nevertheless, Ball resisted the notion of
expanding the commercial operation to a gaudy,
highly developed tourist trap. Over the years, a num¬
ber of movies were filmed at Wakulla including two
Tarzan movies. Creature from the Black Lagoon, and
Airport '77, in which a 70-foot model of a 747 was
perched on the ledge over the cave entrance in the
spring basin. Edward Ball died in 1981 and the land
went to the Nemours Foundation. It was realized that
Wakulla Springs would never make a profit, and the
operation was sold to Florida, in 1986 becoming the
Edward Ball Wakulla Springs State Park.
Today, the park continues its boat tours of the river
where wildlife such as alligators, turtles, deer, and squir¬
rels can be seen at close proximity. Birds abound, includ¬
ing anhingas, which are frequently seen perched,
holding their wings out to dry after diving for fish.
Glass-bottom boats circle the spring basin when the visi¬
bility of the water is good. Swimming from a white sand
beach is encouraged, though scuba diving is generally
not allowed except by special permit.
GEOLOGY
Geologic Framework
Wakulla Spring lies within the Woodville Karst
Plain, a low-lying area underlain by Oligocene- and
Miocene-aged limestone overlain by a thin blanket of
unconsolidated sands (Fig. 1). The Eocene Ocala
Group is a widespread unit that is an important part
of Florida's aquifer system. However, near Wakulla
Spring it lies well below all known cave passages; the
top of the Ocala was encountered between 120 and
180 m below the land surface in oil test wells near
Wakulla (Rupert, 1988). The Lower Oligocene Suwannee
Limestone unconformably overlies the Ocala Group.
The Suwannee Limestone is a calcarenite containing
miliolid foraminifera, mollusks, bryozoans, echinoids,
and corals. Chert is common and was used by Paleo-
Indians for tools and weapons. The passages in
Wakulla Cave are developed in the Suwannee
Limestone. The Lower Miocene St. Marks Formation
unconformably overlies the Suwannee and is the
uppermost bedrock unit at Wakulla Spring. The ledge
above the Wakulla Spring vent is St. Marks Formation.
The unit is primarily a calcilutite-containing quartz
sand, clay stringers, and mollusks. The lower contact
of the St. Marks Formation lies at an approximately
27-m water depth.
Samples collected during exploration in 1987, as
well as visual descriptions and video footage from
within Wakulla Cave, indicate a distinct lithologic and
color change within the Suwannee Limestone at a
NORTH Cody SOUTH
Scarp
FIGURE 1 Geologic cross-section from the Tallahassee Hills
through the Woodville Karst Plain. The approximate location of
Wakulla Spring has been projected onto the cross-section line. After
Rupert, 1993.
depth of about 65 m (Rupert, 1988). Above lies a soft
biocalcarenite, whereas below the contact is a less solu¬
ble, recrystallized dolomitic calcarenite. The dolomitic
rock may have retarded additional downward dissolu¬
tion in the conduits. The 3D map produced during the
1998—1999 diving project shows that the floor level for
much of the cave is uniform at a depth of about 90 m
(Fig. 2).
During the Pleistocene, the shoreline transgressed
across this area, reworking sands from older forma¬
tions and depositing the sediment over the limestone.
Five marine terraces are recognized in Wakulla
County, and Wakulla Spring lies within the Pamlico
Terrace, ranging from 3 to 8 m above sea level. The
Cody Scarp (Fig. 1) is an escarpment marking the
boundary between the northern Tallahassee Hills and
the Coastal Lowlands to the south. The boundary
represents the ancient shoreline location.
Paleontology
The first correct identification of the enormous
bones seen through the clear water on the bottom of
the Wakulla Spring basin was made by Sarah A. Smith
for the Tallahassee Floridian and Journal in 1850. This
publication prompted a local professor to collect a sig¬
nificant portion of a mastodon skeleton, but the bones
were lost in a shipwreck on their way to a museum on
the Atlantic coast.
Additional bones were collected during the next
decades. In 1930, more mastodon bones were discov¬
ered in shallow water during construction of a swim¬
ming area at Wakulla Spring (Rupert and Spencer,
1988). The U.S. Geological Survey was enlisted in its
collection and the skeleton remains on display today
ENCYCLOPEDIA OF CAVES
WAKULLA SPRING UNDERWATER CAVE SYSTEM, FLORIDA
877
FIGURE 2 Perspective view of the entrance area to Wakulla Spring Cave. Each colored dot represents a wall point surveyed with the digi¬
tal wall mapper. Different colors of dots are from different data files. Topographic contours of the ground surface are shown in yellow and
the boundary of the Wakulla River is shown in blue.
FIGURE 3 Mastodon skeleton in the Museum of Natural History,
Tallahassee, assembled with bones found at Wakulla Spring, Florida.
Photo by the author.
in the Museum of Natural History in Tallahassee
(Fig. 3).
Dives by Wally Jenkins, Garry Salsman, and their
buddies in 1955 and 1956 (see exploration section
below) resulted in the discovery of mastodons, mam¬
moths, deer, camels, giant ground sloths, and bears,
along with many spear points from Paleo-Indians. The
divers used pillowcases lined with plastic bags inflated
by air from their tanks to lift the heavy bones to the sur¬
face from depths as great as 60 m. During later explora¬
tion Pleistocene mammal bones were discovered as far
back as 366 m from the entrance.
Hydrology
Wakulla Spring is one of 33 first-magnitude springs
in Florida. Average discharge from 1907 to 1974 was
llm 3 s _1 (Rosenau et al., 1977). Wakulla Spring dis¬
plays the greatest range of discharge of any Florida
spring. A minimum flow of 0.7 m 3 s -1 was recorded
on June 18, 1931, whereas a maximum flow 54 m 3 s -1
was reported on April 11, 1973.
The presence of fossil mammal bones located
deep within the cave (Fig. 4), coupled with lower
Pleistocene sea level (and by extension, base level on
land), constrains hydrologic models. It has been sug¬
gested that mammals wandered into the dry cave
entrance looking for water. Further, it was hypothe¬
sized that the Wakulla Spring may have been a sink
where water entered the aquifer at the time. Indeed, a
report from another Florida spring noted the presence
of an extinct land tortoise with a wooden stake stuck
in its shell at 26 m below the current water level. A
less likely explanation published for the presence of
mastodon bones in Wakulla spring was that "In win¬
ter, mastodons crossing frozen pools broke through
the ice and drowned."
Directly north of the Wakulla area, an unconfined
surficial aquifer system is found in the unconsolidated
sands and gravels of the Tallahassee Hills. Recharge
occurs by direct precipitation. An intermediate aquifer
ENCYCLOPEDIA OF CAVES
878
w
FIGURE 4. Mastodon bone lying nearly 200 m inside Wakulla
Spring Cave at a water depth of about 80 m. Photo courtesy of Wes
Skiles.
system lies below the surficial sediments ranging from
about 15 to 46 m thick. The bedrock contains inter-
layered clayey sediments, limestone, and dolomite
resulting in discontinuous water-bearing zones. Water
recharges into the intermediate aquifer by leakage from
the surficial aquifer and from sinkhole-drained lakes. A
dramatic example of flow into the aquifer from a lake
in the Tallahassee Hills occurred on September 16,
1999. Much of the 1620-hectare Lake Jackson in the
Tallahassee Hills suddenly drained down Porter Hole,
a 5-m-deep sinkhole in the lake bed. The pit leading
down from the sink swallowed the lake's southern half
in a single day. Similar drainage events have occurred
in the past as well.
Below the surficial and intermediate aquifers lies the
Horidan aquifer system, which is a major carrier of water
and extends through much of the northern part of the
state. In the Woodville Karst Plain, the Horidan aquifer
is comprised of the St. Marks Formation, Suwannee
Limestone, and Ocala Group. Transmissivities are high
and range from 465 to 11,613 m 2 per day. Recharge
comes from downward leakage from the intermediate
aquifer in the Tallahassee Hills and through sinkholes.
The Woodville Karst Plain also forms a recharge area via
direct rainfall and through sinkholes. Four streams sink
underground, also contributing to recharge, though one
reemerges.
Regional groundwater flow of the Upper Floridan
aquifer is to the southeast across Wakulla County
(Fig. 5). In the Woodville Karst Plain the Floridan aqui¬
fer is unconfined and no low-permeability units lie
between the surface and carbonate aquifer units.
The primary source of Wakulla Springs was deter¬
mined, through the use of uranium isotopes, to be
southward-flowing Floridan aquifer water. Further,
strontium isotopic ratios were used to conclude that
FIGURE 5 Potentiometric surface map of the Upper Floridan
aquifer in Wakulla County. After Meadows, 1991.
the spring water has not come solely from local
recharge, but rather comes from water that has been in
contact with Floridan aquifer bedrock for considerable
time. Conversely, tritium age dating for shallow and
deep groundwater gave recharge times of less than 30
years.
The conduit flow of groundwater in the Wakulla
Spring cave and the Woodville Karst Plain is remark¬
ably complex considering the relatively uniform piezo¬
metric surface of the area (Fig. 5). A groundwater
divide has been reported 1 to 2 km inside the main
tunnel of Wakulla Spring. The divide marks a diver¬
gence between water flowing north to the Wakulla
Spring vent and water flowing approximately along
the regional groundwater gradient south, presumably
to the Spring Creek Springs Group of 13 submarine
springs. This flow divergence is perplexing because
one would not expect diffuse aquifer percolation to
supply water to a tunnel approximately 30 m in
diameter.
The Spring Creek Springs Group has pulsating
changes in flow where the surface of the water alter¬
nates between flat quiescence and boiling surface tur¬
bulence (Lane, 2001). The alternating surges generally
last for several minutes and are thought to be a result
of flushing through complex, tortuous passages. Since
the Spring Creek Springs Group is likely to be con¬
nected, at least indirectly, with the nearby Wakulla
Spring, much remains to be learned about the flow
within the Woodville Karst Plain.
The importance of understanding the sources of
water for Wakulla Spring is apparent when consider¬
ing the clarity of water discharging from the spring.
The vertigo-inducing clarity of the water in the nine¬
teenth and early twentieth centuries has diminished
during the past few decades. One of the first reports
of low visibility came in 1894: "The water has been
stirred up by the heavy rains, and we could only see
ENCYCLOPEDIA OF CAVES
WAKULLA SPRING UNDERWATER CAVE SYSTEM, FLORIDA
879
down 80 feet [24 m]," less than the maximum of 38 m.
Reports in 1945 and 1946 by the commercial operation
run by Ed Ball noted that visibility was affecting the
ability to run glass bottom boat tours and turn a profit.
The visibility is diminished primarily from tannic and
humic acids in surficial swamp or river waters that
enter the karst. Particulate matter may also contribute
to lower visibility, but the dark color of the tannic
water is a bigger problem. Boat guides at Wakulla
report that dark water has become a more frequent,
longer duration problem. Records of dark water indi¬
cated water visibility was poor for 58% of the time
during the past 12 years and that the poor visibility is
correlated to rainfall events. The best chance for crystal
clear water is during the dry months of May and June.
Another significant problem with water quality at
Wakulla Springs is the increase in nitrates, presumably
the result of runoff from fertilizers used on lawns and
agriculture. Simultaneous has been the introduction of
exotic algae, particularly hydrilla. A virtual explosion
of algae growth has choked much of the Wakulla
Spring basin and river in recent years, probably
enhanced by the high nitrate levels. The state park has
attempted to manage the hydrilla using mowers and
divers to remove as much as they can, but it has been
a losing battle. Recently, the park applied an herbicide
to the spring basin to improve visibility of the spring
by glass bottom boat tours and to regain a healthy eco-
logic balance in the river system. It is hoped that
native plants will reestablish themselves and that the
hydrilla will not be reintroduced.
Speleogenesis
Little work has been conducted on the speleogenesis
of Wakulla Spring cave system considering its magni¬
tude. The reason is probably that the cave system is
entirely flooded and averages a nearly 90-m water depth.
One study suggested that karstification began
approximately 9000 years BP based on the regional
structural geology, sea-level fluctuations, climatic
change, and groundwater flow of the Woodville Karst
Plain and its offshore extension. A model for the origin
of Wakulla Spring described the tunnels as a branch¬
ing, flow-dominated, saturated cave. A four-stage
sequence of development started through self-initia¬
tion. Small, random variations in permeability created
positive feedback loops. The upgradient process was
based on geochemical feedback. Downgradient, hydro-
dynamic processes governed the feedback loop and
was enhanced by corrosion from mixing of waters. The
largest conduits were developed in the cave when sea
level was lower during the Pleistocene. Discharge
occurred in springs that are presently submarine.
while recharge entered through a sinkhole that now is
the current Wakulla Spring.
EXPLORATION AND MAPPING OF
WA KULL A SPRING CAVE
A filming crew arrived at Wakulla Spring in 1955
for one of the many movies that needed a clear water
setting. The movie company hired Garry Salsman, a
Florida State University (FSU) student, to assist
because he owned an air compressor to fill diving
tanks (Burgess, 1999). He and a buddy, Wally Jenkins,
made their initial dive into the cave after completing
the first day's filming. On their second dive into the
cave, the two explorers discovered a mastodon long
bone some 60 m inside and carried it out of the cave.
The manager of the spring was so excited about the
find that he permitted Salsman and his FSU buddies to
continue diving in Wakulla.
Equipment at that time was quite primitive and
techniques for diving at such depths (55 m and head¬
ing down) were not well established. Early dives in
the 69° F water were made in nothing more than swim¬
ming trunks. Hardware-store flashlights wrapped in
plastic bags were used for lighting. The scuba tanks
did not have pressure gauges, so the students would
venture in on each dive just a little bit farther than the
previous one, to ensure an adequate air supply. The
team of divers diligently created safety plans and
avoided decompression sickness. However, by 1957
their diving ended, having reached nearly 300 m into
the cave at a water depth of 73 m. In part because of
limitations in technology, in part because of the stu¬
dents graduating and moving on, and because of lia¬
bility concerns by the springs manager, diving in
Wakulla Springs ended in 1961 for two decades. In
1981 three Florida cave divers used diver propulsion
vehicles and stage bottles to penetrate 335 m into the
cave at 79-m depth. Nitrogen narcosis was a significant
impairment factor, as it had been for the FSU students.
When Wakulla Springs was sold to the state of
Florida, cave divers hoped to obtain access to the
underground tunnels once again. In 1987 the U.S.
Deep Caving Team, Inc. (USDCT), received a permit to
dive in the cave for a 2-month period in the fall (Stone,
1989). State officials were particularly interested in
obtaining a map of the cave, as well as geologic,
hydrologic, and biologic sampling and studies. Bill
Stone, leader of the project, was in the process of
designing a mixed-gas rebreather, the MK-1, and
wanted the opportunity to test the device. During this
project, the rebreather was successfully tested to a
depth of 43 m and Stone was able to remain underwa¬
ter continuously for 24 hours. However, the rebreather
ENCYCLOPEDIA OF CAVES
880
w
was not used for missions into the cave. For this, the
divers used clusters of scuba tanks filled with mixed
gas attached to the diver propulsion vehicles, as well
as tanks on the divers' backs. Dives were made using
dry suits, and a specially designed habitat was built to
allow divers to decompress in a "bubble" of air. The
total surveyed length of the cave was 3310 m at a max¬
imum depth of 110 m. The culmination of diving was
due to time and technology constraints. The intention
was to return to explore the tunnels that beckoned
onward into the darkness.
A group of cave divers known as the Woodville
Karst Plain Project (WKPP) who had been diving in
various sinks and springs in the area began diving in
Wakulla Spring in 1991. During the next few years, the
team explored and surveyed extensively, culminating
in an impressive dive with a reported maximum pene¬
tration in Wakulla Spring of 5488 m using semiclosed
rebreathers. The most recent maps published by the
WKPP extend in maximum distance from the entrance
to about 3000 m.
In December 1998, the USDCT returned to Wakulla
Springs for the Wakulla 2 project (am Ende, 2000).
Stone's rebreather was no longer a prototype and the
MK-5 was commercially available. Diver propulsion
vehicles were powered by electric automobile batteries.
Decompression was conducted in a rented decompres¬
sion chamber deployed in the spring basin. The pur¬
pose of the expedition was not exploration (although
1.1 km of new passage was discovered), but rather
mapping. A special device, the digital wall mapper
(DWM) was designed and built for the project. Also
powered by electric auto batteries, the DWM kept
track of its location and measured distances to the
walls of the cave using sonar (Fig. 6). A total of about
10 million points on the walls of the cave were sur¬
veyed during the 3-month expedition. A total of 6.4 km
of tunnels were imaged this way (Fig. 7). The point
FIGURE 6 Divers entering the tunnel with DWM equipment.
Photo courtesy of Wes Skiles.
cloud generated during the project shows exquisite
detail of passage morphology that had not heretofore
been measured in caves. Current efforts are aimed at
meshing the point cloud into solid polygons (Fig. 8) to
render a more realistic cave.
FIGURE 7 Map of Wakulla Spring Cave showing 3 million wall
points. In B, C, and D Tunnels, segments of passage mapped by
hand are shown by lines.
FIGURE 8 Image of a meshed portion of the Wakulla point cloud.
The points are connected by cream-colored lines to form solid blue
triangles. This section of cave is at the junction of A and C Tunnels.
ENCYCLOPEDIA OF CAVES
WATER CHEMISTRY IN CAVES
881
In a separate effort, a 3D map of the Wakulla Spring
cave was built based on a traditional survey by the
WKPP (Kinkaid, 1999). Solid walls were created from
survey station data of four wall distance estimates
taken about 10 to 100 m apart (269 stations and 1076
estimated wall points). Although this map lacks the
detail of the USDCT map, the sparseness of data makes
it more easily viewable on the average personal com¬
puter. In July, 2007, Jarod Jablonski and Casey
McKinlay of the WKPP made the impressive connec¬
tion between the Leon Sinks underwater cave system
and passage they explored in Wakulla Spring. The
total cave system stands at 51 km in length and 107 m
in depth.
CONCLUSION
The end of Wakulla Spring Cave is nowhere in sight
and exploration will probably continue. However, even
with the best technical diving equipment available, the
current frontier is extremely remote. It may require yet
greater advances in diving technology to make signifi¬
cant additions to the exploration of Wakulla Spring.
Long periods of time with poor visibility of the
water also hampers exploration. The source of low-
quality water in terms of clarity and nitrates has not
been directly identified. It is probably from nonpoint
source pollution such as excessive runoff from cleared
and/or paved areas and residential and agricultural
contaminants. The state of Florida is currently taking
many positive steps to preserve and restore the quality
of water in its many springs. Time will tell if the state
can overcome the ecological pressure created by the
modern lifestyle of a large population in the area
around Wakulla Spring.
See Also the Following Article
Exploration of Caves—Underwater Exploration Techniques
Bibliography
am Ende, B. (2000). Wakulla 2—Building the first fully 3D cave map.
NSS News, 58, 244-260 (270).
Burgess, R. F. (1999). The cave divers. New York: Aqua Quest
Publications.
Kincaid, T. R. (1999). Morphologic and fractal characterization of satu¬
rated karstic caves. Laramie: University of Wyoming.
Lane, E. (2001). The Spring Creek Submarine Springs Group, Wakulla
County, Florida (Special Publication No. 4+). Tallahassee, FL:
Florida Geologic Survey.
Meadows, P. E. (1991). Potentiometric surface of the Upper Floridan aqui¬
fer system in the Northwest Florida Water Management District,
Florida, May, 1990 (U.S. Geological Survey Open File Report
90-586).
Revel, T. J. (2002). Watery eden: A history of Wakulla Springs.
Tallahassee, FL: Sentry Press.
Rosenau, J.C., Faulkern, G.L., Hendry, W.W., Jr. & Hull, R.W. (1977).
Springs of Florida. Tallahassee: Florida Geological Survey,
Bulletin 31 (revised), 461 p.
Rupert, F. R. (1988). The geology of Wakulla Springs (FGS Open File
Report No. 22). Tallahassee, FL: Florida Geological Survey.
Rupert, F. R. (1993). Karst features of northern Florida. In S. A. Kish
(Ed.), Geologic field studies of the coastal plain in Alabama, Georgia,
and Florida: Southeastern section (Meeting Guidebook, Tallahassee,
FL, April 1998, pp. 49-61).
Rupert, F., & Spencer, S. (1988). Geology of Wakulla County, Florida
(Bulletin No. 60). Tallahassee, FL: Florida Geological Survey.
Stone, W. C. (1989). The Wakulla Springs project. Derwood, MD: The
U.S. Deep Caving Team.
WATER CHEMISTRY IN CAVES
Janet S. Herman
University of Virginia
INTRODUCTION
Caves would not exist in limestone bedrock if it were
not for water. Interestingly, water is not only the agent
of formation of the natural openings in the Earth that
we call caves, but it is also the agent of cave decoration.
The duality of water's role in bedrock dissolution to
form the cave and in mineral precipitation to decorate
the cave can be understood through an examination of
the chemistry of cave waters. Just as any natural water
acquires its chemical composition, cave waters acquire
solutes from many different sources, including gases
from the atmosphere, weathering and dissolution of the
minerals of soils and bedrock, and human activities.
LIMESTONE DISSOLUTION
AND THE CHEMICAL EVOLUTION
OF CAVE WATERS
Calcite, CaCCU is the principal component of lime¬
stone. The solubility of calcite in natural waters is
established by chemical equilibria among the solid cal¬
cite, the dissolved calcium and carbonate-bearing spe¬
cies in aqueous solution, and the carbon dioxide in the
gas phase in contact with the solution. Indeed, the pH
of most natural waters is controlled by reactions
involving the carbonate system, calcium is the pre¬
dominant cation in most freshwaters, and most
groundwater in contact with limestone is near equilib¬
rium with respect to calcite solubility. These facts
allow us to understand the chemical evolution of cave
waters using the tools of chemical thermodynamics.
Encyclopedia of Caves.
© 2012 Elsevier Inc. All rights reserved.
882
w
TABLE 1 Chemical Composition for Water in Equilibrium with Respect to Both Calcite and CO 2 Gas at
Two Different PCO 2 Values and at Two Different Temperatures
Pco 2
Temperature (°C)
pH
Ca 2+ (mg L" 1 )
HCO; (mg L -1 )
l 0 -3.5
12
8.3
24
72
Iq-3.5
25
8.3
19
58
10-1-5
12
7.0
120
372
to- 1 - 5
25
7.0
96
298
These values for Ca 2+ and HC0 3 concentration were obtained in a hypothetical calculation of simultaneous equilibria.
Role of CO z Gas in Calcite Solubility
We can describe the dissolution of calcite in natural
waters by
CaC0 3 (s) + H 2 0 + C0 2 (g) <-► Ca 2+ + 2HC0 3 (1)
where the role of carbon dioxide is to generate acid.
Dissolved carbon dioxide is the most common acid in nat¬
ural waters, and the reactions that give rise to free acid
are the formation of carbonic acid from the hydration of
C0 2 gas:
C0 2 (g) + H 2 0 «-► H 2 CO° (2)
and the subsequent dissociation of carbonic acid
according to
H 2 CO^H + HCO 3 (3)
HCO 3 <-> H + + CO 2- (4)
A natural water in contact with a gas phase containing
C0 2/ therefore, has the dissolved carbonate-bearing spe¬
cies (i.e., H 2 C 03 , HCO 3 , and C0 2 ~) in solution regardless
of contact with a solid carbonate phase such as calcite.
The first fact of the chemistry of cave waters is there¬
fore evident. The extent to which C0 2 gas is dissolved
in infiltrating waters determines the extent to which
acid is available to carry out limestone dissolution to
form caves out of previously solid bedrock. Caves in
limestone bedrock in continental settings are typically
formed by the circulation of shallow groundwater.
These waters derive from the infiltration of dilute, fresh
rainfall into soil. Microbial respiration in the soil biode¬
grades natural organic matter (presented simply here
as CH 2 0 from dead vegetation) and generates C0 2 gas
according to
CH 2 0(s) + 02 (g)«- C0 2 (g) + H 2 0 (5)
Infiltrating water dissolves the C0 2 gas available in
the soil environment, thereby generating the acidity for
dissolution of limestone bedrock as the water continues
along its flow path.
The measure of the amount of C0 2 gas in the atmo¬
sphere is its partial pressure (Pco 2 ), where the total
atmospheric pressure is 1 bar. The value of the partial
pressure of C0 2 ranges from the low value of 10 -3 ' 5
bar in the open atmosphere at the Earth's surface, to
the higher values of 10 -2 2 to 10 _2 ° bar in the typical
soil atmosphere of a humid, temperate climate, to a
high value of 10 -15 bar in a notably C0 2 -rich cave.
This range in partial pressure of C0 2 gas is an impor¬
tant factor in the starting point for calcite dissolution.
A pure water in equilibrium with 10 -3 ' 5 bar C0 2 has a
pH of 5.7, whereas a water in equilibrium with 10 -1 ' 5
bar C0 2 has a pH of 4.7 at 25°C. The greater amount
of C0 2 generates a distinctly more acidic solution.
The greater the amount of C0 2 dissolved in the
groundwater, the greater the extent of dissolution of
calcite at equilibrium [see Eq. (1)]. The solubility of a
mineral is the mass of mineral that can dissolve in a
volume of aqueous solution. At equilibrium between
solution and mineral, no further dissolution occurs nor
does mineral precipitate from solution. An idealized
calculation of pure water in equilibrium with respect
to calcite at each of two different partial pressures of
C0 2 and two different temperatures illustrates the
impact these two variables have on solution composi¬
tion (Table 1). The availability of C0 2 gas is clearly an
important factor in determining the extent of calcite
dissolution. Temperature is the other. Like all gases,
C0 2 solubility is greater at lower temperatures than at
higher temperatures. The greater C0 2 gas solubility
that is possible at 12°C in groundwater compared to a
warm 25°C surface water further promotes calcite
dissolution.
Chemical Composition of Groundwater in
Limestone Terrain
The resulting chemical character of the groundwater
in caves is one rich in dissolved calcium (Ca 2+ ) and
bicarbonate (HC0 3 ). The pH of cave waters tends to
be relatively high because the protons released from
carbonic acid are consumed in the dissolution of cal¬
cite. Equation (1) demonstrates the molar relationships
of Ca 2+ and HCO 3 in this system: 2 mol HCO 3
are generated for every mole of Ca 2+ . The typical
ENCYCLOPEDIA OF CAVES
WATER CHEMISTRY IN CAVES 883
TABLE 2 Chemical Composition of Groundwater Collected in Limestone Terrain
Sample Location
pH
Ca 2+
Mg 2+
Na +
K +
HCOJ
so.U
cr
NO 3
State College, PA, U.S.A.
7.36
54.9
23
6
1.4
238
16.5
13.5
22
Mallorca, Spain
7.25
119
25
31
2.9
334
99
58
4.7
Warm River Cave, VA, U.S.A.
6.72
180
32.7
4.3
18.2
341
3.6
ND
Huntsville, AL, U.S.A.
7.5
48
3.6
2.1
ND
152
3.2
8.0
ND
Manatee County, FL, U.S.A.
7.59
93.3
48.9
47.0
8.1
159
352
47.0
ND
Yucatan Peninsula, Mexico
7.23
94
28
16
0.6
382
5.9
34
14
Some samples were collected in caves;
some samples were collected in wells drilled in limestone bedrock. All chemical concentrations are reported in units of
milligrams per liter. ND = not determined.
groundwater collected from limestone terrain, then,
has very large concentrations of dissolved HCO 3 com¬
pared to groundwater in other geological settings
(Table 2).
Unlike the hypothetical case of pure water in con¬
tact with pure calcite and a CC^-containing atmo¬
sphere ( e.g ., Table 1), Ca 2+ and HCOJ are not the only
dissolved chemicals in natural water. Limestone bed¬
rock, although predominantly made up of calcite, ordi¬
narily has some other minerals present. Rainwater
infiltrating through soil and groundwater migrating to
greater depths come in contact with these various solid
phases.
A mineral commonly associated with calcite is
dolomite, CaMg(C 0 3 ) 2 , whose dissolution is described by
CaMg(C0 3 ) 2 (s) + 2H 2 0 + 2C0 2 (g)^Ca 2+ + Mg 2+ + 4HC0 3
( 6 )
In addition to the common occurrence of dolomite
in limestone bedrock, some caves are developed in
dolostone, bedrock made up primarily of the mineral
dolomite. Dissolved Mg 2+ is, therefore, a major con¬
stituent in many cave waters ( e.g ., Table 2). Gypsum,
CaS0 4 ■ 2H 2 0, may be associated with calcite and dolo¬
mite in some sequences of sedimentary rock. If pres¬
ent, this highly soluble sulfate mineral can have a big
impact on the chemical composition of cave waters
CaS0 4 ■ 2H 2 0(s)<-+ Ca 2+ + SO 2- + 2H 2 0 (7)
For instance, in the sample from Manatee County,
Florida, U.S.A. (Table 2), the groundwater circulation
put it in contact with a gypsum layer at depth, and the
resultant groundwater had higher dissolved S0 4 -
concentrations than HCO 3 concentrations.
The other common rock-forming elements of Na +
and K + may enter solution from dissolution of miner¬
als or from ion-exchange reactions whereby the base
cations occupying the surface sites of clay minerals in
the limestone bedrock or in the sediments of the cave
stream can exchange by
Ca 2+ + 2Na + - clay 2Na + + Ca 2+ - clay ( 8 )
This mass action expression illustrates the shift in
the composition of the ion-exchange complex on the
clay mineral surface with shifting concentrations of
dissolved cations. When more dissolved Ca 2+ is pres¬
ent from calcite dissolution, it tends to shift toward
sorption onto the clay [be., Eq. ( 8 ) shifts to the right];
however, when more dissolved Na + is present from
salt water intrusion, Na + will move toward occupying
more surface sites [z'.e., Eq. ( 8 ) shifts to the left].
The anions commonly present in cave waters may
derive from any of several sources. Neither Cl - nor
NO 3 is a major rock-forming element. Both chemicals
have atmospheric sources, especially in regions where
the atmospheric chemistry is impacted by industrial
emissions, and, thus, they may enter with rainfall. Cl -
may be introduced to the groundwater of near-coastal
limestone terrain by the dissolution of sea spray or
marine aerosols into rainfall or by salt water intrusion in
limestone aquifers. Several samples in Table 2 have been
influenced by the proximity of marine waters (Mallorca,
Spain; Manatee County, Florida, U.S.A.; Yucatan
Peninsula, Mexico). Furthermore, both chemicals may
also come in as contaminants in waters influenced by
animal fecal wastes, road salts, and agricultural
chemicals.
Saturation State of the Aqueous Solution
The chemical status of a water capable of dissolving
calcite is termed undersaturated. That is, the solution is
undersaturated with respect to calcite and will dissolve
calcite until equilibrium solubility is obtained [i.e.,
Eq. (1) will shift toward the right]. An undersaturated
solution has less dissolved Ca 2+ and HCO 3 in it than it
would have at equilibrium. Such a solution is also
described as being aggressive toward calcite in that it is
capable of dissolving more calcite in the cave-forming
process. A water initially at equilibrium with respect to
ENCYCLOPEDIA OF CAVES
884
w
FIGURE 1 Solubility of calcite as a function of PCO 2 at two differ¬
ent common groundwater temperatures, 12° and 22° C, for 1 bar total
pressure. For a given PCO 2 value (in bars), water samples with total
dissolved calcium concentrations (in mmol kg -1 ) plotting below the
line are undersaturated with respect to calcite and capable of dissolv¬
ing calcite; solution compositions on the curves are in equilibrium
with respect to calcite and no further reaction can occur without a
change in environmental conditions; solution compositions plotting
above the line are supersaturated with respect to calcite and capable
of precipitating calcite. The retrograde solubility of calcite can also
be observed by noting that at a given PCO 2 value, the equilibrium
solution at 12°C is capable of holding more total dissolved calcium
(greater calcite solubility) than is the 22°C solution.
calcite can be shifted toward undersaturation by
changes in the two important factors discussed above:
Peer? and temperature. An increase in PCO2 or a decrease
in temperature will cause a water to again become
undersaturated with respect to calcite and to further the
dissolution of limestone bedrock. Conversely, an equi¬
librium solution can be shifted toward being supersatu¬
rated and capable of precipitating calcite [i.e., Eq. ( 1 ) will
shift toward the left] with a loss of CO2 gas or an
increase in temperature. A supersaturated solution has
more dissolved Ca 2+ and HCO3 in it than it would
have at equilibrium. An adjustment toward equilibrium
is achieved by precipitating calcite from solution.
A summary of the influences of PCO2 and tempera¬
ture is illustrated in a solubility diagram (Fig. 1 ).
Along each solid curve, the dissolved Ca 2+ concentra¬
tion and the pH are the equilibrium values for the
Pcc>2 and temperature indicated. A groundwater with
lower dissolved Ca 2+ concentration or with lower pH
than the values on the curve is undersaturated and
capable of calcite dissolution. A groundwater with
higher dissolved Ca 2+ concentration or with higher
pH than the values on the curve is supersaturated and
capable of calcite precipitation.
The saturation state of a solution is usually quanti¬
fied through the calculation of the saturation index.
The saturation index is the logarithm of the ratio of the
FIGURE 2 Plot of the log partial pressure of C0 2 (dashed line)
and saturation index with respect to calcite (solid line) as a function
of distance along the flow path for a cave stream. Warm River Cave,
Virginia. From the point of entry into the cave room, the stream
water is outgassing C0 2 from an initial high concentration in the
groundwater toward a lower value in equilibrium with the cave
atmosphere. The decreased PCO 2 thereby drives the solution from an
initial condition near equilibrium with respect to calcite toward
greater and greater supersaturation. Precipitation of calcite begins at
some greater distance (delayed time) from the point of water entry
into the cave. As calcite precipitation proceeds, the solution
approaches equilibrium with respect to calcite at a new, lower Pco 2
level.
actual dissolved Ca 2+ , HCO3, and H + concentrations
to those expected at equilibrium. The log of that ratio
is greater than 0 for a supersaturated solution, less
than 0 for an undersaturated solution, and equal to 0
at equilibrium.
The process of CO2 gas exchange in cave waters
often controls the saturation state of the solution.
Groundwater rich in dissolved CO2 picked up during
infiltration through an organic-rich soil and circulation
in the confined subsurface will have a high PCO2.
When such a groundwater enters a cave, the water
begins to equilibrate with the cave atmosphere that is
usually less rich in CO2 by virtue of circulation and
exchange with the open atmosphere on the Earth's sur¬
face. The outgassing of CO2 from a water entering the
cave can drive the pH up and the saturation state of
the water from being near equilibrium with respect to
calcite to being significantly supersaturated with
respect to calcite. In essence, the loss of C 0 2 to the
atmosphere acts to drive Equation ( 1 ) back to the left. If
calcite does not immediately precipitate, the result is a
supersaturated solution capable of precipitating calcite.
A longitudinal sampling of a cave stream in Virginia
showed just such a trend in saturation state of the solu¬
tion along the flow path (Fig. 2 ). Until the solution can
reach equilibrium by precipitating calcite, it remains
supersaturated with saturation indices significantly
greater than 0.
ENCYCLOPEDIA OF CAVES
WATER CHEMISTRY IN CAVES
885
Environmental Factors Influencing Calcite
Dissolution
The critical role of CO2 gas in shifting the chemical
equilibria toward more calcite dissolution in the pres¬
ence of more CO2 gas or toward less calcite dissolution
in the presence of less CO2 gas is the basis for under¬
standing some of the climatic effects on cave develop¬
ment. In a moist, warm environment, the growth and
decay of terrestrial vegetation occurs to greater extents
and at greater rates than in cool, dry environments.
The overall result is greater availability of CO2 gas to
generate carbonic acid in a warm, organic-rich envi¬
ronment. Cave development in tropical settings, there¬
fore, is recognized to be far greater than in arctic or
desert settings. Some of the most dramatic caves we
know from locations such as Puerto Rico, Mexico, and
Belize developed under tropical climatic conditions.
There are other hydrogeological and geochemical
settings that generate waters undersaturated with
respect to calcite and capable of dissolving limestone
to form caves. The mixing of freshwater and saline
water in coastal regions gives rise to an aggressive
solution that can dissolve more calcite than either indi¬
vidual type of water can. The chemical characteristics
of groundwater in coastal regions are those of a mix¬
ture of two solutions: (1) a dilute, fresh groundwater
such as might be generated by the infiltration of rain
into a soil and into an underlying limestone where cal¬
cite dissolution occurs, and (2) seawater that is intrud¬
ing into the limestone aquifer. The dense saline water
that encroaches into coastal aquifers forms the down-
gradient boundary of the fresh groundwater circula¬
tion system where discharging freshwater mixes with
intruding seawater to form a zone of brackish
groundwater.
The chemical reactivity of the brackish-water mix¬
ture results from the nonlinearity of mineral solubility
as a function of the relative concentrations of dissolved
constituents in the mixture. The great contrasts in sol¬
ute concentrations between the mixing end members
make it possible to generate a mixture that is capable
of further dissolution of calcite from two solutions that
were individually in equilibrium, or even supersatu¬
rated, with respect to calcite (Fig. 3 ). This renewed
potential for calcite dissolution in the coastal mixing
zone has been recognized as an important contribution
to the formation of caves in the near-coastal setting.
Caves developed along the coasts of Florida, Mexico,
Spain, Greece, and the Bahamas have been influenced
by calcite dissolution in the freshwater—saline water
mixing zone. The chemical composition of water in the
resulting coastal caves once again has Ca 2+ and HCO3
as major constituents, but the overall character of the
water could be described as diluted seawater with
FIGURE 3 Saturation state of a solution derived from mixing
fresh and saline waters. The mixing curve was generated for a fresh
groundwater in equilibrium with respect to calcite at Pco 2 = 1CT 22
bar and 22°C. The freshwater was mixed with a standard composi¬
tion of surface seawater for which Pco 2 = 1CT 3 ' 3 bar. The mixture
is undersaturated where the mixing curve plots below value
SIcaicite = 0; the mixture is supersaturated where it plots above the
value SIcaicite = 0. Starting with a saturated freshwater and a super¬
saturated saline water, mixing produces an undersaturated brackish
water which is capable of calcite dissolution that contributes to the
formation of caves in coastal limestone.
sodium (Na + ), chloride (Cl - ), and sulfate (SO4 ) con¬
centrations also being high; however, the source of
these solutes is not in the local bedrock.
Role of Strong Acids in Limestone Dissolution
Although the abundant natural acid in hydrogeologi¬
cal systems is carbonic acid, there are settings in which
other natural acids are important. The migration of
hydrogen sulfide into shallow, oxygenated groundwater
environments from greater depths where reducing con¬
ditions prevailed sets up the potential for the generation
of strong acid in the subsurface environment. For
instance, caves of the Guadalupe Mountains of New
Mexico have been formed by sulfuric acid derived from
the oxidation of hydrogen sulfide migrating from
underlying oil and natural gas reservoirs. As dissolved
H2S gas comes in contact with dissolved O2, oxidation
generates acid in the groundwater:
H 2 S(g) + 20 2 (g)«- 2 H + + SO 2 ( 9 )
Once acid is present, the dissolution of calcite looks
just as it did for the carbonic acid source:
CaC 0 3 (s) + H + <-► Ca 2+ + HCO3 ( 10 )
but the chemical composition of the cave water has a
high concentration of SO4 that is not seen in the more
typical solution caves.
ENCYCLOPEDIA OF CAVES
886
w
The oxidation of other sulfide phases may also gener¬
ate sulfuric acid for limestone dissolution. Iron sulfide
minerals, mainly pyrite (FeS 2 ), and other base-metal sul¬
fide minerals are common in the rocks of the Earth's
crust. Although stable under reducing conditions, with
exposure to 0 2 -containing air and water, the sulfide
minerals undergo oxidation as in this reaction with
pyrite:
FeS 2 (s) + 02(g) + H 2 O^Fe 2 + 2H++ 2SO^ (11)
Just as in the oxidation of hydrogen sulfide gas,
acidity is released to solution that can further react
with limestone bedrock according to Equation (10).
Although natural weathering of bedrock containing
sulfide minerals does occur, the acceleration of this
reaction due to the human activity of mining and pro¬
cessing minerals far outstrips the natural occurrence of
acidity from such reactions. Today, the contact of acid
mine drainage with limestone bedrock in active or
abandoned mining districts may be contributing to the
formation of new conduits and caves in the regional
bedrock.
One environment fosters both reduction and oxida¬
tion reactions involving sulfur. In the saline waters of
coastal caves and blue holes, density stratification of
the groundwater in the cave or aquifer may limit the
mixing of shallow oxygenated waters to greater depth.
At depth, modest concentrations of dissolved and sus¬
pended organic matter can result in reduction of SO4
as the organic matter undergoes microbially mediated
oxidation in the absence of dissolved 0 2 . The use of
SO4 as an alternate electron acceptor results in the
formation of dissolved sulfide:
SOt + 2CH 2 0 <-► H 2 S(g) + 2HCO3 (12)
where the organic matter, CH 2 0, may be dissolved or
particulate. Further migration of H 2 S along a ground-
water flow path may bring the water into contact with
0 2 where the H 2 S oxidizes to sulfate, generating acid¬
ity [Eq. (9)]. The increased capacity to dissolve lime¬
stone has contributed to the formation of caves and
blue holes on the islands of the Caribbean.
PRECIPITATION OF MINERALS AND THE
FORMATION OF SPELEOTHEMS FROM
CAVE WATERS
Not all cave waters derive their final composition
from the process of calcite dissolution. Just as the
increased abundance of C0 2 gas was a factor in shift¬
ing the extent of calcite dissolution in undersaturated
subsurface waters, the same factor in reverse can shift
cave water to being supersaturated and capable of
mineral precipitation. Carbonate minerals precipitate
as a result of Eq. (1) shifting toward the left due to
C0 2 outgassing when soil water or groundwater with
elevated Pco 2 enters the open room of a cave.
Although C0 2 levels were increased in the restricted
environment of the soils, when soil water or ground-
water with high Pco 2 comes in contact with a cave
atmosphere with lower Pco 2 , sudden outgassing can
occur. Precipitation of calcite from drip waters forms
dripstone and flowstone on cave ceilings, walls, or
floors, and precipitation from stream waters forms
travertine. In each case, the resulting chemical compo¬
sition of the cave water is somewhat reduced in dis¬
solved Ca 2+ and FICO3 concentrations and the pFI of
the solution is slightly lower; that is, the solution com¬
position shifts down to the left on Figure 1 as the
supersaturated solution (saturation index greater than
0) establishes equilibrium (saturation index equal to 0).
Although C0 2 outgassing can drive the cave water
to high levels of supersaturation from which calcite
precipitation ensues, the rate of calcite precipitation
may be slow enough to maintain a supersaturated
solution for some time. In the example from a cave
stream in Virginia, for instance, the solution reaches a
high degree of supersaturation before beginning to
adjust toward equilibrium through calcite precipitation
as the stream moves through the cave (Fig. 2). The
principles of thermodynamics are useful in determin¬
ing the ultimate equilibrium composition of the
calcite—water—C0 2 system, but the rates at which the
reactions occur are less known and fall under the pur¬
view of geochemical kinetics.
All cave water is not confined to a stream channel
in which travertine or rimstone dams may be the prod¬
uct of calcite precipitation. Diffuse infiltration waters
commonly emerge from the ceiling or walls of caves
where rapid outgassing of C0 2 from small water
droplets with large surface area drives the solution to
become supersaturated. Generally, the precipitated
mineral is calcite, although in some instances the poly¬
morph aragonite is formed. Being formed from a solu¬
tion with a variety of dissolved chemicals present,
some trace incorporation of cations other than Ca 2 +
contributes to some chemical impurity of the carbonate
speleothems. The chemical and morphological variety
of the precipitated carbonates makes the study of cave
speleothems an endlessly fascinating endeavor.
In addition to the transport and reaction of Ca 2+ and
HC0 3 in the cave environment, chemical reactions in
the soil water and groundwater environment can mobi¬
lize iron and manganese. Ordinarily found in their oxi¬
dized form in insoluble oxide minerals, iron and
manganese can be reduced in much the same way that
0 2 is reduced coupled to the oxidation of organic matter.
In the water-saturated subsurface, when biodegradation
ENCYCLOPEDIA OF CAVES
WATER TRACING IN KARST AQUIFERS
887
of organic matter consumes all of the available 0 2 , bacte¬
ria will turn to the use of alternate electron acceptors.
Oxidized iron (Fe(III)) and manganese (Mn(IV)) may be
present in the form of goethite (FeOOH) and birnessite
(Mn02) as grain coatings in the soil. Here shown for bir¬
nessite, an oxidized manganese mineral is reduced when
coupled to microbially mediated oxidation of organic
matter to C0 2 :
2Mn0 2 (s) + CH 2 0 + 4H+ ^2Mn 21 + 3H 2 0 + C0 2 (g)
(13)
When these elements are used as electron acceptors,
they are reduced to Fe(II) and Mn(II), respectively, and
are then soluble in groundwater. Dissolved Fe 2+ and
Mn 2+ enter the cave in flowing groundwater where
they may remain dissolved in cave waters or they may
be reoxidized to precipitate Fe(III) and Mn(IV) oxide
and carbonate minerals. The presence of oxidized iron
and manganese minerals can be recognized as coatings
on cave walls or on speleothems as black or reddish
brown coatings. For instance, birnessite can be a black,
dendritic deposit on the surface of calcite speleothems.
HUMAN IMPACT ON CAVE WATERS
Just as the activities of crop and animal farming,
industrial production, and fecal, municipal, and haz¬
ardous waste disposal have had deleterious effects on
the quality of surface water and groundwater in many
regions of the world, some cave waters have been neg¬
atively impacted. The variety of chemical contaminants
is too numerous to elucidate in detail, but everything
from fertilizers, pesticides, metals, bacteria, petroleum
hydrocarbons, and chlorinated solvents to exotic syn¬
thetic organic compounds and pharmaceuticals have
been reported to contaminate cave waters. In these
situations, cave waters are not unique but rather are
like other groundwater. Human activities, sometimes
immediate and sometimes at remote times and places,
can impact the chemical composition of groundwater
along its flow path.
There are, however, some dramatic examples in
which the physical characteristics of limestone aquifers
have distinguished the resulting groundwater contami¬
nation from that typical in other types of bedrock. The
large conduits and caves in the subsurface of karst ter¬
rains are openings many orders of magnitude larger
than are present in the subsurface in other geological
settings. These relatively huge openings lend them¬
selves to rapid transmission of high concentrations of
contaminants and allow the migration of large parti¬
cles of contaminants without the benefit of filtration by
the porous media of the subsurface environment.
Some examples of the dramatic contamination of cave
waters include the presence of free-phase gasoline
floating on cave streams and the fecal colliform bacte¬
rial contamination of rural drinking water supplies.
Once introduced to groundwater, the rapid flow
through limestone aquifers makes it particularly diffi¬
cult to limit the spread of contamination. The hard les¬
sons learned about the contamination of water
supplies in karst regions drive home the need to avoid
contamination of groundwater in the first place rather
than to try to remediate contamination after it occurs.
Bibliography
Back, W., Hanshaw, B. B., Pyle, T. E., Plummer, L. N., & Weidie,
A. E. (1979). Geochemical significance of groundwater discharge
and carbonate solution to the formation of Caleta Xel Ha,
Quintana Roo, Mexico. Water Resources Research, 15, 1521—1535.
Butler, J. N. (1982). Carbon dioxide equilibria and their applications.
Reading, MA: Addison-Wesley.
Drever, J. I. (1997). The geochemistry of natural waters (3rd ed.). Upper
Saddle River, NJ: Prentice Hall.
Garrels, R. M., & Christ, C. L. (1965). Solution, minerals, and equilibria.
San Francisco: Freeman, Cooper & Co.
Hem, J. D. (1989). Study and interpretation of the chemical characteristics
of natural water. U.S. Geological Survey.
Krauskopf, K. B., & Bird, D. K. (1995). Introduction to geochemistry
(3rd ed.). New York: McGraw-Hill.
Langmuir, D. (1971). The geochemistry of some carbonate ground-
waters in central Pennsylvania. Geochimica et Cosmochimica Acta,
35, 1023-1045.
White, W. B. (1988). Geomorphology and hydrology of karst terrains.
New York: Oxford University Press.
White, W. B., Culver, D. C., Herman, J. S., Kane, T. C., & Mylroie,
J. E. (1995). Karst lands. American Scientist, 83, 450—459.
Wigley, T. M. L., & Plummer, L. N. (1976). Mixing of carbonate
waters. Geochimica et Cosmochimica Acta, 40, 989—995.
WATER TRACING IN
KARST AQUIFERS
William K. Jones
Karst Waters Institute, Leesburg, VA, USA
DEFINITIONS AND OBJECTIVES OF
TRACER TESTSJN KARST
Water tracer tests are usually conducted to establish
the hydrologic connections between two or more
points. The tracer is an identifiable label or marker
added to flowing water that establishes the links
between the injection point of the tracer and the moni¬
toring points where the tracer reappears. Fluorescent
dyes are the most commonly used tracers in karst
aquifers, but a wide range of substances has been used
successfully. The experimental design of a tracer test
Encyclopedia of Caves.
© 2012 Elsevier Inc. All rights reserved.
888
w
may be qualitative to simply establish if a hydrologic
connection exists between two points, or quantitative to
measure the time-concentration series (breakthrough
curve) generated by the recovery of the tracer. Water
tracer tests usually work well in karst areas because of
the fast groundwater flow rates and the prevalence of
flow paths restricted to discrete conduits.
The level of effort and the cost of conducting a
tracer test are functions of the question being asked.
Many tracer tests are an extension of cave exploration
and are conducted to "see where the water goes." The
tracer test may have the relatively simple objective of
establishing a connection between two cave passages
separated by an impassable reach. Tracer tests may be
used to determine the destination of water flowing
into a sinkhole or blind valley. The karst drainage
basin contributing water to a spring or resurgence may
be estimated based on the results of tracer experi¬
ments. The interbasin transfer of water is frequently
demonstrated by tracer tests. The travel rate of a tracer
may be used to establish some of the hydrologic prop¬
erties of an aquifer. Examination of the breakthrough
curve of a tracer test can aid in the interpretations of
some of the internal flow characteristics of the aquifer.
The movement of a tracer may mimic the subsurface
behavior of a pollutant introduced into the aquifer at
the same point. Tracers may be used to identify the
source of pollution at a spring. Tracer tests are some¬
times used to demonstrate the "vulnerability" of karst
aquifers to chemical spills and pollution associated
with inappropriate land use practices and to establish
"wellhead protection areas" for public water supplies.
A study of the Unica River basin in Slovenia
(Gabrovsek et ai, 2009) provided a demonstration
of the range of hydrologic information that can be
obtained from tracer tests and the applicability to the
protection and management of karst water resources.
Naturally occurring chemicals or isotopes may be
used to determine the age or residence time of water
at sampling points within the aquifer. The sources of
the water in terms of percentages contributed by con¬
duit flow versus drainage from the overlying epikarst
or deeper fracture zones may be studied. These chemi¬
cal markers are considered tracers, but they identify
the various components or storage areas of the aquifer
rather than establish direct hydrologic connections or
paths between specific points in the aquifer.
HISTORY OF WATER TRACING
IN KARST AQUIFERS
A considerable amount of folklore concerning under¬
ground water connections has accumulated throughout
the world's cave areas. Tests have been reported using
tracers such as wheat chaff, duck feathers, marked logs,
tagged eels, and muddy water from storm events. A
2000-year-old tracer test is attributed to tetrarch
Philippus who supposedly established the source of the
springhead of the Jordan River in 10 A.D. Chaff was
thrown into Phialo Pond (Berekhat Ram Crater Lake)
and reemerged from the cave spring Panium, the head
of the Jordan River. Although the results of this test
have been questioned, it does establish that the basic
idea of tracing underground water is quite old.
A scientifically planned tracer test was conducted in
1872 by A. Hagler to determine the origin of a typhoid
fever outbreak in the village of Lausen in Switzerland.
About 800 kg of salt (NaCl) was injected in a sinking
stream on a farm south of the village and the water
supply spring for the village showed a strong reaction
for chloride the following day (Kass, 1998).
The start of modern water tracing studies followed
the discovery of fluorescein by A.V. Baeyer in 1871.
The sodium salt of fluorescein, called uranine in
Europe, rapidly became the most used and probably
the most successful groundwater tracer for cave and
karst studies to this day (Fig. 1). The first reported use
of sodium fluorescein as a tracer was an experiment
conducted in southwestern Germany on October 9,
1877, by Professor Albert Knop. Ten kg of sodium
fluorescein supplied by C. Ten Brink was injected at
infiltration points in the bed of the Danube River and
reappeared two days later and 12 km away at the
Aach Spring, a tributary of the Rhine River. A test con¬
ducted a couple of weeks earlier from the same loca¬
tion used almost 10,000 kg of salt and may have been
the first quantitative tracer test with hourly sampling
at the Aach Spring.
FIGURE 1 Fluorescein sodium powder is red but turns bright
fluorescent green when dissolved in water. Here the dye is being
added to an underground stream just upstream of an impassible
sump. Photo by Philip C. Lucas. Used with permission.
ENCYCLOPEDIA OF CAVES
WATER TRACING IN KARST AQUIFERS
889
Work on the development of methods for tracing
underground waters continued in Europe. Many of the
early tracer tests used large quantities of dye, because
detection depended on visual coloring of the water at
the resurgence. An interesting example of an acciden¬
tal tracer experiment occurred in the French Franche-
Comte in 1901. As the result of a fire at the Pernod dis¬
tillery at Pontarlier, a quantity of absinthe poured into
the Doubs River. A significant amount of the alcohol
seeped through the bed of the river and reemerged
the following day 10 km away at the source spring of
the Loue River. The test of the connection between the
Doubs and the Loue was repeated by E.A. Martel in
1910 using 100 kg of sodium fluorescein. The Loue
was brightly colored two days later for a distance of
100 km downstream. The source of the Garonne River
in southern France was determined by a tracer test in
1931 by Norbert Casteret. He injected 60 kg of sodium
fluorescein at the Trou du Toro on the Spanish side of
the Pyrenees and it emerged 3.7 km away and 10
hours later at the Goueil de Joueou on the French side.
The Garonne River was colored a bright green for over
50 km downstream. Quantitative tests at the same site
in the early 1990s used less than 2 kg of sodium fluo¬
rescein and yielded travel times for the tracers of about
11 hours during high-flow conditions in the summer
and 180 hours for a low-flow test in the winter (Freixes
et al., 1997).
Various tracer techniques were tried in Europe
through the 1920s. A "fluoroscope" was developed
which allowed a rough estimation of dye concentra¬
tions in water samples by visual comparison with pre¬
pared laboratory standards. A number of different
salts, dyes, and even radioactive compounds were
tried. Bacteria were tried as a particulate drift tracer as
early as 1896. Spores from the club moss Lycopodium
calvatum were suggested as a possible tracer in 1910,
and the first reported tracer test using spores is from
Europe in 1940 (Kass, 1998). Spores became a popular
tracer in Europe in the 1950s because they could be
dyed different colors and several sinkpoints could be
traced simultaneously. Plankton nets were suspended
at the springs to passively collect the spores, so con¬
stant surveillance of a number of springs was no lon¬
ger required. Spores dyed with fluorescent dyes
proved to be easier to identify under a microscope.
Drifting tracers such as spores and bacteria are still
used to a certain extent in Europe but have never been
popular in North America.
One novel tracing experiment in Slovenia in 1929
used marked eels. The dorsal fins of 494 eels were
notched and the eels were released in the Reka River
that sinks into Skocjanske Cave and resurges 34 km
to the west at the Timavo Springs in Trieste Bay
(Italy). Twenty-nine of the eels were caught in eel pots
at the Timavo Springs during the one-year observation
period.
The Yorkshire Geological Society carried out a series
of tracer tests in the Ingleborough area of Great Britain in
1904. One principle from this work has been to test "key
sinks" and to guess the remainder in lieu of testing every¬
thing. This approach is still practiced in many karst areas,
but more intensive studies at later dates sometimes pro¬
duce surprising results. Much of the subsequent work on
tracing underground drainage in the British Isles has
been conducted by caving groups associated with various
universities. A study of the caves of northwest County
Clare, Ireland, in the early 1960s by E.K. Tratman and the
University of Bristol Speleological Society, involved a
number of tracer tests using sodium fluorescein and pas¬
sive carbon detectors to monitor the resurgences.
Just as professional cave and karst studies in North
America lagged behind Europe in general, so did the
interest in tracing underground water. The first descrip¬
tion of tracer tests from North America was a U.S.
Geological Survey water supply paper by R.B. Dole in
1906. This paper discussed the use of sodium fluorescein
as a tracer, but few tracer tests are reported from North
America before the late 1950s. Whereas most of the water
tracing studies in Europe were conducted by professional
hydrologists or speleologists, often at considerable
expense, the real beginning of karst water tracing studies
in North America was the contribution of unpaid cavers
as an extension of cave exploration. A brief description of
a method to recover sodium fluorescein sorbed onto acti¬
vated coconut charcoal granules was published by J.R.
Dunn in a cave club newsletter in 1957. This discovery
offered an inexpensive technique to monitor different
resurgences without the necessity of constant surveillance.
The researchers could place carbon packets (called Dunn
Bugs) in the springs, inject the dye in a sinking stream or
cave, and return in a few weeks to collect and test the car¬
bon packets. The concentration of the dye on the carbon
increases with the exposure time to the dye, so recovery
concentrations of sodium fluorescein below the normal
visual threshold could be detected. The testing relied on
visual identification of the dye as a fluorescent yellow-
green sheen floating on top of the carbon in an elutant.
Apparently a very similar technique was described by
Mayrhofer in 1904 (Kass, 1998), but the activated carbon
system was used in Europe more as a technique for
enriching the dye concentration in a sample than as an
unattended monitoring system. The first field test of the
passive detector system in North America was described
by T.D. Turner in 1958. Turner reported a tracer test using
carbon detectors and 4 g of sodium fluorescein over a dis¬
tance of 4.8 km near Pine Grove Mills, Pennsylvania.
Water tracing in North American cave studies really
began in the early 1960s using sodium fluorescein and
passive detectors of activated carbon. Hermine Zotter
ENCYCLOPEDIA OF CAVES
890
w
began a dye tracer study in Pocahontas County in east-
central West Virginia in 1960. Thomas Aley conducted
a tracer test for a court case in California in 1963. This
was probably the first use of underground water tracing
to resolve litigation in the United States. Aley worked
on a study for the U.S. Forest Service in the Ozarks in
Missouri in 1966 and conducted a 64-km-long trace
from the Eleven Point River Basin to Big Spring on
the Current River Basin. William Jones conducted a
study for the U.S. Geological Survey in Greenbrier
County, West Virginia, in 1966. The results of the trac¬
ing tests were coupled with the positions of surveyed
cave streams and topographic divides to delineate karst
drainage basins. Charles Brown was conducting tracer
tests in the Maligne basin in Alberta, Canada, by 1969.
James Quinlan became the park geologist for Mammoth
Cave National Park in Kentucky in 1973 and started
an intensive series of water tracing tests. The results
of the tracing tests were combined with data from
cave surveys and potentiometric contours to produce a
detailed map of drainage basins and subbasins around
Mammoth Cave (Quinlan and Ewers, 1989).
The development in the early 1960s of fluorometers
that control or measure light emissions in both the
excitation and emission wavelengths of different fluo¬
rescent compounds enabled dye analytical methods to
become much more quantitative. Quantitative water
tracing using fluorometers dates from the early 1960s.
Much of the early quantitative tracer work was for sur¬
face water time-of-travel studies. Fluorometers also
allowed several different fluorescent compounds to be
used simultaneously because the instruments could be
set to narrow specific wavelengths.
Much of the development of quantitative tracer techni¬
ques was presented in a series of conferences in Europe
beginning with the Specialists Conference on Tracing of
Subterranean Waters in Graz, Austria, in 1966. This con¬
ference series has continued at about four-year intervals
as the Symposium on Underground Water Tracing
(SUWT) conferences. The published proceedings from the
series contain a wealth of information on various tracer
techniques for both karst and porous media water trac¬
ing. Much of the material from these conferences is dis¬
tilled in the book by Kass (1998). A U.S. Geological
Survey paper by Wilson (1968) and a paper on the evalu¬
ation of fluorescent tracers by Smart and Laidlaw (1977)
were probably the most influential papers on quantitative
water tracing published in North America.
TRACERS
Groundwater tracers are generally classified as:
(1) naturally occurring (often accidental in their addition
to the water); (2) artificial tracers deliberately introduced
into the aquifer; and (3) pulses. Naturally occurring tra¬
cers include the chemical constituents and isotopes natu¬
rally found in water or present due to human activities.
Tritium, fluorocarbons, pollutants from waste sites, and
thermal waters may all serve as tracers in some circum¬
stances. Isolated thunderstorms may generate a distinc¬
tive flow response at springs some distance from the
storm. Organisms unique to a particular subsurface
drainage basin may be considered tracers if their distri¬
bution can be used to define the boundaries of the basin.
Naturally occurring tracers are often used to "date" the
water, artificial tracers are used to label a specific sink or
cave, and pulses send an identifiable signal through the
conduit or sometimes a large part of the aquifer.
Most tracer tests in karst areas are conducted using
artificially injected tracers. These tracers may be
broadly classified as (1) water-soluble tracers (dyes,
salts); (2) particulate or drift material tracers (club
moss spores); and (3) physical pulses (flood waves).
The ideal tracer should be:
1. Nontoxic
2. Not normally present in the study area (low
background)
3. Detectable at very low concentrations
4. Conservative (minimal sorptive and decay losses)
5. Have the same density of water (or be neutrally
buoyant)
6. Relatively inexpensive
7. Easy to sample using passive collectors
8. Easy to analyze and quantify in the laboratory
The vast majority of tracer tests in cave and karst areas
are conducted using fluorescent dyes. All fluorescent
dyes have some tendency to adsorb to clay minerals and
some have high photochemical decay rates. The relative
fluorescent intensity of many dyes is pH- and tempera¬
ture-dependent. These considerations are usually not a
problem for underground tracing in cavernous aquifers
with high flow rates. Tracer tests in diffuse-flow aquifers
often involve the use of more conservative isotopes where
the longer residence times make fluorescent dyes less
suitable. A list of the principal fluorescent tracer dyes
commonly used in North American karst studies is pre¬
sented in Table 1.
Tracer tests from Europe have largely been con¬
ducted quantitatively and required direct sampling of
the resurgences. Studies in the British Isles and North
America through 1990 have mostly been qualitative
and used passive collectors to monitor the resurgences.
Qualitative tracing is still used for reconnaissance-type
studies and work in remote areas, but quantitative
tracing techniques are now the norm for most scientific
studies throughout the world.
ENCYCLOPEDIA OF CAVES
WATER TRACING IN KARST AQUIFERS
891
TABLE 1 Principal Fluorescent Tracers Used in North American Karst Studies
Tracer Colour Index
Name Color
Excitation
Emission
Maximum
Passive
Detector
Remarks
Phorwite AR Solution
FB # 28 UV blue
349 nm
430 nm
Unbleached
cotton
Invisible in solution. High photochemical decay rate.
Often high background fluorescence. May be present
due to household detergents.
Solophenyl Direct
»;397 nm
Unbleached
Sensitive to pH. No coloring of water. Usually low
Yellow 96 UV yellow
^490 nm (pH
sens.)
cotton
background.
Sodium Fluorescein
Acid Yellow 73 Yellow-
491 nm
512 nm
Activated
carbon 6—14
High photochemical decay and sensitive to pH.
Probably the most frequently used tracer for karst
green
mesh
studies. Low sorption.
Eosin Acid Red 87 Red
516 nm
538 nm
Activated
carbon 6—14
mesh
High photochemical decay. Can overlap with sodium
fluorescein if both are present in the same sample.
Rhodamine WT Cl Acid
Red 388 Red
554 nm
580 nm
Activated
carbon 6—14
mesh
Low background. Good stability in sunlight. Moderate
sorption. Generally requires a fluorometer for detection.
Some environmental concerns.
The fluorescent dyes commonly used as tracers are
believed to be quite harmless to people and the envi¬
ronment. Reviews of the toxicity data for tracer dyes
were presented by Smart (1984) and Behrens et al.
(2001). Several dyes in the rhodamine group, especially
rhodamine B, are not recommended due to environ¬
mental concerns, but the high sorptive tendency for
some of these dyes makes them a poor choice for
groundwater tracing under any circumstances.
A lot of confusion exists over the names of many of
the tracers. Sodium fluorescein is known as uranine in
Europe and commonly called fluorescein in the North
American literature. Only the sodium salt of fluores¬
cein is soluble enough to make it an efficient tracer.
When describing a dye, the Color Index name should
be cited at least once. Sodium fluorescein has a Color
Index constitution number Cl 45350 and a generic
name of acid yellow 73. Also, all dye shipments should
be checked when they are received to make certain the
correct product is being used.
Note that a few words of caution are in order for any¬
one planning on conducting tracer tests. Specialists now
do most dye tracing, and water tracing by untrained
people may be very damaging. In the United States,
some states require a permit for conducting a tracer test,
and the local water inspectors should always be notified
in advance of any test so a hazardous materials team is
not called out at great expense. A few states do not
allow any water tracing tests. Care should be taken to
avoid coloration at water intakes. All work in a given
area should have one coordinator so various tests cannot
cross tracers. Some epikarst aquifers may take several
years to clear a given dye from an injection point, so
careful planning is needed to avoid confounding future
tests using the same tracer. The following description of
the basic procedures for conducting tracer tests in karst
areas is not intended to serve as a training manual.
Tracers, such as antibiotics, will lose their usefulness if
overused or improperly used through time.
QUALITATIVE TRACING USING
PASSIVE DETECTORS
Passive collectors such as packets of activated carbon
or unbleached cotton are placed in the monitoring
points and collected at regular intervals. The dye, if
present, is then eluted from the carbon or exposed by
examining the cotton under ultraviolet light. The use of
a fluorometer to analyze the elutant increases the mini¬
mum detectable concentration of the tracer by several
orders of magnitude. It is not possible to reliably obtain
the dye concentrations at the resurgences from this
technique, because exposure time of the detector to the
dye and other variables affect the elutant dye concentra¬
tion. Time of travel for the tracer can only be approxi¬
mated based on the changing interval for the detectors.
Detailed instructions for conducting qualitative tracer
tests are presented in Aley and Fletcher (1976), Jones
(1984), and Alexander and Quinlan (1996).
The principal groups of passive detectors and tracers
are:
1. Activated carbon (activated charcoal)
a. Sodium fluorescein
b. Eosin
c. Rhodamine WT, Sulforhodamine B
ENCYCLOPEDIA OF CAVES
892
w
2. Cotton
a. Tinopal CBS-X (FB 351)
b. Phorwite
c. Calcophor white (FB number 28)
d. Direct yellow 96
3. Plankton netting (about 25-micron mesh openings)
a. Colored spores of Lycopodium calvatum
The fluorescent dyes have been used extensively in
North America and Europe while tracing using spores
has been primarily a European technique. The follow¬
ing discussion outlines the basic procedure for con¬
ducting a qualitative tracer test using sodium
fluorescein and carbon detectors.
The execution of a tracer test should involve an ini¬
tial survey of the study area to identify the hydrologic
boundaries and all possible springs and resurgences.
Any previous tracer tests from the area and all cave
survey data should be studied. All landowners poten¬
tially affected by the test should be contacted and
access secured to injection and monitoring points.
Local and state agencies should be notified about the
study and any required permits obtained. Measure¬
ments of background fluorescence from the springs
should be used to help determine the most suitable
tracers. Initial tests in an area should be conducted
during average flow conditions; tests under minimal
discharge should generally be avoided.
The amount of dye to use for any given test is an edu¬
cated guess at best. The quantity of tracer to be injected
is a function the hydrologic conditions of the study area
and the characteristics of the tracer and the analytical
procedure. The possible travel distances and discharges
of the resurgences may be used to help estimate the
travel time and tracer quantity required for the test.
Several formulae to estimate dye quantities are pre¬
sented in Kass (1998), but these provide only rough
guidelines. Tests in conduit aquifers involving distances
less than 1 km and discharges less than 30 L s -1 should
require less than 200 g of sodium fluorescein. Most of
the other commonly used tracer dyes probably require
two to four times the mass used for a sodium fluores¬
cein test. Less dye is usually needed for qualitative tests
than for quantitative tests. If the possible coloration of
the resurgences is not objectionable, the estimated
amount of dye should be increased. Care must be taken
to keep maximum tracer recovery concentrations below
the visible threshold (about 30 pg L 1 for sodium fluo¬
rescein) if the tracer could affect a water supply. The
U.S. Geological Survey suggested a maximum dye con¬
centration of 10 pg L -1 at water intakes, but this is a
conservative number and is not based on toxicological
considerations (Alexander and Quinlan, 1996).
The monitoring locations should include all possible
resurgences and one or two "impossible" ones to serve
as controls for the test. Carbon detectors are usually
made of envelopes of plastic screening material contain¬
ing a few grams of activated carbon granules (about
6—14 mesh). Cotton detectors are usually balls of
unbleached cotton. The passive detectors should be
changed at regular intervals, usually between 2 and 14
days. The detectors should be in place for at least one
changing interval prior to dye injection to establish
background information. They should be placed to
remain submerged with changing water levels, be read¬
ily recoverable, and remain out of sight to casual obser¬
vers. An interesting "hanger" for detectors in deeper
water was developed by James Quinlan and is shown
in Figure 2. The detectors should be placed so medium
velocity current flows through them and they are not
exposed to direct sunlight. Detectors lose their ability to
adsorb the tracers through time, but they will retain the
dye for periods of at least one year following exposure
when kept in total darkness. Caution must be used to
prevent contamination of the detectors and they should
be placed before the dye is injected.
Injection points for the tracer are ideally flowing
streams at surface sinkpoints or upstream of impassible
reaches in a cave. Powdered dyes are often mixed with
water prior to injection to avoid unwanted spread of
the dye. The injection points should be chosen to mini¬
mize the surface exposure of the tracers to sunlight.
FIGURE 2 A "Quinlan Gumdrop" anchor for passive dye collec¬
tors. The base is concrete molded in a coffee can and the detectors
are suspended above the bottom of deeper springs or streams. The
screen envelops contain unbleached cotton (right arm) and activated
carbon (left arm and bottom). The packets can also be secured with
wire to rocks or other solid anchors such as old railroad spikes.
ENCYCLOPEDIA OF CAVES
WATER TRACING IN KARST AQUIFERS
893
Tests involving injection points without rapid con¬
tinuous flow such as a dry sinkhole or water well are
much more difficult and more prone to failure. Tracer
tests from sinkholes or wells are usually only con¬
ducted by professional hydrologists. The basic proce¬
dure is to use a tank truck to deliver water to the site.
Several thousand liters of water are run into the sink¬
hole before the dye is added. The dye is then flushed
into the aquifer system with about an additional 10,000
liters of water. Injection sites that use water wells often
provide unsuccessful tests, which means that the dye
injected as the tracer will not be usable in the area for
a number of years. Caution and careful planning are
needed for this type of test.
Another dye injection method involves leaving the
dye in a safe storage area in a presently dry sink or
streambed and letting the next storm wash the tracer
into the flow system. This "dry set" system works best
in remote areas where people or animals are unlikely
to encounter the tracer before it is removed by
precipitation. The analysis of the carbon for the pres¬
ence of sodium fluorescein involves the following steps:
1. Rinse the carbon in clean water to remove sediment.
2. Fill a test tube about half full with the carbon.
3. Pour in the test solution, called the eluent, to cover
the carbon about 3 cm deep (5—10 mL).
4. Let the samples sit undisturbed in the dark for
between 30 min and 24 hr.
5. Shine a concentrated white light source such as a
focusable flashlight through the test tube and look
for the characteristic fluorescent green sheen on top
of the carbon granules (do not shake or disturb the
solution for a visual examination).
Some workers prefer to use a larger diameter vial
to increase the length of the light path through the elu¬
ent. The most commonly used eluent for sodium
fluorescein is 5% potassium hydroxide (KOH) in 70%
isopropyl alcohol. Many other eluents have been used
including the "Smart Solution" (50% 1-propanol, 20%
NH 4 OH, and 30% distilled water by volume). It takes
some practice to distinguish weakly positive sodium
fluorescein tests from some naturally occurring back¬
ground. Numerous trials at varying dye concentrations
should be run under controlled conditions before mov¬
ing on to actual field tests. One of the advantages of
the passive collector system is that the samples are
time-integrated for the exposure period of the detec¬
tors at each monitoring point.
QUANTITATIVE WATER TRACING
IN KARST AREAS
Many of the professionally conducted water tracing
tests are quantitative. Resurgences are sampled at short
discrete time intervals and the water is analyzed using
a calibrated fluorometer to determine the concentration
of the dye in each sample. Several different dyes may
be injected simultaneously so multiple sink points may
be tested at the same time and under the same flow
conditions. The results are plotted on a graph showing
the concentration of the dye through time (breakthrough
curve). The expense and complexity of quantitative tests
are much higher, but considerably more information is
obtained about the nature of the subsurface flow sys¬
tems. The combination of quantitative dye tracing
results with discharge measurements allows a more cer¬
tain delineation of karst drainage basins and the calcula¬
tion of the percentage of discharge from different
sources at a given resurgence. Good results are depen¬
dent on using an appropriate tracer, adequate sampling
frequency, and accurate analysis of the samples. A paper
by Field (2003) provides formulas for determining dye
mass and sampling frequency, but a good estimation of
hydraulic and transport characteristics of the system is
needed. This usually requires advance knowledge of the
flow conditions and the resurgence(s). An estimate of
the expected travel time from sink to resurgence is also
helpful in planning tests. Flow velocities for conduit
karst aquifers are typically in the range of 0.2 to 7 km
per day, but some tracer tests have shown results far
outside of these values.
The planning and dye injection procedure is essen¬
tially the same as described for qualitative traces.
Qualitative tests are often conducted first to identify
which resurgences to monitor and to estimate the
travel time for the tracer. Most tracer tests in karst
areas use an instantaneous injection of the dye,
but continuous (over several hours) releases are occa¬
sionally used to facilitate the calculation of discharge
at the recovery point. Fluctuations in background fluo¬
rescence are often a problem in the blue and green
wavelengths, so the red dye Rhodamine WT is the
most commonly used fluorescent dye for quantitative
studies. Quantitative tests generally require about
twice as much dye as qualitative tests. The much
greater sampling effort required for quantitative test¬
ing is the major reason for the increase in cost.
The instrument used to determine fluorescent dye
concentrations is a fluorometer (Fig. 3). Fluorometers
pass light through the sample at a controlled excitation
wavelength and measure the relative fluorescent inten¬
sity of the output signal at the emission wavelength.
Laboratory standards at different dye concentrations
must be prepared for each batch of dye to calibrate the
fluorometer to determine dye concentrations. Filter
fluorometers may be laboratory or field instruments
and use various filters to set the excitation and emis¬
sion wavelengths, while scanning spectrofluoropho-
tometers are laboratory instruments that are better
ENCYCLOPEDIA OF CAVES
894
w
FIGURE 3 A scanning spectrofluorometer used for laboratory quantification and identification of fluorescent dyes and a battery-powered
hand held fluorometer (lower left) used for field monitoring of a dye pulse.
suited to analyzing samples containing multiple dyes
or having high fluorescent backgrounds. A continuous
scan through the wavelengths of interest with about a
20-nm separation between the excitation and emission
settings allows a single scan to be analyzed for
several different dyes. Some workers use computer
enhancement of the instrument output signal to help
identify and separate overlapping fluorescent peaks.
Spectrofluorophotometers are more selective in their
ability to isolate specific fluorescent compounds, but
both types of fluorometers are highly sensitive to low
concentrations of dye. Recording filter fluorometers
provide a high resolution record of the passage of the
dye pulse, but spectrofluorophotometers are more reli¬
able in the identification and quantification of individ¬
ual tracers.
Filter fluorometers may be fitted with a flow¬
through door which circulates water past the instru¬
ment's light source. This provides a continuous
record of fluorescent intensity in one set of wave¬
lengths, but this is usually only practical for tests
with travel times of less than 24 hours. Fluorometric
probes compatible with data loggers make continuous
fluorometric measurements more cost-effective for
many field studies and record a nearly continuous
picture of the dye pulse through time. Handheld fluo¬
rometers with several "channels" set for specific
tracer dyes are robust enough to take into caves, but
the samples must be taken manually. Most quantita¬
tive tracer studies at present involve the use of auto¬
matic water samplers at several resurgences over a
period of days to weeks. The sampling interval is
dependent on the expected travel time of the tracer
and may range from 15 min to one per day. Some res¬
olution in the shape of the dye recovery curve may be
lost if the sampling interval is too long. Sampling
should continue for a time sufficient to ensure that
most of the dye pulse has been recovered. Samples
should be stored in the dark in glass containers to
minimize photochemical decay and sorption of the
dye to the sample bottles.
INTERPRETATION OF WATER
TRACING TESTS
Qualitative tracer tests are positive if the dye is
detected at one or more monitoring points. A success¬
ful tracer test may prove a hydrologic connection
between the injection point and recovery point(s) of
the tracer, but the exact flow path(s) will remain
unknown. Depending on the changing interval for the
detectors, some information of the travel time for the
tracer may be available. If the tracer is not detected at
a monitoring station, it is listed as "none detected"
rather than "negative." Negative tests do not prove
that a connection does not exist and the percentage of
the injected tracer actually recovered can only be
determined by quantitative measurements of the dye
concentration and discharge at the resurgences during
the recovery period of the dye pulse.
Tracer tests are representative of the karst flow system
for the hydrologic conditions prevailing at the time of the
test. Overflow routes may exist that function only during
high-flow conditions, so a thorough study of a karst
ENCYCLOPEDIA OF CAVES
WATER TRACING IN KARST AQUIFERS
895
FIGURE 4 Graph showing a simple breakthrough curve from a tracer test from a cave to a spring. The principal components of the curve
are (A) first arrival of the dye at 30 hours; (B) peak (maximum) recovery concentration at 42 hours; (C) dye centroid (center of mass and
median recovery point) at 50 hours; and (D) total time of dye cloud passage from 30 through 120+ hours. Note the four-hour sampling
interval and that sampling was stopped before the complete passage of the dye cloud.
region will require multiple tests at different water levels.
Some resurgences may only function during high-flow
conditions. The travel times for the tracers will be less at
high water levels, and the dye will move much faster if
the conduits are completely flooded. Quantitative tests
between the same injection and recovery sites under a
range of flow conditions may be used to develop the rela¬
tionship between travel time and groundwater discharge
and to predict solute transport characteristics. Discharge-
dependent thresholds may be identified.
The graph of dye recovery concentration plotted
through time is called a dye recovery or breakthrough
curve (Fig. 4). This curve presents a picture of the pas¬
sage of the dye "cloud" or pulse through time at the
stationary sampling point. Breakthrough curves are
typically skewed to the right with rapid rise to peak
(maximum) dye concentration and a long trailing edge
of declining dye concentrations. The shape or form of
the breakthrough curve is a function of the characteris¬
tics of the tracer, the flow conditions during the test,
and the characteristics of the aquifer or conduit sys¬
tem. The mean flow velocity is usually computed
using the straight-line distance between sink and
resurgence and the time between dye injection and
recovery of half of the tracer. This is the centroid or
center of recovered mass on the breakthrough curve.
Some workers use the time to peak concentration if the
resurgences are not completely sampled through the
long trailing period of declining tracer concentrations.
If discharge is known throughout the dye recovery
period, the amount of dye recovered can be calculated
by integrating the area under the breakthrough curve
and multiplying by the discharge:
M = Q
o
Cdt
where M = mass of dye recovered, Q = discharge,
C = dye concentration at time t.
Discharge is usually based on current-meter measure¬
ments during the test. If the tracer is conservative (not
subject to decay or sorptive losses) and all of the water
from the injection point is accounted for, the mass of dye
recovered should be equal to the mass injected. This is
the dye budget for the test. None of the fluorescent trac¬
ers is completely conservative, but tracer tests with travel
times of less than a few days should balance reasonably
well if all of the resurgences have been monitored.
Assuming complete mixing and a conservative tracer,
the area under the time-concentration breakthrough
curve should be constant even as the shape of the curve
changes in the downstream direction. A significant loss
of the tracer from the system suggests that additional
resurgences are present or some of the tracer has moved
into storage areas along the flow route. Less conservative
tracers usually produce a longer travel time estimate due
to retardation caused by sorption and desorption along
the flow route.
Breakthrough curves for many tracer tests in karst
areas exhibit multiple peaks, and the time between the
peaks may be dependent on discharge. The interpreta¬
tion of these tests is somewhat subjective, but much
ENCYCLOPEDIA OF CAVES
896
w
insight into the internal flow characteristics of the karst
aquifer and the conduit system can be gained from
these tests. The best breakthrough curve data are
obtained using continuous flow fluorometry during
the passage of the dye pulse. The shape or form of the
dye cloud is dependent on the interaction of disper¬
sion, dilution, divergence and convergence, and stor¬
age (Smart, 1988). The quantitative interpretation of
breakthrough curves is generally based on solute
transport equations. An introduction to this type of
analysis was presented in Benischke et al. (2007) and a
computer program to analyze tracer recovery data was
developed by Field (2002).
Traces involving longer distances and longer travel
times show lower peak concentrations and longer
persistence of the dye pulse as the tracer becomes
increasingly more dispersed with time and distance. The
shape of the tracer cloud becomes more asymmetrical as
dispersion is influenced by the tracer moving in and out
of pockets of dead zone storage along the flow route.
Dispersion is also affected by the retardation factor for
tracers that are less than completely conservative.
Divergence occurs where the underground flow
routes divide or break away from the main conduit.
Convergence is where the routes rejoin the main
conduits. The divergent routes generally are longer
flow paths than the main conduit, so the portion of the
flow and tracer diverted may reappear as second or
multiple pulses at the sampling station.
Dilution is caused by unlabeled tributary water mix¬
ing with the water containing the original tracer injec¬
tion. This will cause a lower concentration of dye in
the downstream water samples.
Storage of the tracer in the conduit system may be
very complex. Storage may just represent very low
velocity flow through one section of a branching con¬
duit, called inline storage. A test done under conditions
of diminishing flow may leave some of the dye aban¬
doned in upper level pockets to be remobilized when
the flow increases at a later date. Some dye may become
stored offline in lateral areas along the route such as
small voids or in pores in the sediments. Inline storage
tends to create dye pulses that correlate with flushing
due to storm events. Dye from offline storage tends to
drain slowly during groundwater recessions. Some con¬
ceptual models of breakthrough curves for different
flow levels and conduit geometries are presented in
Figure 5 (adapted from Smart and Ford, 1982).
HIGH FLOW
LOW FLOW
STORAGE
a
ANASTOMOSING
CHANNELS
FIGURE 5 Sketches showing conceptual models of possible dye breakthrough curves. The pattern of the recovery curve is a function of
prevailing water levels at the time of the test and the geometry of the conduit system. From W.K. Jones, 1984; after C.C. Smart and D.C. Ford,
1982. With permission.
ENCYCLOPEDIA OF CAVES
WETLANDS IN CAVE AND KARST REGIONS
897
The best picture of the conduit aquifer system for
any karst area is from the actual surveying of the cave
passages. Tracer tests, interpreted in the context of the
regional geology and hydrologic setting, can be used
to fill in the blank place on the map where actual
exploration of the caves is not possible. The rather
unusual mixture of quantitative techniques and subjec¬
tive interpretation of the unseen part of the aquifer
makes water tracing a fascinating endeavor.
See Also the Following Articles
Hydrogeology of Karst Aquifers
Springs
Wilson, J. F. (1968). Fluorometric procedures for dye tracing. Techniques
of water-resources investigations of the U.S. Geological Survey (Book 3,
Chapter A12, pp. 1—34) (Revised).
WETLANDS IN CAVE AND
KARST REGIONS
Tanja Pipan and David C. Culver 1
Karst Research Institute ZRC SAZU, ^American University
INTRODUCTION
Bibliography
Alexander, E. C., Jr., & Quinlan, J. F. (1996). Practical tracing of ground-
water with emphasis on karst terranes. In Guidelines for wellhead and
springhead protection area delineation in carbonate rocks (EPA 904-B-97-
003, Appendix B, pp. 1—38). Nashville, TN: Eckenfelder, Inc.
Aley, T., & Fletcher, M. W. (1976). The water tracers cookbook.
Missouri Speleology, 16(3), 1—32.
Behrens, H., Beims, U., Dieter, H., Dietze, G., Eikmann, T., Grummt, T.,
et al. (2001). Toxicological and ecotoxicological assessment of water
tracers. Hydrogeology Journal, 9, 321—325.
Benischke, R., Goldscheider, N., & Smart, C. (2007). Tracer tech¬
niques. In N. Goldscheider & D. Drew (Eds.), Methods in karst
hydrogeology (pp. 147—170). London: Taylor and Francis.
Field, M. S. (2002). The QTRACER2 program for tracer-breakthrough
curve analysis for tracer tests in karstic aquifers and other hydrologic
systems. U.S. Environmental Protection Agency.
Field, M. S. (2003). A review of some tracer-test design equations
for tracer-mass estimation and sample collection frequency.
Environmental Geology, 43, 867—881.
Freixes, A., Monterde, M., & Ramoneda, J. (1997). Tracer tests in the
Joeu karstic system (Aran Valley, Central Pyrenees, NE Spain).
In A. Kranjc (Ed.), Tracer hydrologi/ 97 (pp. 219—225). Rotterdam,
The Netherlands: A. A. Balkema.
Gabrovsek, F., Kogovsek, J., Kovacic, G., Petrie, M., Ravbar, N., &
Turk, J. (2009). Recent results of tracer tests in the catchment of
the Unica River (SW Slovenia). Acta Carsologica, 39(1), 27—37.
Jones, W. K. (1984). Analysis and interpretation of data from tracer
tests in karst areas. National Speleological Society Bulletin, 46(2),
41-47.
Kass, W. (1998). Tracing technique in geohydrology. Rotterdam, The
Netherlands: A. A. Balkema.
Quinlan, J. F., & Ewers, R. O. (1989). Subsurface drainage in the
Mammoth Cave area. In W. B. White & E. L. White (Eds.), Karst
hydrology: Concepts from the Mammoth Cave Area (pp. 65—103).
New York: Van Nostrand Reinhold.
Smart, C. C. (1988). Artificial tracer techniques for the determinat¬
ion of the structure of conduit aquifers. Ground Water, 26(4),
445-453.
Smart, C. C., & Ford, D. C. (1982). Quantitative dye tracing in a gla-
cierized alpine karst. Beitrage Geol. Schweiz-Hyrdrologie, 28(1),
191-200.
Smart, P. L. (1984). A review of the toxicity of twelve fluorescent
dyes used for water tracing. National Speleological Society Bulletin,
46(2), 21-33.
Smart, P. L., & Laidlaw, I. M. S. (1977). An evaluation of some fluo¬
rescent dyes for water tracers. Water Resources Research, 13, 15—33.
A general definition of wetlands is an area of land
where the soil is saturated with moisture either perma¬
nently or seasonally, which may also be covered par¬
tially or completely by shallow bodies of water. They
include swamps, marshes, bogs, fens, and peatlands. A
more restricted and legalistic definition is that of the
U.S. Environmental Protection Agency: "those areas
that are inundated or saturated by surface or ground-
water at a frequency and duration sufficient to sup¬
port, and that under normal circumstances do support,
a prevalence of vegetation typically adapted for life in
saturated soil conditions." The emphasis of this defini¬
tion on vegetation makes identification of wetlands
easier. Many wetlands, both freshwater and saltwater,
play an important role in water purification, are highly
productive in terms of carbon fixation, and play a key
role in the life cycle of many vertebrates, such as anad-
romous fish and migratory waterfowl.
At the same time, wetlands are highly susceptible to
destruction by human activities, particularly by drain¬
ing. Wetlands are of great interest and concern to both
international and national agencies. The UNESCO
Ramsar Convention on Wetlands is designed to high¬
light important wetlands internationally. In the United
States, wetlands as defined by the U.S. Environmental
Protection Agency are probably the most regulated
and protected (and threatened) general landscape
features.
In karst areas, wetlands have received less recogni¬
tion and attention than elsewhere, probably because
wetlands are not especially common, due in part to the
nature of karst and cave areas. Since precipitation rap¬
idly sinks into the ground, wetlands are not usually
formed by the accumulation of surface runoff in areas
of recharge. Euliss et al. (2004) suggest a continuum of
wetlands along a groundwater axis (ranging from
recharge to discharge) and atmospheric water axis (pre¬
cipitation). In karst areas, wetlands usually only occur
at the discharge portion of the groundwater continuum,
near springs and resurgences. Secondly, since most
Encyclopedia of Caves.
© 2012 Elsevier Inc. All rights reserved.
898
w
streams are underground, there is less opportunity for
development of surface wetlands.
The broad patterns of distribution of major wetland
areas (Mitsch and Gosselink, 2007) are nearly the
reverse of the distribution of karst areas (Fig. 1).
Wetlands are especially common at far northern lati¬
tudes (50°—70° N) and near the equator. These are
regions with poor soil drainage where hydric soils
form. In contrast, karst regions are concentrated at inter¬
mediate northern latitudes (30°—50° N). Of the 65 major
wetlands enumerated by Mitsch and Gosselink in their
textbook on wetlands, only two are unequivocally in
karst areas. One is the Hudson Bay lowlands in
Canada, one of the most extensive wetlands in the
world. However, it is probably the relatively low relief
and Pleistocene glaciations that are the defining features
of this wetland. The other large karst wetland listed by
Mitsch and Gosselink is Cuatro Cienegas in Mexico, a
1500-km 2 wetland in a valley of the Sierra Madre
Oriental formed by more than 200 springs and home to
a number of endemic obligate aquatic subterranean
(stygobiotic) invertebrates.
At smaller geographic scales, karst wetlands fit uneas¬
ily into wetland classification schemes. Leibowitz (2003)
suggests that isolated wetlands (ones without direct sur¬
face connection to rivers and streams) be defined as wet¬
lands "completely surrounded by uplands." By this
definition every sinkhole and dolina is potentially an iso¬
lated wetland and it almost becomes a synonym for sink¬
hole and dolina. The number of dolinas and sinkholes in
karst areas can easily exceed 100 per km 2 . Karst wetlands
are also different because large connections (e.y., cave
entrances) exist between surface water and groundwa¬
ter, and these connections allow for the rapid move¬
ment of water onto or away from the surface. Hence,
rapidly disappearing lakes and ponds can be found
FIGURE 1 Qualitative view of the distribution of wetlands and
karst areas with respect to latitude. Data for wetlands was based on a
graph in Mitsch and Gosselink (2007) and data for karst areas was taken
from Bala'sz (1977), and values estimated at 10° intervals.
in many karst regions. Finally, there is a complex
terminology of landscape features in karst wetlands
( e.g ., poljes, turloughs, estavelles), one that is almost
completely different from the terms used by wetland
scientists. Mitsch and Gosselink list 35 common terms
for wetlands ( e.g ., billabongs, bogs, bottomland), only
one of which (furlough) is a karst feature. The general
wetland type that is most useful in framing discussion
about karst wetlands is the isolated wetland, especially
temporary freshwater wetlands. Temporary wetlands
are categorized on the basis of
1. Periodicity of flooding, whether it occurs every
year, occasional years, or several times throughout
the year
2. Seasonality of flooding
3. Duration of flooding, the average length of time the
wetland persists
4. Predictability of flooding
Numerous authors have pointed out that not only is
the vegetation of wetlands adapted to saturated soil,
but the animals of wetlands must cope with the peri¬
odic drying of the habitat (Wiggins et al., 1980).
Among the adaptations to life in temporary water are
burrowing, resistant resting stages, and migration. In
spite of the differences in distribution, definition, and
focus, wetlands and karst systems do share some key
features which make their joint consideration impor¬
tant. Both are highly dependent on the hydrologic
cycle, both involve groundwater/surface water interac¬
tions, and both are very fragile systems that are espe¬
cially susceptible to human impacts.
Depending on the definition used (see above), wet¬
lands in karst areas could include not only relatively
large wetlands of the types we review below, but also
smaller features, often less than one hectare in size, that
are sometimes covered with water but may lack a char¬
acteristic flora adapted to life in saturated soil. These
smaller features include some sinkholes, dolinas, uvalas,
and springs. We limit our description to the larger sur¬
face wetlands, as the logical starting point for the incor¬
poration of wetlands concepts into the physical
geography of karst.
The Dinaric karst, ranging from extreme northeast
Italy south to Albania, is both the "type locality" and
the site of an extensive variety of landscape features
that are the benchmark against which other karst
areas, especially in temperate regions, are compared. A
series of karst wetlands in the Dinaric karst of
Slovenia, although they have not been called wetlands,
are well described, and serve as a useful starting point
of analysis of karst wetlands. Karst wetlands, espe¬
cially poljes, are the dominant wetland type through¬
out the Dinaric karst. They comprise about 4% of the
72,000 km 2 of the Dinaric karst and are critical habitat
ENCYCLOPEDIA OF CAVES
WETLANDS IN CAVE AND KARST REGIONS
899
for many waterbirds (Stumberger, 2010). Undoubtedly,
other kinds of large karst wetlands occur in tropical
and polar regions, but they have been little studied.
CERKNISKO JEZERO/POLJE
Called Cerknica Lake in English, the site is given the
name jezero (lake) when flooded and polje (field) when
it is dry (Gaberscik, 2003). This is perhaps the best-
known seasonal lake, and scientists have sought expla¬
nations for its sudden appearance and disappearance
since Valvasor's diagram from the seventeenth century
(Fig. 2). The lake featured prominently in Valvasor's
epic description of the Slovenian landscape. Die Ehre
dess Herzogthums Crain. The entire closed depression
(Cerknisko polje) in which Cerknisko jezero occurs is
38 km 2 . The lake itself can occupy up to 27 km 2 and
reach a depth of more than 2 m, or be completely dry.
When the lake is present, it is used for recreation.
When it is dry, the land is used for grazing and
cropland. Typically, the lake is completely dry for
80 days in the summer (August to October). Occasionally
(about once a decade), the lake does not completely
dry out anytime during the year. It reaches its maxi¬
mum extent in April and May, and again in December.
The transition from wet to dry (and vice versa) is rapid,
with the lake changing level of up to 8 cm per day.
One of the water exits, a spectacular ponor (swallow
hole), appears during drying (Fig. 3). In the past there
have been attempts both to permanently drain the lake
for agricultural purposes and to make the lake perma¬
nent to enhance recreational opportunities. All these
projects failed.
Not surprisingly, the flora of Cerknisko jezero is
very diverse. Aquatic plant communities are very
restricted because of the intermittent nature of the
lake. Most species die back during the dry phase.
Among such species are Charetea fragilis and several
species of Potametum. Amphibious species, ones where
the species occur in both wet and dry phases and
where the duration of the aquatic phase is at least 3—4
months, include Rorippa amphibian and Shim latifolinm.
Typical wetland species include Phragmites australis,
Equisetum fluviatile, many species of Carex, and Nymphaea
alba. There are also species living on the edge of the
inundation zone, including Deschampsia caespitosa and
Plantago altissima. Many of these wetland species are
rare or endangered in Slovenia.
The Crustacea reported from Cerknisko jezero
include a fairy shrimp, Chirocephalus croaticus, a rare
species limited to temporary karst waters (Fig. 4). The
population was destroyed because pools were filled
with debris, and its recent occurrence has not been
demonstrated. A total of 21 cladoceran species and 15
copepod species have been recorded from the lake.
FIGURE 2 Diagram of the subterranean drainage of Cerknisko jezero as conceived of by Janez Vajkard Valvasor in 1689. It was the first
attempt to provide a rational explanation for the disappearance and appearance of Cerknisko jezero. From J.V. Valvasor (1689) Die Ehre dess
Herzogthums Crain. I. Th., Laibach-Nurnberg.
ENCYCLOPEDIA OF CAVES
900
w
FIGURE 3 Ponor in Cerknisko polje, a major drain for the lake.
Photo courtesy of S. Polak. Used with permission.
FIGURE 4 The fairy shrimp (Anostraca), Chirocephalus croaticus, a
rare species limited to temporary karst waters. Photo courtesy of
S. Polak. Used with permission.
While copepods are ubiquitous, cladocerans tend to be
most common in temporary waters. One of the cope-
pod species, Diacyclops languidoides, is an obligate sub¬
terranean dweller (stygobiont). The lake is also a
hotspot for both Odonata and Trichoptera, with sev¬
eral rare and endangered species in Slovenia.
In the seventeenth century, three fish species were
known to inhabit Cerknisko jezero—pike ( Esox Indus),
tench ( Tinea tinea), and burbot ( Lota lota). Unfortunately,
as in many other lakes worldwide, several species have
been introduced. In a stream in the polje, 11 species of
fish are known, the majority of which were introduced.
The cave salamander Proteus anguinus has been
reported from Cerknisko jezero, where it washes in
from subterranean channels as the lake fills.
PLANINSKO POLJE
Planinsko polje is the second most studied polje. In
comparison to Cerknisko jezero/polje, it is much drier,
and the period of flooding is much shorter, up to a
maximum of 2 months. It is typically referred to as a
polje, not a jezero. The geological setting for Planinsko
polje and poljes in general is quite complex, typically
aligned along tectonic axes, and covered with alluvium.
Two main springs drain into Planinsko polje (Fig. 5)
from the Javornik and Sneznik massifs, a string of poljes
including Cerknisko polje, and from the Pivka River
(Petrie, 2010). One of the springs, Unica, drains into the
large Postojna—Planina Cave System. Unica River flows
across Planinsko polje and plunges underground in a
series of ponors at the northern end of the polje.
Since the polje is dry for most of the year, it is an
important terrestrial habitat for mammals such as the
wildcat, Felix sylvestris, many birds, butterflies, and bee¬
tles. The Postojna—Planina Cave System, whose waters
exit at Unica spring, is the richest cave in the world in
terms of the number of obligate cave-dwelling species.
Especially during high water, the cave salamander
Proteus anguinus washes out of the cave and can be
found in the surface part of the Unica River. A number
of fish live in the Unica River, including the brown
trout Salmo trutta and seven other fish species native to
the river, and is an important recreational fishery.
Since there is a permanent stream across the polje,
there is a persistent zonation of plants, from terrestrial
to aquatic. Much of the land is used for grazing and
some crops. Its position at the junction of the Dinaric
and Alpine regions accounts in part for the high floris-
tic diversity. A species of hyacinth, Scilla literdierei,
is limited to Dinaric karst wetlands and is common
in Planinsko polje. Other species characteristic of
Planinsko polje are Gladiolus illyricus, Iris sibirica, Iris
pseudacorus, and Ophioglossum vulgatum.
PIVKA INTERMITTENT LAKES
In the drainage of Pivka River, which flows for only
26 km across a classic karst landscape before it empties
into Postojnska jama, a total of 17 intermittent lakes of
widely varying sizes appear with different periodicities
ENCYCLOPEDIA OF CAVES
WETLANDS IN CAVE AND KARST REGIONS
901
Malenscica spring
/
Unica spring
/,
l - .
- i ( - - A
1
Planina karst polje
FIGURE 5 Planinsko polje, with the two major springs (Malenscica and Unica) indicated. Photo courtesy of M. Petrie. Used with permission.
(Fig. 6). Unlike Cerknisko jezero, none of these lakes is
associated with a polje, but rather are in closed depres¬
sions (dolinas). The largest, Palsko jezero, is 1.03 km 2 ,
about 3% of the size of Cerknisko jezero. The actual
number of lakes is difficult to determine because there
are a number of very small dolinas (<0.01 km 2 ) that occa¬
sionally fill with water. For the entire area there is a close
connection between underground and surface water. The
karst aquifer is a rather shallow one, probably the result
of underlying flysch rocks. During most of the year
water flows underground but after periods of intensive
or long-lasting precipitation, the water table rises and
emerges at the surface in different locations (Petrie and
Kogovsek, 2005). Some of them appear annually and last
for up to 6 months, as is the case for Palsko jezero and
Petelinjsko jezero, the two largest lakes. In Palsko jezero,
the lake is present for an average of 3 months, but may
fill more than once a year if heavy rainfall occurs. Most
of the other lakes appear rarely, and only fill up in times
of big floods. The water table fluctuates more than 40 m
at the estevelle that both fills and drains Palsko jezero
(Fig. 7). In the smaller lakes, they recharge and discharge
mainly through the alluvial sediments which cover their
limestone lake bed. Some of the smaller lakes probably
fill only once a decade on average, although there are
few long-term data on the presence or absence of these
lakes. During and after the Second World War the area
was an important military base. In the past, the area was
used primarily for seasonal sheep grazing. At present,
agriculture is less intensive than in the past, and forests
are gradually covering the landscape, including some of
the intermittent lakes (Mulec et al., 2005). In 1985, in the
Pivka valley as a whole, 44% of the surface was pasture
and meadows, 26% forest, 9% cultivated fields, and 2%
wetlands.
Vegetation of the intermittent lakes has an interest¬
ing pattern. Soil is thin and rocky on the lake margins
and becomes thicker toward the center of the lake,
making the center more suitable for cultivation. In gen¬
eral, open meadow-like flooded areas and dry mea¬
dows at the margin of the lakes are important habitats.
Species characteristic of dry margins of the lakes
include Galium verum and Lotus corniculatus. In wetter
areas with thicker, nutrient-poor soil, species such as
Allium angulosum and Carex panacea are found. Some of
the species characteristic of wetter areas, such as Carex
hirta and Ranunculus repens, are becoming less common
because of drying out of some areas in recent decades.
In the areas that are flooded for longer periods, willows
{Sallx) overgrow meadows and pastures in the absence
of cultivation or grazing. The impact of reforestation on
the hydrological regime and the appearance of the tem¬
porary lakes are not precisely known but apparently
lead to reduced frequency of lake formation and
reduced residence time of the lake.
The crustacean fauna of the Pivka lakes includes two
species of fairy shrimp ( Branchypus schafferi and
Chirocephalus croaticus; Fig. 4), three Cladocera species,
nine Copepoda species, two Ostracoda species, one
Amphipoda, and one Isopoda. While the species list is
very incomplete and nonexistent for insects, it includes
stygobionts, generalist species, and species adapted to
temporary waters (Pipan, 2005). One of the copepod
species ( Diacyclops charon) and one isopod subspecies
(Asellus aquaticus cavernicolus) are stygobionts. The
European cave salamander Proteus anguinus can also be
found in the lakes as a result of washing out from
springs and estavelles. Typical of temporary wetlands,
fairy shrimp have thick-shelled winter eggs that are
resistant to desiccation. They also have a very high
reproductive rate, and in the case of C. croaticus, they
can complete their life cycle in 14 days, an obvious
advantage in temporary waters such as the Pivka lakes.
Pivka intermittent lakes may be representative of
karst wetlands in general than either poljes or disap¬
pearing lakes. While Cerknisko jezero and Planinsko
polje are large wetlands of major importance in
Slovenia and elsewhere in the Dinaric karst, the geo¬
logical conditions needed for polje formation (large
valley along tectonic axes) are not that common
(Bonacci, 2004). Pivka lakes, on the other hand, occur
in a landscape of dolinas (sinkholes) in a river valley,
an extremely common feature of karst landscapes.
Intermittent lakes in the Sinkhole Plain of central
ENCYCLOPEDIA OF CAVES
902
w
Postojnska
djama
Kobtljih gnzah
LEGENDA* LEGEND
1
6
M 2
7
■■ 3 H
8
4 A
9
5
_ _
—
Petelinjsko
jezero
izvir poo Paicjem
▲ •*
Ma'.ijeva jama
Palsko jezero
• vMiinik
Parski studenecY Parsko jezero
Drskovsko jezero
2 Zaoorsko jezero
\
• ° Backo jezero
/^VKlemi&ca
v *.* 4 *o ■
A L
/
■-no.-i-
lenjiek
: Sembijsko jezero
FIGURE 6 Hydrogeological map of the area of Pivka intermittent
lakes. Legend: (1) karst aquifer; (2) porous aquifer; (3) rock of very
low permeability; (4) permanent surface flow; (5) intermittent surface
flow; (6) intermittent karst lake; (7) karst spring; (8) cave; (9) ponor.
Map courtesy of M. Petrie. Used with permission.
Kentucky have many of the features of the intermittent
Pivka lakes. Vessels and Jack (2001) recognize these
intermittent lakes as wetlands, and note that the fauna
of the lakes includes components similar to what
Pipan (2005) reported for Pivka Lakes.
Sheehy Skeffington and Gormally (2007) and Gunn
(2006) discuss a karst wetland feature in Ireland, called
furloughs, found in grasslands in Ireland. They are also
intermittent lakes and differ in some details from the
Pivka intermittent lakes (Sheehy Skeffington et ah, 2006).
They suggest that the term furlough be used for the
Pivka intermittent lakes, but Gunn (2006) on the other
hand emphasizes the importance of glaciation in the
development and functioning of turloughs. Turloughs
have priority status under the European Habitats
Directive and therefore the term is a useful tool for con¬
servation. Perhaps the term intermittent lakes should
also be retained for these features to recognize their
connection to other kinds of wetlands.
PROTECTION OF KARST WETLANDS
There are a variety of country- or region-specific
laws that are important in wetlands protection, such as
the European Union's Habitat Directive and the United
States' Clean Water Act. Internationally, the most
important structure is the Ramsar Convention on
Wetlands, adopted in 1971. Utilizing a broad definition
of wetlands and a process by which individual countries
nominate a site, nearly 1900 freshwater and marine
wetlands have been designated as of March 2010. One
of the wetland types is "karst and other subterranean
hydrological systems—freshwater." A total of 69 sites
include this designation as one of the wetland types,
and for 23 it was listed as the dominant type (Table 1).
The 46 sites for which the karst designation is not the
primary one serve as a reminder that a variety of wet¬
land types can occur on karst, but are not unique to
karst. Of the 23 designated as primarily karst wetlands,
6 correspond to the poljes and intermittent lakes
described above. Ramsar sites of this type include
Cerknisko polje/jezero and Bayerische Wildalm and
Wildalmfilz in Austria. Another six focus primarily on
surface features, but permanent bodies of water rather
than intermittent lakes. These include a groundwater-
fed lake in Turkey—Kizoren Obrouk—and alpine karst
streams in Austria—Nationalpark Kalkalpen. To a
lesser extent than intermittent lakes, these kinds of
wetlands are not necessarily unique to karst. Several
large ecoregions were also designated as karst wetland
types, including a large region in Yucatan, Mexico.
The final category is that where caves are the pri¬
mary wetlands feature in the designation. These
include the binational Baradla Cave/Domica Cave
System in Hungary and the Slovak Republic, and
Skocjanske jame in Slovenia. Exactly what can be con¬
sidered as a subterranean wetland is not clear, but the
Ramsar Convention has an extraordinarily broad defi¬
nition of wetland, basically including any temporary or
permanent body of water less than 6 m deep. Under
the Ramsar definition, almost all subterranean waters,
except for deep phreatic lakes, could be called wet¬
lands. All of the caves on the Ramsar list are
ENCYCLOPEDIA OF CAVES
WETLANDS IN CAVE AND KARST REGIONS
903
FIGURE 7 Hypothetical stream profile showing sequential reversal of flow in estavelles: (A) low flow, only lower section of karst drainage
system flooded and the estavelles act as ponors (swallow holes); (B) medium flow, headward flooding of the system and the lowest estavelle
acts as a spring; (C) high flow, storage capacity of karst system exceeded and all estavelles become springs. From J.D. Vineyard and G.L. Feder
(1982). Springs of Missouri. Missouri Department of Natural Resources, Division of Geology and Land Survey.
TABLE 1 Primary Wetland Type for Rarnsar Sites on Karst
Wetland Type
Number
Marine/ saline
14
Permanent lakes
11
Tree or shrub-dominated wetland
3
Marsh
3
Peatlands
4
Rivers
3
Intermittent lake
2
Karst—primarily cave
8
Karst—primarily surface
6
Karst—intermittent lakes
6
Karst—large-scale landscapes
3
Other/ unknown
6
Source: From www.ramsar.org. (Accessed 10 March 2010).
substantial caves with unique biological or hydrologi¬
cal features, often more than 1 km long, and typically
including either insurgence or resurgences of water.
An alternative way to use the term wetland for sub¬
terranean water bodies is to look for analogies with
surface wetlands. For example, there can be under¬
ground floodplains and even oxbows (abandoned
stream channels). Of course, subterranean floodplains
and oxbows will lack any vegetation and any migra¬
tory waterfowl. Their biological significance is that
these are areas of significant organic carbon that was
deposited by lateral transport. Many obligate cave¬
dwelling terrestrial species depend on this resource as
their primary source of organic carbon.
See Also the Following Articles
Closed Depressions in Karst Areas
springs
ENCYCLOPEDIA OF CAVES
904
w
Bibliography
Balaszs, D. (1977). The geographical distribution of karst areas.
Proceedings of the Seventh International Congress of Speleology,
Sheffield, 1, 13-15.
Bonacci, O. (2004). Poljes. In J. Gunn (Ed.), Encyclopedia of caves and
karst science (pp. 599—600). New York: Fitzroy Dearborn.
Euliss, N. H., LaBaugh, J. L., Fredrickson, L. H., Mushet, D. M.,
Laubhan, M. K., Swanson, G. A., et al. (2004). The wetland contin¬
uum: A conceptual framework for interpreting biological studies.
Wetlands, 24, 448-458.
Gaberscik, A. (Ed.), (2003). In A. Gaberscik (Ed.), Jezero, ki izginja
Ljubljana: Monografija o Cerkniskem jezeru. Drustvo Ekologov
Slovenije.
Gunn, J. (2006). Turlough and tiankengs: Distinctive doline forms.
Speleogenesis and Evolution of Karst Aquifers, 4, 1—4.
Leibowitz, S. G. (2003). Isolated wetlands and their function: An eco¬
logical perspective. Wetlands, 23, 517—531.
Mitsch, W. J., & Gosselink, J. G. (2007). Wetlands (4th ed.). New York:
Wiley.
Mulec, J., Mihevc, A., & Pipan, T. (2005). Intermittent lakes in the
Pivka basin. Acta Carsologica, 34, 543—565.
Petrie, M. (2010). Characterization, exploitation, and protection of the
Malenscica karst spring, Slovenia. In N. Kresic & Z. Stevanovic
(Eds.), Groundwater hydrology of springs (pp. 428—441).
Amsterdam: Elsevier.
Petrie, M., Ik Kogovsek, J. (2005). Hydrogeological characteristics of the
area of intermittent karst lakes of Pivka. Acta Carsologica, 34, 599—618.
Pipan, T. (2005). Fauna of the Pivka Intermittent Lakes. Acta
Carsologica, 34, 650—659.
Sheehy Skeffinton, M., Moran, J., O'Connor, A., Regan, E., Coxon,
C. E., Scot, N. E., et al. (2006). Turloughs—Ireland's unique wet¬
land habitat. Biological Conservation, 133, 265—290.
Sheehy Skeffinton, M., & Gormally, M. (2007). Turoughs: a mosaic of
biodiversity and management systems unique to Ireland. Acta
Carsologica, 36, 217—222.
Stumberger, B. (2010). A classification of karst poljes in the Dinarides
and their significance for waterbird conservation. In D. Denac,
M. Schneider-Jacoby & B. Stumberger (Eds.), Adriatic flyway —
clasoing the gap in bird conservation (pp. 151—154). Radofzell,
Germany: Euronatur.
Vessels, N., & Jack, J. D. (2001). Effects of fish on zooplankton com¬
munity structure in Chaney Lake, a temporary karst wetland in
Warren County, Kentucky. Journal of the Kentucky Academy of
Science, 62, 52—59.
Wiggins, G. B., MacKay, R. J., & Smith, I. M. (1980). Evolutionary
and ecological strategies of animals in annual temporary pools.
Archieves de Hydrobiologie Supplement, 2, 97—206.
WHITE-NOSE SYNDROME: A FUNGAL
DISEASE OF NORTH AMERICAN
HIBERNATING BATS
Marianne S. Moore and Thomas H. Kunz
Boston University
INTRODUCTION
White-nose syndrome (WNS) is an epizootic currently
affecting several species of hibernating bats in North
America. The most devastating wildlife disease in
recorded history, WNS is causing unprecedented mor¬
tality and threatening regional extinction in at least one
previously common bat species (Frick et al., 2010). The
syndrome is named for a white, filamentous growth of
fungal hyphae and conidia (spores) on the nose, ears,
wings, and tail membranes of affected bats (Blehert
et al., 2009; Fig. 1). The first evidence of this fungal
growth on hibernating bats was photographed on 16
February 2006 at Howes Cave, located approximately
50 km west of Albany, New York (Blehert et al., 2009).
In the four years since it was first observed, the putative
fungal pathogen, Geomyces destructans ( Gd ), associated
with WNS, has spread rapidly from its epicenter,
southward to North Carolina, westward to Missouri
and Oklahoma, and northward to Canada. As of April
2011, Gd has been reported from bats in 17 states
(Connecticut, Delaware, Indiana, Maryland, Massachusetts,
Missouri, New Hampshire, New Jersey, New York, North
Carolina, Ohio, Oklahoma, Pennsylvania, Tennessee,
Vermont, Virginia, and West Virginia) and three Canadian
provinces (New Brunswick, Ontario, and Quebec),
although confirmed infections based on histopathology
have not been determined for bats from Missouri and
Oklahoma (United State Fish & Wildlife Service, 2011).
In contrast to stable or increasing pre-WNS bat
populations, declines of hibernating bats in the north¬
eastern United States, ranging from 30—99% annually
and averaging 73%, have been documented (Frick
et al., 2010). Assuming that current rates of mortality
continue, the once common little brown myotis ( Myotis
lucifugus) is expected to be extinct in the northeastern
United States by the year 2026 (Frick et al., 2010), and
at least five other hibernating bat species (M. s eptentrio-
nalis, M. sodalis, M. leibii, Eptesicus fuscus, and
Perimyotis subflavus) in this region are at risk (Blehert
FIGURE 1 Hibernating little brown myotis ( Myotis lucifugus )
infected with a putative fungal pathogen, Geomyces destructans, asso¬
ciated with white-nose syndrome. Photo by A.C. Hicks, New York
Department of Environmental Conservation. Used with permission.
Encyclopedia of Caves.
©2012 Published by Elsevier Inc.
WHITE-NOSE SYNDROME: A FUNGAL DISEASE OF NORTH AMERICAN HIBERNATING BATS
905
et ah, 2009; Turner and Reeder, 2009; Courtin et ah,
2010; K.E. Langwig, pers. comm). Based on morpho¬
logical and molecular (polymerase chain reaction,
PCR) criteria, Gd was recently reported from three
other hibernating bat species in North America (M. gri-
sescens [listed as federally endangered], M. austroripar-
ius, and M. velifer ) although infections have not been
confirmed in M. grisescens from Missouri or M. velifer
from Oklahoma based on histopathology. Two feder¬
ally endangered subspecies of Townsendi's long-eared
bat, Corynorhinus townsendii (C. f. virginianus and C. t.
ingens), are within the geographic range of hibernating
bats affected by Gd, but to date there is no evidence
that this fungus has spread to these two species.
No introduced pathogen has caused such a precipi¬
tous decline in any mammal species in recorded his¬
tory. Low reproductive rates and long life span of bats
(Kunz and Fenton, 2003) will make their recovery
extremely slow unless the syndrome can somehow be
controlled. The recent decline of bat populations in the
northeastern United States caused by WNS is likely to
have adverse economic consequences on the ecology of
both natural and human-derived agroecosystems.
BAT MORTALITY AND THE
PUTATIVE PATHOGEN
Beginning in late winter 2007/2008, wildlife biologists
began to document unusual physical conditions, beha¬
viors, and mortality in bats hibernating in caves and
abandoned mines in New York State. Subsequently, post¬
mortem histopathological examinations revealed a dis¬
tinct cutaneous fungal infection where hyphae invaded
the epidermis, hair follicles, sebaceous and sweat glands,
and breached the basement membrane of hibernating
bats (Blehert et ah, 2009). Examination of these bats with
cutaneous fungal infections revealed a previously unde¬
scribed fungus with an asymmetrically curved conidia
(Blehert et ah, 2009). Laboratory studies later documented
that this fungus is psychrophilic {i.e., cold loving) and
grows optimally between 5 and 10°C, well within the sea¬
sonal temperature range of most hibernacula (2—14°C)
(Blehert et ah, 2009). Phylogenetic analysis has placed
these isolates within the inoperculate ascomycetes (order
Helotiales) and near the anamorphic genus Geomyces (tel-
eomorph Psendogymnoascus). The closest relative of the
fungal isolates from bats was identified as the ubiqui¬
tous G. pannorum, a keratinophylic species of which a
subspecies, G. pannorum var. pannorum, has been iden¬
tified in superficial skin and fingernail infections in
humans. The combination of a unique morphology
and genetic composition (specifically at the identical
internal transcribed spacer region and small subunit
ribosomal RNA gene) distinguishes the bat isolate
from other members of Geomyces. This discovery led to
the recognition of a new species within the genus, and
subsequently the fungus associated with WNS was
aptly named Geomyces destructans. To date, it is not
entirely clear whether Gd is the causative agent of
WNS, but it appears that this fungus alters the physiol¬
ogy of hibernating bats (Boyles and Willis, 2009) and
thus may have a direct pathogenic role (Meteyer et ah,
2009; Cryan et ah, 2010). Aside from the histopatholog¬
ical evidence that Gd causes severe cutaneous lesions
(Meteyer et ah, 2009), secretory proteases involved
in tissue digestion have been identified from Gd sup¬
porting the etiologic role of the fungus in WNS
(Chaturvedi et ah, 2010). Additionally, genotyping of
five genetic markers using Gd isolates collected from
bats within a 200-km survey distance from its epicen¬
ter provide preliminary evidence that a single strain
was introduced into New York and a clonal popula¬
tion is infecting all bats within the affected region
(Chaturvedi et ah, 2010).
DIAGNOSTICS
Histopathology is currently the gold standard of
diagnostic tools used to confirm WNS, which with spe¬
cialized training and appropriate staining and visuali¬
zation of tissues, allows the identification of Gd and
associated cutaneous lesions (Meteyer et ah, 2009).
Culture, isolation and morphological examination can
also be used to identify presence of the fungus from
tissue or swab samples (Blehert et ah, 2009; Gargas
et ah, 2009; Chaturvedi et ah, 2010); however, a number
of researchers have noted low rates of isolation (e.g.,
54%) (Lorch et ah, 2010) despite an abundance of avail¬
able Grf-affected tissue, a complication that may be due
to competition from other microorganisms colonizing
bat skin (Chaturvedi et ah, 2010; Lorch et ah, 2010).
PCR amplification and sequencing of genetic material
is a promising, economical, and rapid alternative that
can also be used to identify the presence of Gd on skin,
hair, and other tissue samples (Chaturvedi et ah, 2010;
Lorch et ah, 2010), and trials using this method on
the skin of bats inoculated with various concentrations
of Gd showed 96% diagnostic sensitivity and 100%
diagnostic specificity (Lorch et ah, 2010). However,
although this technique can be used to less invasively
screen individual bats for the presence of Gd, small tis¬
sue samples usually collected for DNA extraction and
PCR amplification may not provide accurate results
because the fungus is not equally distributed across all
tissue surfaces (Lorch et ah, 2010). Additionally, the
use of PCR amplification and culture/isolation/exami¬
nation methods only allow for confirmed presence of
the fungus, which may occur in the absence of
ENCYCLOPEDIA OF CAVES
906
w
cutaneous lesions, and do not provide the ability to
determine the extent of Gd invasion into cutaneous tis¬
sues as compared with histopathological examination.
BAT SPECIES AFFECTED BY WNS
To date, six hibernating cave-roosting bat species have
experienced varied rates of mortality associated with
WNS, including the little brown myotis ( Myotis lucifu-
gus), northern long-eared myotis (M. septentrionalis) , the
federally listed endangered Indiana myotis (M. sodalis),
the small-footed myotis (M. leibii), the tricolored bat
(Perimyotis subflavus), and the big brown bat ( Eptesicus
fuscus ) (Blehert et al., 2009; Courtin et al., 2010; A.C. Hicks
and K.E. Langwig, pers. comm.). More recently, Gd, the
fungus associated with WNS, has also been identified
morphologically and genetically (using PCR) from the
southeastern myotis ( Myotis austroriparius), the federally
listed endangered gray myotis (M. grisescens), and cave
bat (M. velifer), although to date no other symptoms asso¬
ciated with WNS, including mass mortality, have been
observed in these three species.
CHARACTERISTICS OF WNS
Fungal Infections and Tissue Damage
The primary characteristic of WNS is the presence
of cutaneous lesions from Gd, where the invasion of
fungal hyphae causes cup-like epidermal erosions,
ulcers, and degradation of regional connective tissue
(Meteyer et al., 2009; Fig. 2). Invasion by Gd leads to
the replacement of cutaneous glands, blood, and lym¬
phatic vessels and muscle as the fungus digests tissues
and also causes a reduction in tone, tensile strength,
and elasticity of wing membranes (Cryan et al., 2010).
Additionally, tissue damage and necrosis, apparently
due to oxygen depletion (i.e., infarcted tissue), has
been observed in regions of wing membrane distant
from the area of fungal invasion (Cryan et al., 2010).
Given the importance of bat wings in many physiolog¬
ical processes, these observations suggest that a mas¬
sive homeostatic imbalance caused by the destruction
of wing tissue may be a possible proximate cause of
death in hibernating bats (Cryan et al., 2010).
Depleted Fat Reserves
WNS is also manifested in hibernating bats by an
apparent premature mobilization of fat reserves
(Blehert et al., 2009; Meteyer et al., 2009; Courtin et al.,
2010; J.D. Reichard, pers. comm.) suggesting that
starvation due to depleted fat reserves may also be a
FIGURE 2 Histological section of a wing membrane collected from a
hibernating little brown myotis ( Myotis lucifugus) affected by white-nose
syndrome immediately following euthanasia. Exuberant growth of
Ceomyces destructans and conidia (spores) are present on the skin surface
(arrow) as well as penetrating wing membrane (arrowheads) without
associated inflammation. PAS stain. Bar =15 |im. Photo by C.U. Meteyer,
U.S. Geological Survey; Meteyer et al., 2009. Used with permission.
proximate cause of death in bats affected by WNS.
Depletion in energy reserves may be due to the poten¬
tial homeostatic imbalance noted above or an initial
dearth of fat reserves in bats as they enter hibernation
possibly due to a lack of fatty acids in the diet because
of changes in insect abundance. It is also possible that
depleted fat reserves are the result of accelerated
energy consumption if bats are arousing from torpor
more frequently to groom the fungus from their skin,
or to mount immune responses. Arousing from torpor
and maintaining euthermic body temperatures at typi¬
cal hibernaculum temperatures are energetically expen¬
sive activities (Boyles and Willis, 2009). If normal
torpor-arousal patterns are disrupted during the devel¬
opment of WNS, increases in metabolic rates could eas¬
ily require more fat reserves than bats typically store in
preparation for hibernation and result in the observed
depletion of these reserves and death by starvation.
Atypical Winter Behavior
In addition to wing damage, depleted fat reserves,
and fungal infections, bats affected by WNS may
also exhibit atypical behavior including flying outside
hibernacula in midwinter (A.C. Hicks, pers. comm.)
and the inability to arouse from torpor when disturbed
by human activity (A.C. Hicks, J.G. Boyles, and
J.D. Reichard, pers. comm.). In fact, infrared thermal
imaging of clusters of hibernating little brown myotis
shows that many torpid bats affected by WNS do not
elevate their body temperatures upon anthropogenic
ENCYCLOPEDIA OF CAVES
WHITE-NOSE SYNDROME: A FUNGAL DISEASE OF NORTH AMERICAN HIBERNATING BATS
907
(A)
FIGURE 3 Wings of little brown myotis ( Myotis lucifugus ) unaf¬
fected by and affected by white-nose syndrome captured in early
spring from a maternity colony. (A) Undamaged wing (wing-damage
index 0); (B) wing showing moderate scarring and spotting (wing-
damage index 1+); (C) wing showing severe scarring and necrosis
(wing-damage index 3). Wing-damage index after Reichard and Kunz,
2009. Photo by N.W. Fuller, Boston University. Used with permission.
disturbance (J.D. Reichard, unpubl. data; J.G. Boyles,
unpubl. data).
Atypical behavior of hibernating bats infected with
Gd also appears to manifest increased frequencies of
arousal from torpor (D.M. Reeder, pers. comm.).
Because elevating body temperature (T b ) from near
ambient to ~37°C and sustaining this elevation of T b
during arousals is extremely energetically costly
(Boyles and Willis, 2009), this behavioral change is
likely to incur extreme costs in the form of fat depletion. It
is possible that increased frequencies of arousal from
torpor are due to irritation caused by the cutaneous fungal
infection, or to elevate immune responses in an attempt to
resist fungal invasion. It is also possible that the depletion
of fat reserves prompts bats to arouse from torpor to feed,
which is supported by the observations of bats emerging
from hibemacula in midwinter potentially in an attempt
to forage (Blehert et al., 2009; A.C. Hicks, pers. comm.; J.D.
Reichard, pers. comm.).
Changes in Immune Response during
Hibernation
As true hibernators, bat species affected by WNS
may experience levels of immunocompetence con¬
strained by prolonged periods of deep torpor, given
that research on other hibernating mammals suggests
that deep torpor depresses numerous mechanisms
associated with immunity (Bouma et al., 2010). Because
of attempted resistance against fungal invasion, the
immune function in hibernating bats affected with
WNS may be altered, with some aspects apparently
elevated and others reduced in comparison to unaf¬
fected bats (M.S. Moore, unpubl. data; R. Jacob and
D.M. Reeder, unpubl. data). Based on histopathological
examination, the little brown myotis does not appear
to mount a morphologically detectable inflammatory
response to Gd (Meteyer et al., 2009); however, the
extent to which WNS-affected bats attempt to resist
invasion by Gd through other immunological
responses remains unknown.
Wing Damage
Bats that survive hibernation while being affected
by WNS may return to maternity colonies with ulcer¬
ated, necrotic, and scarred wing membranes (Reichard
and Kunz, 2009; N.W. Fuller, pers. comm.). In fact,
severe wing damage was observed in the majority of
little brown myotis roosting in two New Hampshire
maternity colonies during the summer of 2008 (Fig. 3).
Damage to wings caused by wounds or infections can
impair the ability of wings to perform their functions
and may result in reduced foraging success and
increased vulnerability to predation. However, the
results from a recent study indicate that complete heal¬
ing from WNS-associated damage can occur in some
individuals during summer months (N.W. Fuller, pers.
comm.). If bats with damaged wings experience
reduced foraging success during the active season, fur¬
ther reductions may occur in the relative body condi¬
tion of bats arriving at maternity colonies in an already
compromised state. This reduction in body condition
may also result in an inability to properly care for
young, avoid predation, and enter the following
ENCYCLOPEDIA OF CAVES
908
w
hibernation period with sufficient fat reserves. In fact,
reduced body condition and lower recapture rates
have been observed in relation to WNS-induced wing
damage (Reichard and Kunz, 2009). In addition to the
massive mortality that has been observed during
hibernation, this physiological challenge incurred dur¬
ing the active season also has presumably contributed
to the approximately 73% decline in summer activity
of little brown myotis within regions affected by WNS
(Dzal et al., 2011; Brooks, 2011).
FUNGAL TRANSMISSION
The pattern of expansion away from its epicenter, the
known behavioral patterns of bats throughout the year,
and the fact that some gated caves and mines have had
no human visitors for several years prior to becoming
affected, support the hypothesis that bat-to-bat contact is
the primary mode of transmission. However, the possi¬
bility of anthropogenic transmission was acknowledged
when researchers first became aware that the majority of
newly affected sites in 2008 had been visited by humans,
either as biologists or recreational users (Turner and
Reeder, 2009). However, in 2009, bats infected with Gd
appeared in areas distant from other affected sites, sug¬
gesting that human transmission of the fungus was a
possibility. These infected sites have high recreational
use and small hibernating populations or may have
been visited by people with gear used at other affected
sites. Thus, it appears that while bat-to-bat transmission
facilitates the movement of Gd into new sites adjacent or
near to affected hibernacula, anthropogenic transmission
may accelerate the spread and introduce the fungus into
relatively distant sites and areas.
A EUROPEANCONNECTION?
As early as the 1980s, repeated observations of white
fungal growth on the muzzles of hibernating bats were
reported in Europe; however, no unusual mortality
events have been documented in association with these
reports (Martinkova et al., 2010; Wibbelt et al., 2010).
Since the advent of WNS in North America and the
identification of the associated fungus, Gd, in nine spe¬
cies of bats in Europe, sampled in Germany,
Switzerland, Hungary, the Czech Republic, Slovakia,
and France (Martinkova et al., 2010; Puechmaille et al.,
2010; Wibbelt et al., 2010), no mass mortality has been
associated with these isolations. Some researchers have
postulated that the fungus may be native to Europe or
at least that it predates its presence in North America
(Martinkova et al., 2010; Wibbelt et al., 2010).
Alternatively, it is also possible that the fungus had its
origin in North America and was recently introduced
to Europe. However, because considerable genetic vari¬
ability was observed in Gd isolates from the Czech
Republic and Slovakia (Martinkova et al., 2010), in con¬
trast to the identical clones found in New York
(Chaturvedi et al., 2010), these results suggest that the
fungus had its origin in Europe. One hypothesis is that
Gd caused mass mortality in European bats long before
human observations were recorded and that they
evolved resistance while being exposed to this fungus.
Different hibernation strategies may also affect the sus¬
ceptibility of various species to the syndrome. As sug¬
gested by Wibbelt et al. (2010), competition between Gd
and other microbial flora colonizing bat skin or roost
surfaces and soils in hibernacula in Europe may have
contributed to the evolution of a nonpathogenic form
in Europe, while a lack of similar competition in North
America may have led the evolution of a pathogenic
variant of Gd that affects hibernating bats.
CURRENT AND FUTURE RESEARCH
A multicontinent research effort has been mobilized
to study the effects of WNS on North American and
European bats, and includes researchers from numer¬
ous academic, governmental and nonprofit institu¬
tions. One of the primary areas of focus is in
determining if the fungus is the cause of mortality or
is a secondary symptom. Controlling the spread and
mitigating the effects of WNS are also high priority
areas of research. As an example, based on modeling
efforts, Boyles and Willis (2009) suggested that local¬
ized "thermal refugia," or warm areas inside hibernac¬
ula, could reduce heat loss and energy expenditure
during periodic arousals, possibly providing a stop¬
gap measure to employ while other control measures
are being developed. However, this unconventional
practice would require that bats somehow could detect
and travel to refugia and that these practices would
not alter conditions within the rest of hibernacula
where refugia are installed. Culling has also been sug¬
gested as a possible means to reduce transmission
(Arnold Air Force Base, 2009), although Hallam and
McCracken (2011) modeled the effects of culling on
transmission cycles and concluded that this mitigation
practice will have little effect on the spread of the syn¬
drome. Some groups are also conducting trials to test
the ability of fungicides to eradicate Gd from contami¬
nated gear and the efficacy of compounds to treat
affected bats (H. Barton, pers. comm.; D.M. Reeder,
pers. comm.; A.H. Robbins, pers. comm.).
Additional studies initiated shortly after the emer¬
gence of the syndrome include investigations into pat¬
terns of arousal and torpor, thermoregulatory changes.
ENCYCLOPEDIA OF CAVES
WHITE-NOSE SYNDROME: A FUNGAL DISEASE OF NORTH AMERICAN HIBERNATING BATS
909
variation in metabolic rates, behavioral changes,
determination of variation in body condition and types
of fat throughout the hibernation period, immunologi¬
cal correlates, and examination of the bacterial flora of
digestive systems in WNS-affected bats (Turner and
Reeder, 2009). The possible role of environmental con¬
taminants has been investigated to a degree, showing
high levels of persistent organic pollutants (Kannan
et al., 2010), lead, and arsenic (Courtin et al., 2010) in
WNS-affected bats. However, high levels of organic
pollutants were also found in bats hibernating in unaf¬
fected hibernacula (Kannan et al., 2010), making the
interpretation of these results ambiguous. Primers
developed for the detection of Gd on bat tissues
have been tested on environmental samples to better
understand the geographic distribution of this newly
described fungus, although high levels of cross-reaction
with other species of fungi have prevented this method
from being used to successfully identify Gd in samples
other than those directly collected from bat tissue
(Lorch et al., 2010). Modeling efforts have compared
demographics in the affected little brown myotis to
predict the risk of extinction as a result of mortality
caused by WNS (Frick et al., 2010), and demographic
models are being developed to help explain the dif¬
ferential rates of mortality among affected species
(K.E. Langwig, pers. comm.; A.P. Wilder, pers. comm.).
More recently, researchers have begun to investigate
additional components important to the development
of WNS. Quantitative PCR methods are being devel¬
oped to determine fungal load on affected bats.
Investigations are being conducted into the pathogene¬
sis, microbial ecology and phylogeography of the fun¬
gus, including sequencing the genome of Gd (J.T.
Foster, pers. comm.). Population genetic structure and
gene flow in the little brown myotis is being evaluated
to help predict the further spread of the syndrome
(C.M. Miller-Buttersworth, pers. comm.; A.P. Wilder,
pers. comm.). Studies are also under way to compare
the North American and European strains of the fun¬
gus and differences in susceptibility between North
American and European species.
CAVE CLOSURES AND
DECONTAMINATION PROTOCOLS
Because of the likelihood that human visitations facili¬
tate the transmission of Gd to unaffected sites, a number
of states have closed caves on state-owned property and
the U.S. Fish and Wildlife Service (USFWS) recommends
that all cavers observe cave closures, advisories, and
decontamination protocols. Decontamination procedures
and a list of current cave closures can be found at www
.fws.gov/WhiteNoseS 5 mdrome/cavers.html. For the most
current list of WNS-affected and adjacent states, visit
www.fws.gov/WhiteNoseSyndrome/. Most importantly,
bats should not be handled by people unless authorized
to do so. If live or dead bats are encountered in caves
and mines with the appearance of WNS, contact the
nearest state wildlife agency, USFWS Ecological
Services Field Office, or email WhiteNoseBats@fws.gov.
State office listings can be found at www.fws.gov/
offices/statelinks.html. USFWS office listings can be
found at: www.fws.gov/offices/.
Acknowledgments
We wish to thank H.A. Barton, J.G. Boyles, N.W. Fuller, A.C. Hicks, R.
Jacob, K.E. Langwig, D.M. Reeder, J.D. Reichard, A.H. Robbins, and
A.P. Wilder for sharing unpublished information referenced in this
chapter. We are grateful to the American Society of Mammalogists,
Bat Conservation International, the Eppley Foundation for Research,
National Science Foundation, Morris Animal Foundation, National
Speleological Society, U.S. Fish and Wildlife Service, and the Woodtiger
Fund for supporting our research on WNS.
See Also the Following Article
Bats
Bibliography
Arnold Air Force Base (2009). Wlute nose syndrome cooperative monitor¬
ing and response plan for Tennessee 2009. Tullahoma, TN: Arnold
Air Force Base.
Blehert, D. S., Hicks, A. C., Behr, M., Meteyer, C. U., Berlowski-Zier,
B. M., Buckles, E. L., et al. (2009). Bat white-nose syndrome: An
emerging fungal pathogen? Science, 323(5911), 227.
Bouma, H. R., Carey, H. V., & Kroese, F. G. M. (2010). Hibernation: The
immune system at rest? Journal of Leukocyte Biology, 88, 619—624.
Boyles, J. G., & Willis, C. K. R. (2009). Could localized warm areas inside
cold caves reduce mortality of hibernating bats affected by white-
nose syndrome? Frontiers of Ecology and the Environment, 8, 92—98.
Brooks, R. T. (2011). Declines in summer bat activity in central New
England 4 years following the initial detection of white-nose syn¬
drome. Biodiversity and Conservation, 20(1), 2537—2541.
doi:10.1007/sl0531-011-9996-0.
Chaturvedi, V., Springer, D. J., Behr, M. J., Ramani, R., Li, X., Peck,
M. K., et al. (2010). Morphological and molecular characteriza¬
tions of psychrophilic fungus Geomyces destructans from New
York bats with white nose syndrome (WNS). PLoS ONE, 5(5),
el0783. doi:10.1371/journal.pone.0010783.
Courtin, F., Stone, W. B., Risatti, G., Gilbert, K., & Van Kruiningen,
H. J. (2010). Pathogenic findings and liver elements in hibernating
bats with white-nose syndrome. Veterinary Pathology, 47, 214—219.
Cryan, P. M., Meteyer, C. U., Boyles, J. G., & Blehert, D. S. (2010).
Wing pathology of white-nose syndrome in bats suggests life-
threatening disruption of physiology. BMC Biology, 8, 135.
Dzal, Y., McGuire, L. P., Veselka, N., & Fenton, M. B. (2011). Going,
going, gone: The impact of white-nose syndrome on the summer
activity of the little brown bat ( Myotis lucifugus). Biology Letters, 7
(3), 392-394. doi:10.1098/rsbl.2010.0859.
Frick, W. F., Pollock, J. F., Hicks, A. C., Langwig, K. E., Reynolds,
D. S., Turner, G. G., et al. (2010). An emerging disease causes
ENCYCLOPEDIA OF CAVES
910
w
regional population collapse of a common North American bat
species. Science, 329(5992), 679—682.
Gargas, A., Trest, M. T., Christensen, M., Volk, T. J., & Blehert, D. S.
(2009). Geomyces destructans sp. nov. associated with bat white-
nose syndrome. Mycotaxon, 108, 147—154.
Hallam, T. G., & McCracken, G. F. (2011). Management of the pan¬
zootic white-nose syndrome through culling of bats. Conservation
Biology, 25, 189-194.
Kannan, K., Yun, S. H., Rudd, R. J., & Behr, M. J. (2010). High con¬
centrations of persistent organic pollutants including PCBs, DDT,
PBDEs and PFOS in little brown bats with white-nose syndrome
in New York, U.S.A. Chemosphere, 80, 613—618.
Kunz, T. H., & Fenton, M. B. (2003). Bat ecology. Chicago, IL:
University of Chicago Press.
Lorch, J. M., Gargas, A., Meteyer, C. U., Berlowski-Zier, B. M.,
Green, D. E., Shearn-Bochsler, V., et al. (2010). Rapid polymerase
chain reaction diagnosis of white-nose syndrome in bats. Journal
of Veterinary Diagnostic Investigations, 22, 224—230.
Martinkova, N., Backor, P., Bartonicka, T., Blazkova, P., Cerveny, J.,
Falteisek, L., et al. (2010). Increasing incidence of Geomyces
destructans fungus in bats from the Czech Republic and Slovakia.
PLoS ONE, 5(11), el3853. doi:10.1098/rsbl.2010.0859.
Meteyer, C. U., Buckles, E. L., Blehert, D. S., Hicks, A. C.,
Green, D. E., Shearn-Bochsler, V., et al. (2009). Histopathologic
criteria to confirm white-nose syndrome in bats. Journal of
Veterinary Diagnostic Investigations, 21, 411—414.
Puechmaille, S. J., Verdeyroux, P., Fuller, H., Gouilh, M. A.,
Bekaert, M., & Teeling, E. C. (2010). White-nose syndrome fun¬
gus (Geomyces destructans ) in bats, France. Emerging Infectious
Diseases, 16, 290—293.
Reichard, J. D., & Kunz, T. H. (2009). White-nose syndrome inflicts
lasting injuries to the wings of little brown myotis (Myotis lucifu-
gus). Acta Chiroptera, 11, 457—464.
Turner, G. G., & Reeder, D. M. (2009). Update of white-nose syn¬
drome in bats, September 2009. Bat Research News, 50, 47—53.
United States Fish & Wildlife Service (USFWS). (2011). White-nose
syndrome: Something is killing our bats, www.fws.gov/white-
nosesyndrome/. Accessed 02.11.11.
Wibbelt, G., Kurth, A., Hellmann, D., Weishaar, M., Barlow, A., Veith,
M., et al. (2010). White-nose syndrome fungus (Geomyces destruc¬
tans) in bats, Europe. Emerging Infectious Diseases, 16, 1237—1242.
WORMS
Elzbieta Dumnicka
Institute of Nature Conservation, Polish Academy of Sciences
GENERAL CHARACTERISTICS
Several types (phyla) of small invertebrates are usually
included under the common term worms: flatworms
(Turbellaria), roundworms (Nematoda), ribbon worms
(Nemertea), and annelids (Annelida). Sometimes the
wormlike Onychophora are also included. The shape
of worms is rather simple: flat, filiform, or tube-like.
The body is usually small: from half a millimeter to
2—3 cm, exceptionally more. Nevertheless true
"giants" can be found among worms—some reaching
lengths in the meters. The majority of worm species do
not show morphological changes connected with life
in the cave environment, because they are usually
unpigmented and blind. Only in flatworms, leeches,
and a few ribbon worms, which are normally occu-
lated and pigmented, do these changes occur. The
body coloration is determined principally by the con¬
tents of the digestive tube and by blood pigments (as
in annelids). Features characteristic of typical cave
fauna other than morphological ones have been
observed, including aperiodicity of breeding and pro¬
longation of all stages of the life cycle (Dumnicka,
1984).
The majority of worms living in caves are the same
species that live in the soil or in marine or freshwater
environments. Nevertheless, in all phyla there are also
species found exclusively in various kinds of under¬
ground habitats (including caves), and they are classified
as troglobionts (terrestrial forms) or stygobionts (aquatic
forms). The distribution of particular stygobiotic species
is often wider than that of species found only in caves
since stygobionts can live in various types of under¬
ground waters (e.g., interstitial waters, hyporheic waters
of rivers, etc.; (Botosaneanu, 1986). They can even migrate
from the drainage area of one river to the other—due to
so-called stream capture (or stream piracy )—via an under¬
ground connection between streams from different river
catchments. Stygobionts also live in regions without
caves; in such a situation they inhabit fissures and small
spaces between the grains of sediments filled with water.
Their presence in underground waters of such areas is
noticed during studies made in wells or springs, where
they have been caught after being washed out from
underground waters following heavy rains.
HISTORY OF STUDIES
Though studies on cave worms started many years
ago (e.g., Jeannel, 1926; Hyman, 1937), there is still a
lot that is unknown. World data on the distribution,
biology, and ecology of various taxonomic groups
were summarized in Encyclopaedia Biospe'ologica
(Juberthie and Decu, 1994, 1998) and Stygofauna Mundi
(Botosaneanu, 1986) (for water fauna). From the
United States, Culver et al. (2000) listed 927 species liv¬
ing obligatory in caves, but Nematoda were omitted
from this list, whereas other worms were represented
by only a small number of species: Tricladida, 24 spe¬
cies; Oligochaeta, 5 species; and Branchiobdellida, 7
species. From that time a few new stygobiotic worms
were described in this region but more studies of this
group are needed. Since the publication of those books,
many new species or even higher taxa have been
described from other regions, especially Europe. For
example, 66 stygobiotic oligochaete species were cited
Encyclopedia of Caves.
© 2012 Elsevier Inc. All rights reserved.
WORMS
911
in Stygofauna Mundi (1986) but that number is now
more than 130. Recently, a special European research
project (PASCALIS—Protocols for the Assessment and
Conservation of Aquatic Life in the Subsurface) was
carried out in five Mediterranean countries and
Belgium that allowed the finding of many new species
as well as to elucidate the origin of some stygobiotic
taxa. Moreover, in the studied karstic regions, map¬
ping of biodiversity was performed to aid in the choice
of the best conservation strategies. Thus the majority
of published information on the number and distribu¬
tion of troglobiotic and stygobiotic species is incom¬
plete. According to our present state of knowledge,
troglobiotic and stygobiotic worms are most numerous
in Europe and North America, but these are also the
continents with the most intensive biospeleological
studies. Studies of worms in karstic regions of
Malaysia and South America started only recently, and
their results should increase our knowledge about this
group. Recently in Australia, the most interesting dis¬
coveries of the underground fauna were made in vari¬
ous regions, mainly in the western part of the country
(Austin et al., 2008). Troglobiotic and stygobiotic fauna
inhabiting small fissures up to 100 m deep were sam¬
pled from boreholes but worm fauna in this country
has been little studied.
BRIEF CHARACTERIZATION
OF INVERTEBRATE TYPES
CONSIDERED WORMS
Phylum: Turbellaria (Flatworms)
The body of flatworms is elongated with a delicately
ciliated surface, with only one alimentary canal opening
(no anal opening). Representatives of two orders (dis¬
cussed next) have been found in cave waters, whereas
so-called microturbellaria (small representatives of vari¬
ous turbellarian groups) are known predominantly
from interstitial waters.
Class: Tricladida
The Tricladida are almost exclusively aquatic forms
feeding on living or dead animals. The triclad engulfs
food particles extruding a long and muscular pharynx
through its mouth (Fig. 1) and sucks up the food into
the gastrovascular cavity. These invertebrates are very
resistant to lack of food—the starvation period can
reach some months. Morphological adaptations to cave
life are clearly visible in this group of worms, expressed
in depigmentation of the body and disappearance or
reduction of the eyes. They have been found mainly in
Europe and North America, but recent discoveries
have been made in Tahiti, Mexico, and Australia. Of
| 500 pm
FIGURE 1 Dinizia sanctaehelenae (Tricladida). From Botosaneanu,
1986.
FIGURE 2 The Temnocephalid, an external parasite of crustaceans.
the approximately 400 triclad species currently known,
about 150 are stygobionts. In recent systematics of this
group, a suborder named Tricladida Cavernicola was
separated (Sluys, 1990).
Class: Temnocephalida
The Temnocephalida are external parasites of crus¬
taceans (Decapoda and Amphipoda), attaching to the
host's gills, antennae, or legs (Fig. 2). The body is sim¬
ple, with one or more (up to six) acetabula allowing
for attachment to the host animal (Fig. 3). They pro¬
duce a viscous substance to better attach and glue their
eggs, from which small specimens, similar to adult
forms, hatch. Those worms feed mainly on secretions
or body fluids of the host; only the predatory genus
Bubalocerus uses crustaceans as a means of dispersal.
ENCYCLOPEDIA OF CAVES
912
w
FIGURE 3 Troglocaridicola capreolaria (Temnocephalida). Redrawn
from Juberthie and Decu, 1994.
So far only 15 species of these highly specialized
worms are known from caves in southern Europe, but
one undetermined species was found on a cave crab in
Papua New Guinea.
Phylum: Nematoda (Roundworms)
Roundworms show high homogeneity in their mor¬
phology and anatomy. The body is filiform, usually
small in free-living forms (length 0.2—3 mm), pointed
at one or both ends, covered by smooth chitin cuticle,
and without appendages (Fig. 4). This very large group
(about 100,000 known species) lives in all habitats: ter¬
restrial and aquatic (sea bottom and all kinds of fresh¬
water), and as parasites on plants, animals, and
humans. The parasitic species living in cultivated
plants or breeding animals are well known whereas
the free-living nematodes are poorly known. Many
free-living species are eurytopic; such species can be
found in different environments (terrestrial or aquatic)
as well as in various climatic zones. Their high resis¬
tance to unfavorable life conditions (the eggs are resis¬
tant to desiccation and extreme temperatures) and
ability to form cysts means that they can be easily
transported for long distances by wind or water. That
is why common species of nematodes are abundant
in cave environments where they feed on bacteria,
vegetal debris, and bat guano; some of them are pre¬
dators. Among nematodes the troglobiotic species are
extremely rare; only about 20 are known.
Phylum: Nemertea (Ribbon and Proboscis
Worms)
In Nemertea, the body is unsegmented, strongly elon¬
gated, slender, soft, and usually without appendages
(Fig. 5). A retractable, muscular proboscis apparatus
at the anterior body end is characteristic of these worms
only. It is used for catching the prey—almost all ribbon
worms are predators or scavengers. The length of the
majority of species does not exceed a few centimeters
but the longest marine species reach 30 m. Almost all
species of ribbon worms (about 1200) live in marine
environments (Sundberg and Gibson, 2008), but there
are some exceptions: about 20 freshwater species are
known and among them 3 or 4 are blind, depigmented
inhabitants of various kinds of subterranean waters.
Their dimensions are smaller than these of epigean spe¬
cies. Stygobiotic species have been found mainly in
Europe (France, Austria, Switzerland, Germany) but
one species has been described from New Zealand also.
Moreover some marine species inhabiting deep layers of
sandy beaches could be treated as stygobionts, but the
morphological adaptations for subterranean life have
not been observed in them.
Phylum: Annelida
The Annelida are tube-like invertebrates with a regu¬
larly segmented body. Usually they are bilaterally sym¬
metrical, elongated, mainly hermaphroditic (bisexual)
with gonads situated in some segments only.
Class: Polychaeta
Segments have muscular, paired, laterally situated
"pseudolegs" (parapodia) with setae on their ends.
ENCYCLOPEDIA OF CAVES
WORMS
913
FIGURE 5 Ototyphonemertes sp. (Nemertina). From Botosanearm,
1986.
Polychaeta are an almost exclusively marine group
represented by about 9000 known species, but the
degree of knowledge of this group is insufficient.
Many polychaetes are sessile forms, living in tubes;
others are mobile and move actively. Numerous
marine species live in conditions similar to cave condi¬
tions: inside the bottom material, under stones, and in
the rock fissures. In submarine (anchihaline) caves,
stygobiotic forms very tolerant to changes of environ¬
mental conditions were found (Brito et al., 2009). In
Central and South America rich polychaete fauna
was found in a special type of sinkhole, cenotes, which
connect with the ocean by flooded cave corridors.
Nevertheless, due to their lack of morphological adap¬
tations to cave environment, only on the basis of their
habitat can they be classified as stygobionts.
About 2% of all polychaete species live in freshwa¬
ter and some curious stygobionts are known from this
habitat. The most famous one is Marifugia cavatica
(Fig. 6), known from caves situated in Dinaric karst
(northeastern Italy, Slovenia, Croatia, and Bosnia and
Herzegovina) but occurring in small, isolated areas
(Kupriyanova et al., 2009). Even though some locali¬
ties are situated a few kilometers from the seashore,
this species has never been found in brackish or
saltwater. Individuals live in calcareous tubes which
are attached to the walls, roofs, and bottoms of water-
filled corridors forming colonies of various dimen¬
sions. In recent years live animals were collected in a
few caves only, whereas aggregations of their broken
tubes were found in many localities (Kupriyanova
et al., 2009). Other freshwater stygobiotic polychaetes
belong to the genus Namanereis. Namanereis cavernicola
is known from Mexican and Peruvian caves, while
N. tiriteae from New Zealand and Fiji (Gray et al., 2009)
lives in interstitial waters. The disjunct distribution of
FIGURE 6 Marifugia cavatica (Polychaeta): (A) Anterior part of the
body (from Botosaneanu, 1986). (B) Habitus of a female individual
(from Juberthie and Decu, 1998).
these species according to the Gondwana continent
indicates their ancient origin (Glasby and Timm, 2008).
Class: Oligochaeta
All segments (except for the first and last ones) of
Oligochaeta have four bundles of setae (Fig. 7). They
live in both aquatic and terrestrial (earthworms) envir¬
onments. About 5000 species have been described, with
only 130 from caves and subterranean waters. The
smallest ones (Fig. 8) live among the grains of sand
and gravel saturated with water (interstitial waters) at
the seashore as well as along streams and in stagnant
water bodies in caves. The bigger ones "walk" on the
bottom surface or swim using setae as locomotory
organs. The representatives of naidids and lumbricu-
lids generally are "anchored" inside the sediments with
the anterior part of the body, waving the posterior end
for better oxygenation of the surrounding water. They
move slowly when the food is used up or environmen¬
tal conditions deteriorate, which is when oxygenation
decreases or products of metabolism accumulate. Like
nematodes, some stygobiotic oligochaete species can
form cysts, which help them to survive in small, peri¬
odically desiccated cave pools. These invertebrates are
bisexual. Normally cross-fertilization occurs but self-
fertilization is also known. The majority of oligochaetes
feeds on organic matter from sediments or soil. In the
cave environment, organic matter usually forms only a
small part of the sediment, so oligochaetes ingest big
ENCYCLOPEDIA OF CAVES
914
w
500 jum
(C)
FIGURE 7 Representatives of cave oligochaetes: (A) Aeolosoma sp.
(B) Nais sp.; (C) Haber turquini ( from Botosaneanu, 1986).
FIGURE 8 Gianius aquedulcis (Oligochaeta), anterior segments.
amounts of substrate but utilize only a small part of it.
Some species feeding on algae in surface waters can
change their food preference in subterranean waters
and eat small particles of organic matter. Only a few
stygobiotic species are predators. Stygobiotic crayfish-
worms (Branchiobdellida) are also known from caves.
Species of this group (classified as oligochaetes or
leeches, or even a separate class among annelids) are
usually ectoparasites of crustaceans, some of them
(similarly to flatworms Temnocephalida) were found
on 1—3 host species only. Troglobiotic and stygobiotic
oligochaetes are known from the caves of all continents
(except for Antarctica), but the most diversified fauna
have been found in the Mediterranean region (Creuze
des Chatelliers et al., 2009).
Class: Hirudinea (Leeches)
The Hirudinea have segments without parapodia and
setae. All leeches have two suckers: usually a small one
around the mouth and a big one at the end of the body
(Fig. 9). About 500 species of leeches are known; they
live principally in freshwater, but some are terrestrial
and these live in humid habitats. Usually the species
from surface waters are found in caves, but they can
form separated cave populations, depigmented to vari¬
ous degrees, and with eyes partially or totally reduced
(e.g., Dina sp.) (Fig. 9A). An unusual species of leech,
Croatobranchus mestrovi, was found in deep caves in
Croatia (Sket et al., 2001). This species has head morphol¬
ogy not found in other Hirudinea. The rim surrounding
the oral sucker forms four pairs of conical lobes, with
five finger-shaped papillae on each lobe. Their role is
not known yet. Moreover, the posterior half of the body
has 10 or more pairs of lateral processes (Fig. 9B) not
serving as locomotory organs, but most probably as
respiratory organs. This species lives in deep vertical
1 cm
FIGURE 9 Representatives of cave leeches: (A) Dina sp. ( from
Botosaneanu, 1986). (B) Croatobranchus mestrovi, ventral side ( from Sket
et al, 2001).
ENCYCLOPEDIA OF CAVES
WORMS
915
FIGURE 10 Speleoperipatus speleus (Onychophora). Redrawn from
Juberthie and Decu, 1994.
caves, usually near their bottom, where temperatures
are low and constant (4—6°C). Most leeches were seen
on the rocky walls in a thin layer of flowing water; rarely
they can be found on the bottom of cave pools. All cave
leeches are predators, feeding on other small inverte¬
brates; bloodsuckers have not been found in subter¬
ranean waters. Leeches are known from caves of the
northern hemisphere.
Phylum: Onychophora
Onychophora are worm-like, predatory invertebrates
with one pair of antennae and 13—43 pairs of locomo-
tory appendages (Fig. 10). This is a small group (about
100 species) of "living fossil" terrestrial animals, which
are very interesting from the systematics and phyloge¬
netic point of view. They are found in tropical and
subtropical regions but in very restricted areas. Cave
species are known from the Southern Hemisphere
except for Speleoperipatus speleus, which was found in
caves in Jamaica.
ECOLOGICAL AND PHYLOGENETIC
REMARKS
As mentioned above, among worms living in caves
there are both common, widely distributed species and
highly specialized forms known exclusively from very
restricted areas or from only a few localities. This first
group is not very interesting to biologists working on
the systematics and evolution of cave fauna, but their
role in the functioning of cave biocenoses is usually
important. Very often detritivorous, small worm spe¬
cies occur at high densities and sometimes their
biomass is also high so they constitute a good source
of food for predators.
Troglobiotic and stygobiotic worm species origi¬
nated from soil, freshwater, and marine forms. Their
ancestors colonized the underground environment in
various geological epochs and in various places. Some
of them are very old; for example, Onychophora
formed in the Paleozoic. Among the oligochaetes
there are cave forms at various systematic levels,
from the exclusively troglobiotic or stygobiotic families
(Ocnerodrilidae in South America, Dorydrilidae in a
restricted area of Europe, and Parvidrilidae from
North American and European caves) and genera
( Delaya , Krenedrilus), and genera with mainly subterra¬
nean species ( Rhyacodrilus , Trichodrilus), to genera with
only a few or single stygobiotic species. Troglobiotic or
stygobiotic species belong to almost all oligochaete
families of various origins (marine, freshwater, and
terrestrial) and of various ages—both old and recent.
This tells us that multiple migrations to the subterra¬
nean environment took place at various times and
places whereas in other groups such as Temnocephalida
and Hirudinea the settlement in caves was an excep¬
tional event. Among stygobiotic worms many species
seem to be endemic to one cave or one karstic region,
but with advances in research sometimes it turns out
that the distribution of some species previously consid¬
ered as endemic is actually quite—for example, an oli¬
gochaete Gianius aquedulcis that was originally known
only from interstitial waters of Weser River in Germany
was recently found in some countries of southern
Europe, Poland, and even in the United States.
Bibliography
Austin, A. D., Cooper, S. J. B., & Humphreys, W. F. (Eds.), (2008).
Subterranean connections: Biology and evolution in troglobiont
and groundwater ecosystems. Invertebrate Systematics, 22, 1—310.
Botosaneanu, L. (Ed.), (1986). Stygofauna mundi. A faunistic, distribu¬
tional, and ecological synthesis of the world fauna inhabiting subterra¬
nean waters (including the marine interstitial). Leiden, The
Netherlands: E. J. Brill.
Brito, M. C., Martinez, A., & Nunez, J. (2009). Changes in the stygo-
biont polychaete community of the Jameos del Agua, Lanzarote,
as a result of bioturbation by the echiurid Bonellia viridis. Marine
Biodiversity, 39(3), 183—187.
Creuze des Chatelliers, M., Juget, J., Lafont, M., & Martin, P. (2009).
Subterranean aquatic Oligochaeta. Freshwater Biology, 54(4),
678-690.
Culver, D. C., Master, L. L., Christman, M. C., & Hobbs, H. H.
(2000). Obligate cave fauna of the 48 contiguous United States.
Conservation Biology, 14(2), 386—401.
Dumnicka, E. (1984). Laboratory studies on biology of cave-dwelling
Enchytraeidae: Enchytraeus dominicae Dumnicka and Fridericia bul-
bosa (Rosa). Me'moires de Biospe'ologie, 11, 199—205.
Glasby, C. J., & Timm, T. (2008). Global diversity of polychaetes
(Polychaeta; Annelida) in freshwater. Hydrobiologia, 595(1),
107-115.
ENCYCLOPEDIA OF CAVES
916
w
Gray, D. P., Harding, J. S., & Winterbourn. M. J. (2009). Namanereis
tiriteae, New Zealand's freshwater polychaete: New distribution
records and review of biology, www.biol.canterbury.ac.nz/ferg/
people/Duncan/Natural%20history/Polychaete%20paper.pdf.
Hyman, L. H. (1937). Studies on triclad Turbellaria, VIII. Some cave
planarians of the United States. Transactions of the American
Microscopal Society, 44, 51—89.
Jeannel, R. (1926). Faune Cavernicoie de la France, avec une Etude des
Conditions d'Existance dans le Domain Souterrain. Paris: Lechevallier.
Juberthie, C., & Decu, V. (Eds.), (1994). Encyclopaedia biospeologica
(Vol. I). Moulis, France: Societe de Biospeologie (in French).
Juberthie, C., & Decu, V. (Eds.), (1998). Encyclopaedia biospeologica
(Vol. II). Moulis, France: Societe de Biospeologie (in French).
Kupriyanova, E. K., ten Hove, H. A., Sket, B., Zaksek, V., Trontelj, P., &
Rouse, G. W. (2009). Evolution of the unique freshwater cave-dwelling
tube worm Marifugia cavatica (Annelida: Serpulidae). Systematics
and Biodiversity, 7(4), 389—401.
Sket, B., Dove, P., Jalzic, B., Kerovec, M., Kucinic, M., & Trontelj, P.
(2001). A cave leech (Hirudinea, Erpobdellidae) from Croatia
with unique morphological features. Zoologica Scripta, 30(3),
223-229.
Sluys, R. (1990). A monograph of the Dimarcusidae (Platyhelminthes,
Seriata, Tricladida). Zoologica Scripta, 29(1), 13—29.
Sundberg, P., & Gibson, R. (2008). Global diversity of nemerteans
(Nemertea) in freshwater. Hydrobiologia, 595, 61—66.
ENCYCLOPEDIA OF CAVES
Index
A
A2, 703
Abandoned buildings, 643
Abandonment suite, 137
Acari, 793-795
Acaridae, 795
Acarina, 214
Accelerated corrosion, 144
Accidentals, 276—277
Acid mine drainage (AMD), 161—162
Actinomycete colonies, 494, 497f
Activated carbon passive detector, 890
Active colonization model, 275, 405
Active dispersal models
of invasion, 405
active colonization model, 405
adaptive-shift model, 405
climatic-relict model, 405
competitive pressure, 405
deep sea colonization, 405
epigean limnic system model, 405
escape from epigean predators, 405
marine shallow-water ecotones
model, 405
Adaptation. See also Behavioral adaptations
to cave environment, 341
Cavefish of China, 113—121
adipose storage, 117,118f, 119f
body shape changes, 108f, 113—114,
113f
eye degeneration, 114f, 119f,
117-120, 121f
humpback and horn, 114—116, 114f,
115f
pigmentation loss, 108f, 120
scale disappearance, 120—121
sensory apparatus, 113f, 116—117,
116f
specialized appendages, 108f, 117
Gammarus minus
cave population origin, 345, 346f
cave streams and, 344
compound eyes and, 345—346, 347f
foods of, 343-344
general ecology of, 343-344
idealized subterranean drainage
basin and, 343, 343f
karst window and, 344—345
as model system, 347
population density of, 343—344
population structure of, 345
reproduction, 343—344
springs and, 344
variation in, 344—345, 344f
to low food
energy demand and, 4—6
factors concerning, 2f
food resources and, 1 —3
food scarcity types and, 1
food-finding and, 3—4
life history characters and, 6—8
overview, 1, 2f
periodic starvation and, 8—9
mechanism hypotheses, 341—342
directly adaptive hypothesis, 342
indirectly adaptive hypothesis, 342
nonadaptive hypothesis, 341—342
model organisms and, 37
preadaptation, 10
regressive evolution and, 341
study criteria, 342
Adaptive shift
case studies, 12—16
cixiid planthoppers, 13—14, 13f
continents, 16
crickets, 15
Hawaii, 12—16, 12t
moths, 15—16
other islands, 16
terrestrial Isopoda, 14—15
wolf spiders, 15
conclusions about, 16—17
factors underlying, 10—12
ancestral habitats, 12
cavernicolous habitats, 11
environmental stress, 12
extrinsic, 11—12
food resources, 11—12
founder events, 10—11
genetic repertoire, 10
intrinsic, 10—11
mating behavior and hybridization,
11
preadaptation, 10
stress response, 11
future regarding, 17
theory of, 9—10
Adaptive-shift model, 405
Aeolian caves, 105f, 106
Aerobes, 490f, 491
Agagonite, 777
Agelenidae, 789
Aggradation, 136f, 137
Agoristenidae, 796
Ailuropoda, 594
Akravidae, 787
Allochthonous sediments, 135
Allogenic recharge, 387
Allophane, 5011, 505
Alluviated springs, 799, 799f
Alpine
karrenfields, 430
karst, 436-437
Alpinotype caves, 724
Alunite, 5011, 503
Amarnath Cave, 322
Amblyopsidae, 522-523, 522f, 525, 525f
Amblyopsis rosae
life history characters of, 7—8, 8t
life history evolution of, 465, 466, 466t,
467-468
Amblyopsis spelaea, 465, 466, 466t, 467—468
Amblypygi, 214, 789, 789f
AMD. See Acid mine drainage
Amensalism, 743, 744f
Ammonia, chemoautotrophy and, 129
Amphibia, 260, 261t
Amphibian chytrid fungus, 624
Amphipoda, 178t, 186, 187t, 257
anchihaline habitats and, 22
Appalachian stream competition and,
746- 747, 746f, 747f
in Australia, 207
in Dinaric karst, 233
AMT. See Audio-magnetotelluric soundings
Anaerobes, 490f, 491
Anapidae, 790
Anaspidacea, 178t, 184,184f
Anastamose, 422, 423f
Anastomose, 422
Anastomotic caves, 395-396, 637, 638f
Anastomotic maze, 602f
Anchialine pools, 18
Anchihaline and marine caves, 240
Anchihaline habitats
in Australia, 209, 210, 214f
biogeography, 22-23
biology and inhabitant distribution,
23-24
cave morphology and hydrology, 18—20
horizontal caves, 18
U-shaped tube effect, 19
vertical caves, 18—19
Crustacea, 20-22, 209, 210, 214f
Amphipoda, 22
Copepoda, 21
Decapoda, 21
Isopoda, 22
Leptostracans, 21
Mictaceans, 22
Mysids, 21-22
Niphargus , 22, 22f
917
918
INDEX
Anchihaline habitats ( Continued )
Ostracods, 21
Remipedia, 21
Thermosbaenaceans, 21
fauna and humans, 25
geography, 18
noncrustacean groups, 20
filter feeders, 20
fishes, 20
gastropods, 20
sessile chaetognaths, 20
sponges, 20
overview about, 17
trophic relations, 24
Ancient cavers, eastern North America
archaeology and, 26
conclusions about, 30
Mammoth Cave
archaeological evidence, 27—28, 27f,
28f
archaeology, 26-27
chronology and, 27
evidence interpretation, 28—30, 29f
prehistoric archaeology, 27—30
overview about, 26
Androniscus dentiger, 615
Angel hair, 375
Angular passages, 604-605, 604f
Anhydrite, 374
Annelida, 911—914
Hirudinea, 913-914, 913f
Oligochaeta, 912-913, 913f
Polychaeta, 911-912, 912f
Anomura, 178t, 192
Anophthalm beetles, 56—57, 56f
Anoxia survival times, 653
Anthodite Hall, 715f
Anthodites, 379-380, 380f, 782
Anthuridea, 178t, 188, 189t
Antroherpon scutulatum, 519f
Antrolana lira, 854—855, 854f
Anvil Cave, 739, 739f
Aphaencips Pluto, 519f
Aphenopsian beetles, 56—57, 56f
Appalachian Mountains, 734f, 739— 740, 740f,
741f
Appalachian Plateaus, 734f, 738—739, 739f
Aqua Cave, 74, 77
Aquatic cave animals, energy demand and,
4, 4t
Aquatic interstitial subterranean habitats,
240
Aquatic stygobites distribution, 853—857,
854f, 856f
Antrolana lira, 854—855, 854f
Crangonyx antennatus, 855, 856f
fishes, 854
Stygobromus, 853—854
Aquifers. See also Calcrete aquifers; Edwards
Aquifer; Karst aquifer, hydrology of;
Karst aquifers, modeling of; Karst
aquifers, water tracing in
NAPLs storage in aquifers, 171
fractures, 171
matrix and vugs, 171
pools in conduits, 171
sediment, 171
NAPLs transport into aquifers, 168—171
climate and, 169, 170
epikarst and, 169—170
fractures and, 170—171
nonwetting fluid and, 169
open drains and, 170—171
sinkhole formation and, 170
sinking streams and, 170—171
soil and regolith pores and, 169
volume and rate of release and,
168-169
Arabika Massif
deep cave evolution, 449—450
deep caves in, 443—447
geology, 447—448, 447f
hydrogeology, 448—449
location and physiography, 443, 444f
Ortobalagan Valley in, 443, 445f
Arachnida, 256—257
mites of, 256
Opiliones of, 256
Pseudoscorpions of, 256—257
Schizomida of, 257
scorpions of, 257
spiders of, 256
tropical diversity patterns, 245t
Aragonite, 500, 5011
calcite problem, 783—784, 783f
massive column of, 783, 784f
stalagmites, 809
Araneae, 256, 789— 791
in Australia, 215
in Dinaric karst, 232
Araneomorphae, 789
Arctodus pristinus, 593
Arctodus simus, 593, 594f
Arctotherinm, 593—594
Argasidae, 794
Arsenates, 506
Artesian hypogene speleogenesis, 759
Artesian springs, 799—800
Ascending vertical exploration techniques,
307, 316f, 317-318
Asellota, 178t, 188-190,189t
Asellus aquaticus, 653, 654t
Asellus kosswigi and, 34—35, 34f
broader geographic distribution, 33—34,
34f
continental scale, 35-36, 35f, 35t
future prospects, 36
habitats, 30—31, 31f
natural selection and, 545
overview, 30—31
phenotypic dines and, 545—546, 546f
PPCS and, 32-33
respiratory area of, 31, 31f
subspecies, 31
troglomorphies development, 31—32, 32f
Asellus kosswigi, 34—35, 34f
Assamiidae, 796
Astacidea, 178t, 192,192t
Astigmata, 795
Astyanax fasciatus
aggressive behavior of, 65—66, 66f, 67f
egg size, 6-7, 7f
energy demand, 4, 4t
excitement and aggression and, 5
feeding behavior of, 63
neutral mutations
albinism and, 550
dark melanin and, 550
deleterious risk hypothesis and,
552-553
energy economy hypothesis and, 553
eyes and, 550, 552f, 554
guanine and, 550
neuromasts and, 553
overview, 549—550
pleiotropy hypothesis and, 553
surface and cave forms regarding,
550
reproductive behavior of, 64—65
Astyanax mexicanus
developmental stages, 39f
melanophore differentiation and, 41—42,
41 f
as model organism, 37
natural history of, 37—38
populations and distribution, 37—38, 38f
troglomorphic traits, 38—40
constructive, 38t, 39, 40f
developmental basis of, 40—42
evolution of, 42—43
genetic basis of, 42
regressive, 38-39, 38t
Atemnidae, 791
Atlantida Cave, 829t
Aturidae, 794—795
Audio-magnetotelluric soundings (AMT),
349-350
Austin blind salamander, 669
Australia. See also Diversity patterns, in
Australia; Nullarbor Caves, Australia
root communities in, 661t
Austrochilidae, 790
Autochthonous sediments, 135
Autotroph, 490—491
Avens, 423
Aves, 261
B
Back to network, 6
Bacteria
conservation and, 624
cyanobacteria, 453
iron and manganese, 133
Proteobacteria, 495—496
saltpetre mining and, 676
symbiosis, 560
Bad air zone, 662—663
Balitoridae, 107
Ball, Edward, 874-875
Bamboo, 402, 403f
Banta, A. M., 404-405
Barberry Cave, 80, 80f
Barenschacht, 699—700
Barite, 501t, 503
Barometric airflow. Jewel Cave, 411
INDEX
919
Barr, T. C., 276
Barton Springs salamander, 669—670
Barychelidae, 789
Basic Cave Diving: A Blueprint for Survival
(Exley), 310
Basins, 427—428
Basswood Cave, 81—82
Bathynellacea, 178t, 184, 184f, 257
Batrachochytrium dendrobatidis, 624
Bats
in Australia, 218
bell holes and, 423-424, 424f
cave climate influenced by, 357
conservation, 625
decline of, 625
mining and, 625
overview, 625
Shelta Cave, 625
water projects and, 625
WNS and, 625-626
in Dinaric karst, 232—233
distribution of, 46
echolocation and, 45
ecosystem services provided by, 51—54
flight and, 45
food sources and, 331t, 333
guano, 358
carnivorous, 358-359, 359f
community structure and succession
in, 360-361
conservation and, 242—243, 249, 626
frugivorous, 358-359, 358f
hematophagous, 358-359, 359f
habitat destruction and alteration and, 52
hibernation of, 48-49, 48f
human disturbance and, 52
Jewel Cave, 415—416
life-history traits, 45
living in caves
benefits, 49-50
costs, 49-50
management, 50—51
overview about, 45—46
paleontology and, 592
roosting habits, 45, 46—48
courtship and mating, 47
rearing young, 47—48
roost functions, 47
Rousette fruit, 46
Townsend's big-eared, 46f
tropics and, 242—243, 249
in United States, 261—263, 262t
vampire, 53—54
wind turbines and, 53
WNS and, 52-53, 625-626
atypical winter behavior, 905—906
cave closures and decontamination,
908
characteristics, 905—907
conservation and, 625—626
current and future research,
907-908
depleted fat reserves, 905
diagnostics, 904-905
European connection, 907
fungal infections and tissue damage,
905, 905f
fungal transmission, 907
immune response changes, 906
mortality and, 904
overview, 903—904
pathogen, 903f, 904
species affected by, 905
wing damage, 906—907, 906f
Battered Bar Cave, 80—81
Battery operated light sources, 305-306
Beaded helictites, 379, 380f
Bears, 593-594
Ailuropoda, 594
Arctodus pristinus, 593
Arctodus simus, 593, 594f
Arctotherium, 593—594
tremarctine, 593
Ursus americanus, 593
Ursus speleus, 593
Bear's Shaft, 699—700
Bedrock-collapse sinkhole, 720f, 722
Beetles
anophthalm and aphenopsian, 56f, 56—57
biogeography and, 851-852, 852f
cricket egg prey of, 744-746
development cycles of, 56, 56f
ecology and, 59-61
fecal food sources and, 331t, 333
importance and protection of, 61
Leptodirini group, 57, 58f
Leptodirus hochemvarthii, 55, 55f, 231f,
519f
phylogenetic Mediterranean tree, 59,
60f
other types of, 57, 58f
overview about, 54—55
Pholeuon
flooding periods influencing, 60, 61f
proportion of cave, 55, 55f
systematics of, 56-58, 56f, 58f
troglomorphic features of, 56—57, 57f
White Cave food web and, 60—61, 61f
Behavioral adaptations
aggressive, 65-68, 66f, 67f
Astyanax fasciatus
aggression, 65-66, 66f, 67f
feeding, 63
reproduction, 64—65
discussion, 68
food and feeding, 63, 64f
Munidopsis polymorpha
aggression, 66-67, 67f
reproduction, 63, 64f
overview, 62—63
Poecilia mexicana
aggression, 67-68
reproduction, 64, 65f, 67—68
Proteus anguinus
aggression, 67-68, 67f
feeding, 63, 64f
reproduction, 64—65
reproductive, 63—65, 64f, 65f
Bell holes, 423-424, 424f
Benarat Caverns System, 536
Berchilskaja Cave, 443, 444f
Berry cave salamander, 673f, 674
Best management practices (BMPs), 627—628
Better Forgotten Cave, 77—78
Biantidae, 796
Bible, 322
Big river caves, 308
Biogeography
anchihaline habitats, 22—23
aquatic stygobites distribution and,
853-857, 854f, 856f
Antrolana lira, 854—855, 854f
Crangonyx antennatus, 855, 856f
fishes, 854
Stygobromus, 853—854
of cave animals, 850-851
cavefish of China, 121—124
cladistic, 850, 850f
epikarst communities and, 293—294
overview of, 849-850
panbiogeography, 850
Sinocyclocheilus, 121—124
summary about, 857
terrestrial troglobites distribution and,
851-853, 852f
beetles, 851-852, 852f
of thalassoid limnostygobionts, 483, 484f,
485f
types of, 849-850
vicariant phenomena, 855
calcrete aquifers and, 857
spring failure, 856-857
stranding, 855-856
stream capture, 856
Biospeologica, 219—220
Birds, as cave visitors
in Australia, 218
cave ecology impact of, 847
cave-dwelling species, 845—846
facultative visitors, 845
regional cavernicoles, 845
regular cavernicoles, 845—846
den sites and, 846—847
foraging habitats and, 847
guano and, 358
knowledge gaps, 846
man and, 848—849
rare species, 848
resting and mating and, 846
wintering sites and, 847
Bird's nest soup, 532
Black Hills caves, 417. See also Jewel Cave,
South Dakota
Blanco blind salamander, 669
Blarney Stone Cave, 79
Blattodea, 216-217
Blepsimolge, 668f, 669-670
Blind Faith Cave, 80, 280
Blind salamanders, 668—669, 668f
Block breakdown, 69
Block-fault structure, 445—447
Blowholes, 570
Blue holes, 159-160
defined, 160
exploration, 160
920
INDEX
Blue holes ( Continued )
formation hypotheses, 160
locations, 160
overview about, 159—160, 160f
settings, 160,160f
speleothems and, 160
BMPs. See Best management practices
Bobcat Cave, 78—79, 79f
camps, 87, 87f, 88f
description, 78—79, 79f
entranceless caves and, 280
hypothermia and, 87—88
nylon used in, 88
sleeping and, 88
team sizes, 88
Bochicidae, 791—792
Bochusacea, 178t, 186
Bogaz, 428—429
Bogli, Alfred, 726
Boneyard maze, 461—462, 462f
Boon's Blunder, Castleguard Cave, 91—93,
91f
Bougainvilliidae, 234
Bouguer gravity anomalies, 348
Boundless Cave, 78
Brachyura, 178t, 192, 193t
Branchiopods, 177—179, 178t
Breakdown
block, 69
breakout domes, 69—70, 70f
chip, 69, 72-73
geologic influences on, 72—73
crystal wedging, 72—73
initiation of, 72
limestone replacement, 72— 73, 73f
Jewel Cave and, 412—413
mechanics, 70—72, 71f
fixed-beam model, 70, 71f
microcracks propagation model, 71
mines and, 70
morphology, 69-70
classification, 69
large-scale features, 69-70, 70f
small-scale features, 69
overview, 68
recent, 71
slab, 69
speleological processes and, 73
enlargement phase, 73
stagnation and decay phase, 73
terminal, 69
triggering processes, 72
base-level back flooding, 72
buoyant support removal, 72
crystal wedging, 72—73
ice wedging, 72
phreatic passage enlargement, 72
vadose water action, 72, 72f
Breakout domes, 69—70, 70f
Breakthrough curves, 894, 894f, 895f
Breathing Cave, 74, 77
Brushite, 501t, 504—505
Buckets of Smoke Cave, 82
Buckwheat Cave, 80, 280
Bugs, 362
Bukovinka Cave, 829t
Burial dating, 173-176, 174f
caves and human evolution and, 176
conclusions, 176
Mammoth Cave and, 175—176
requirements for, 174—175
Buried sinkhole, 719f, 722
Burns Chestnut Ridge Cave, 80
Burnsville Cove, Virginia
background about, 74
cave descriptions, 77—82
Aqua Cave, 77
Barberry Cave, 80, 80f
Basswood Cave, 81—82
Battered Bar Cave, 80—81
Better Forgotten Cave, 77—78
Blarney Stone Cave, 79
Blind Faith Cave, 80
Bobcat Cave, 78— 79, 79f
Boundless Cave, 78
Breathing Cave, 77
Buckets of Smoke Cave, 82
Buckwheat Cave, 80
Burns Chestnut Ridge Cave, 80
Butler Cave, 74, 77, 78f
By-the-Road Cave, 81
Chestnut Ridge Cave System, 78— 79,
80
Fuhl's Paradise Pit, 82
Helictite Cave, 81, 81f
Robins Rift, 78
Water Sinks Cave, 82
Wishing Well Cave, 82
exploration history, 75—77
map, 74f
physical and geologic setting, 75
special attributes, 82—83
Bush-like caves, 395, 395f
Buthidae, 787
Butler Cave, 74, 77, 78f
Butler Cave Camp, 86
Bypass tubes, 423
By-the-Road Cave, 81
c
C-3 Expedition. See Crystal Cave Expedition
Calabozoidea, 178t, 189t, 190
Calanoida, 178t, 181,181t
Calcite, 500, 5011, 503f, 777
aragonite problem, 783—784, 783f
chemistry and, 769— 770, 770f
Jewel Cave, 411, 412f
Nullarbor Caves and, 574
rafts, 783
Calcium
epikarst and, 289
equilibrium, 508-509, 509f
flux rate, 509-510, 509f
Calcrete aquifers, 857
Calcretes, groundwater, 207—208, 213f
Cambalopsidae, 249, 249f
Millipedes, 242, 244f
Campeche State, 833-834
Camps
Bobcat Cave and, 87, 87f, 88f
Butler Cave Camp, 86
community gear used in, 88
Crystal Cave Expedition and, 85—86
eating and, 88
Ellison's Cave and, 86
European model for, 85
first, 85
hypothermia and, 87—88
LED lamps and, 88
Mexico and, 86—87, 87f
Midnight Cave and, 86
nylon and, 88
Project SIMMER and, 86
sleeping and, 88
Sullivan Cave, 86
tarp, 89
Wind Cave, and, 86
Cango caves, 695f
Canyon-like passages, 734, 735f
Capitan Formation, 457-458, 457f
Carabidae, 221, 222f, 230-232
Carbide lamps, 305
Carbon fixation pathways, 126t, 127
Carbonates, 500, 5011
aragonite, 500, 50It
calcite, 500, 501t, 503f
coralloid multiaggregates and, 500, 503f
deposition of, 500
hydromagnesite, 500
malachite, 500, 5011
sulfate strata dissolution, 759
Carbonic acid dissolution, 756— 758
Caridea, 178t, 191,192t
Carlsbad Cavern, 816—817, 816f
Carrot-shaped stalactite, 779—781, 781f
Cascade Caverns salamander, 670
Cassowary System, 731—732, 732f
Castleguard Cave, Canada
Boon's Blunder, 91—93, 91f
Central Cave, 89-90, 92f
climate, 93—94
conclusions about, 94
dating, 93
Downstream Complex, 89—90
fauna, 94
genesis, 89, 90
Headwater Complex, 89—90
Ice Crawls, 92f
modern hydrology, 90—93, 90f, 93f
morphology, 89—90, 90f
overview, 89, 90f
phreatic looping, 90, 92f
sediments, 93
speleothems, 93
Subway, 92f
Cats, 595, 595f
Caves. See also specific cave
aeolian, 105f, 106
beads and, 779, 780f
burial and, 487-488
cotton, 375
darkness and, 104
definitions and points of view, 103
exploration and, 103
eyeless and depigmented species and, 104
INDEX
921
as geological repositories, 103-104
glacier, 105f, 106
as groundwater system parts, 103—104
as habitat, 104—105
hygropetric, 236
interstitial aspect of, 105
pearls and, 782— 783, 783f
radiolocation, 351—352, 351f
rock shelters and, 105f, 106
sea, 105f, 106
shields and, 381
solution, 105f, 106—107
suffosional, 105f, 106
talus, 105f, 106
tectonic, 105-106, 105f
temperature buffering and, 104
types of, 105-107, 105f
volcanic, 105f, 106
water sources and, 105f, 107
Cave Minerals of the World (Hill & Forti), 499
Cavefish of China
biogeography, 121—124
biology of, 108—112
characteristics of, 112—113
distribution of, 107-108, 109t, 112f
diversity of, 107-108, 108f, 109t
morphology and adaptation, 113—121
adipose storage, 117,118f, 119f
body shape changes, 108f, 113—114,
113f
eye degeneration, 114f, 117—120,
119f, 121f
humpback and horn, 114—116, 114f,
115f
pigmentation loss, 108f, 120
scale disappearance, 120—121
sensory apparatus, 113f, 116—117,
116f
specialized appendages, 108f, 117
Onychostoma macrolepis, 107, 109t
phylogenesis, 121—124, 123f
reproduction, 108—112
research and conservation of, 124—125
Sinocyclocheilus hyalinus, 107, 108f, 109t
speciation mechanisms, 121—124
Cavernous edging, 761, 762
CCRs. See Closed circuit rebreathers
Celestine, 374, 501t, 503
Cenotes, 151, 836-837
Centipedes, 539-540
Craterstigmatomorpha, 539
Geophilomorpha, 539
Lithobiomorpha, 539—540, 539f
overview, 539
Scolopendromorpha, 539
Scutigeromorpha, 539
Central Cave, Castleguard Cave, 89-90, 92f
Cerknica, polje of, 433, 435f
Cerknisko jezero/polje, 898-899, 898f, 899f
Chactidae, 788
Chaerilidae, 788
Chandelier Maze, 464
Chandeliers, 374—375
Channels, 428
Chappuis, Alfred, 219—220
Character reduction, 6
Charontidae, 789
Cheiridiidae, 792
Chelicerata
in Australia, 214—215
Acarina, 214
Amblypygi, 214
Araneae, 215
Opiliones, 215
Palpigradi, 214
Pseudoscorpiones, 215
Schizomida, 215
Scorpiones, 215
Cheliferidae, 792
Chemical remanent magnetism (CRM), 588
Chemoautotrophy, 100—101
bacteria symbiosis and, 560
carbon fixation pathways, 126t, 127
in cave and karst settings, 130—133
chronological summary, 131t
iron and manganese bacteria, 133
methane cycle, 133
nitrogen cycle, 132-133
sulfur cycle, 130—132, 132f
cave ecosystems and, 133—134
energy reactions, 126t
as food resource, 3
future and, 134
microbial groups associated with, 126t,
127-130
ammonia and nitrite, 129
hydrogen, 127—128
iron and manganese, 129
one-carbon compounds, 129—130
overview of, 127
sulfur compounds, reduced
inorganic, 128—129
overview about, 125—126
terminology regarding, 126—127
Chemolithoautotrophs, 126—127, 490—491.
See also Chemoautotrophy
Chemoorganotrophs, 126—127, 490—491
Chemosynthesis, 125. See also
Chemoautotrophy
Chernetidae, 792
Chester Upland, 736
Chestnut Ridge Cave System
Blarney Stone Cave, 79
Bobcat Cave, 78—79, 79f
Burns Chestnut Ridge Cave, 80
entranceless caves and, 280
Chilopoda, 258
in Australia, 216
Centipedes, 539-540
Craterstigmatomorpha, 539
Geophilomorpha, 539
Lithobiomorpha, 539—540, 539f
overview, 539
Scolopendromorpha, 539
Scutigeromorpha, 539
Chimney effect, 401
Chimneys, 423
Chinese cavefish
biogeography, 121—124
biology of, 108—112
characteristics of, 112—113
distribution of, 107-108, 109t, 112f
diversity of, 107-108, 108f, 109t
morphology and adaptation, 113—121
adipose storage, 117,118f, 119f
body shape changes, 108f, 113—114,
113f
eye degeneration, 114f, 117—120,
119f, 121f
humpback and horn, 114—116, 114f,
115f
pigmentation loss, 108f, 120
scale disappearance, 120—121
sensory apparatus, 113f, 116—117,
116f
specialized appendages, 108f, 117
Onychostoma macrolepis, 107, 109t
phylogenesis, 121—124, 123f
reproduction, 108—112
research and conservation of, 124—125
Sinocyclocheilus hyalinus, 107, 108f, 109t
speciation mechanisms, 121—124
Chinese folklore, 322
Chip breakdown, 69, 72—73
Chirocephalus croaticus, 899f
Chisholm Trail salamander, 670
Cholevidae, 230
Chologaster comuta, 7—8, 8t
Chologaster comuta, 465, 466t
Chordeumatida, 540t, 541
Chthoniidae, 792
Cirolanidae, 212—213
Cixiid planthoppers, 13—14, 13f
Cladistic biogeography, 850, 850f
Cladonychiidae, 796
Clastic sediments
abandonment suite and, 137
aggradation and, 136f, 137
geomorphic perspective, 139
grain size and, 135,136f
Karst aquifer and, 390
paleomagnetic records and, 586-588, 587f
physical properties, 134-136, 135f, 135t
production, 137—138
sedimentary structures and, 135—136, 136f
stratigraphic facies and, 135t, 136f, 137
stratigraphy and, 135t, 136—137, 136f
transport and deposition, 138—139
gravity sediments, 139
Hjulstrom's Diagram and, 138, 138f
Shield's Diagram and, 138f, 139
stream sediments, 138—139,138f
Clearwater River Cave, 532—534
Clearwater River Passage, 534, 535f
Cleaveland Avenue, 472f
Clemmer, Gregg, 76—77
Climate. See also Paleoclimate records, from
speleothems
aquifers and, 169,170
bats influencing cave, 357
Castleguard Cave, 93-94
Dinaric karst and, 195—197
Nullarbor Caves, 570—571
solution caves and, 724
surface Karren and, 427
922
INDEX
Climate ( Continued )
Vjetrenica and, 860—861
Climatic-relict model, 405
Clint, 428-429, 429f
Closed circuit rebreathers (CCRs), 313
Closed depressions, 434—435
compound hollow, 152—153
dolines
collapse, 150-151,150f, 151f
cover, 150f, 152
intersection, 150f, 152
other types of, 150-152, 150f
subsidence, 150f, 151—152
dolines, solution
conclusions about, 150
defined, 140
drainage and, 140
drawdown, 143f, 144
evolution examples, 144f, 145—147
Eyre Peninsula, 147
filling material evacuation and, 145
frost shattering and, 145
inception, 143f, 144
as karst surface diagnostic forms,
140-150
locations, 141
midlatitude, 149
Monte Baldo, 147
Montello, 145-147, 146f, 148f
Monti Lessini, 144f, 145
morphology and size, 140—141
other evolution processes, 145—147
overview about, 140
point-recharge, 142—144, 143f, 145
Santa Ninfa Plateau, 145
slope processes and, 145
soil and, 145
specific environmental conditions
and, 147-149
structure, 141-142, 142f
subsoil, 141
tropical landscape and, 147—149,
149f
Velebit Mountains, 147
Waitomo district, 145
poljes, 140,153-155
classification, 154
defined, 153
Dinaric karst and, 153,199—200
lakes and, 153,153f
lithological contact, 154, 154f
polygenetic sink, 152—153
types of, 140
uvala, 152—153
Coastal caves
blue holes, 159-160,160f
categories of, 155
conclusions about, 161
flank margin caves, 156—159,157f, 158f,
159f, 160f
overview, 155
sea caves, 155-156,156f
sea level and, 155
Coastal karren, 425f, 430
Coastal karst, 437
Cobitidae, 107
Cobra-Bridge-Cloud Cave, 534
Cockroaches, 241, 242f, 362, 663
Cold air trapping, 400—401
Coleoptera, 230, 259—260. See also Beetles
Carabidae, 230—232
Cholevidae, 230
Dystiscidae, 217—218
Pselaphidae, 232
Collapse dolines, 150—151, 150f, 151f
Collapse shafts, 761, 762
Collapsed tiankeng, 821, 821f, 822f
Collembola
Australian diversity patterns, 216
European diversity patterns, 221 —222,
224f
morphological adaptations in, 523, 523f
tropical diversity patterns, 245t
Collins, Floyd, 85—86, 648
Colonization
active model of, 275, 405
deep sea, 405
Dinaric karst human, 229
ecological classification and, 275
natural selection and, 547—548
Comal blind salamander, 670
Combined artesian/deep-rooted
speleogenesis, 759
Commensalism, 743, 744f
Competition, 743, 744f
Appalachian stream, 746— 747, 746f, 747f
eutrophication and, 747—748
initial flow routes, 599—600, 600f
Niphargus and, 559—560, 560t
Competitive pressure, 405
Components, 269—270
Composite 3D systems, 762—763
Compound depressions, 140, 152—153
Compound hollow, 152—153
Condensation
corrosion hollows, 425, 425f
rills/flutes, 425
Conduit, 104
diffuse flow springs and, 798
hypogene, 367f
NAPLs and, 171
networks evolution, 767— 769, 768f
permeability, 384, 385-387, 385f, 385t
single-conduit caves, 394
systems carrying capacity, 389
tubular, 734-735, 735f
Cones, 435-436, 436f
Congeria kusceri, 512—513, 517
Conical stalactite, 807, 807f
Conn, Herb, 411
Conn, Jan, 411
Conservation, 309
bats, 625
decline of, 625
guano and, 242-243, 249, 626
mining and, 625
overview, 625
Shelta Cave, 625
water projects and, 625
WNS and, 625-626
of cave life, 624—626
bacteria and, 624
chemical spills and, 625
extinct and endangered species and,
625
fungus and, 624
human impact and, 624
threats, 624—625
water pumping and, 625
cave preserve design, 626-631
gate styles, 628-630, 629f, 630f
gating criteria, 628
karst BMPs and, 627—628
large cave buffer zones, 627
security systems, 631
signs and, 626, 627f
small cave buffer zones, 627
gating decision guide, 630-631
final results, 631
good reasons not to gate, 630
good reasons to gate, 630-631
poor reasons to gate, 630
poor reasons to not gate, 630
management, 631-633
education and, 631—633
prioritizing caves, 632—633
Vjetrenica, 864
Contamination. See Heavy metals;
Nonaqueous phase liquids
Copepoda, 178t, 181-182, 291-293
anchihaline habitats and, 21
in Australia, 211
Calanoida, 178t, 181,181t
Cyclopoida, 178t, 181-182
in Dinaric karst, 233
Gelyelloida, 178t, 182
Harpacticoida, 178t, 182
Misophrioida, 178t, 181, 182t
Platycopioida, 178t, 181
Coral Sea, 464
Coralloid multiaggregates, 500, 503f
Corkscrews, 423
Corridors, 428—429
Corrosion hollows, 425, 425f
Corrosion notch, 425, 429—430
Cosmogenic isotope dating
burial dating and, 173—176, 174f
caves and human evolution and, 176
conclusions, 176
Mammoth Cave and, 175—176
requirements for, 174—175
dating techniques
relative vs. absolute, 173
overview, 172—173
Cotton passive detector, 891
Cover dolines, 150f, 152
Cover-collapse sinkhole, 718—719, 719f, 720f
Cover-subsidence sinkhole, 719f, 721—722,
721 f
Crangonyx antennatus, 855, 856f
Crash cycles, 10-11
Craterstigmatomorpha, 539
Crawford Upland, 736
Crayfish, 521, 522f
Creep closure, 354-355, 355f
INDEX
923
Crete, 321
Crevice Cave, 738, 738f
Crickets, 15, 333, 331t
eggs, 330, 744- 746
guano and, 358
heterozygosity in, 615—617, 616t
root communities and, 663
CRM. See Chemical remanent magnetism
Cross-bedding, 135—136, 136f
Crustacea. See also Gammarus minus
anatomical variety, 177
in anchihaline habitats, 20—22, 209, 210,
214f
Amphipoda, 22
Copepoda, 21
Decapoda, 21
Isopoda, 22
Leptostracans, 21
Mictaceans, 22
Mysids, 21-22
Niphargus, 22, 22f
Ostracods, 21
Remipedia, 21
Thermosbaenaceans, 21
in Australia, 209—214
anchihaline habitats, 209, 210, 214f
Copepoda, 211
Decapoda, 213—214
Isopoda, 212—213
Ostracoda, 211—212
overview, 209
Spelaeogriphacea, 213
Syncarida, 209—211
low oxygen and, 653
overview about, 177
stygobitic taxa synopsis, 177—193, 178t
Branchiopods, 177—179, 178t
Hexapoda, 258-259
Insecta, 259-260
Malacostraca, 178t, 183—193,
257-258
Maxillopoda, 178t, 180-182, 257
Ostracoda, 178t, 182-183, 182f
Remipedia, 178t, 179-180,180f, 180t
taxonomy, 177, 178t, 290, 290t
troglomorphy, 177
tropical diversity patterns, 245t
Crusts, gypsum, 374, 375
Cryogenic cave micropearls, 401—402
Cryptostigmata, 795
Crystal Cave Expedition (C-3 Expedition),
85-86
Crystal wedging, 72-73
Cueva de Villa Luz, Mexico, 813f, 813—815,
814t, 815f
Culturing, 492
Cumacea, 190,191t
Cumberland River incision pulses, 530, 530f
Cunaxidae, 794
Cupola, 395, 424, 425f
Curl, Rane, 278
Curvilinear branchwork, 602f
Curvilinear passages, 604—605, 604f
Cusps, 425, 425f
Cut-and-closure caves, 354-355, 355f
Cutter, 429
Cvijif, Jovan, 433
Cyanobacteria, 453
Cybaeidae, 790
Cycloctenidae, 790
Cyclopoida, 178t, 181-182
Cyphophthalmi, 795—796
Cyprinidae, 234
Cyrtaucheniidae, 789
D
Damaeidae, 795
Danielopolina kornickeri, 214f
Dan's Cave, 311f
Danube River, 707
Darwin, Charles, 543—544
Das Karstphanomen, 433
Dating. See also Uranium series dating, of
speleothems
burial, 173-176, 174f
caves and human evolution and, 176
conclusions, 176
Mammoth Cave and, 175—176, 737
requirements for, 174—175
Castleguard Cave, 93
ice in caves, 402
paleoclimate records from speleothems,
577-578
relative vs. absolute, 173
from stygofauna, 484—486
Davis, Nevin W., 76—77
Deal, Dwight, 411
Decantation runnel, 428
Decapoda, 178t, 191,191f, 258
anchihaline habitats and, 21
Anomura, 178t, 192
Astacidea, 178t, 192,192t
in Australia, 213—214
Brachyura, 178t, 192, 193t
Caridea, 178t, 191,192t
Deep sea colonization, 405
Deep Secrets, 463
Deepest caves, 443
Deep-rooted hypogene speleogenesis, 759
Deep-seated karst, 365, 367
Deer Cave, 532, 534f
Degeneration, cave, 603
Delay a bureschi, 231 f
Deleterious risk hypothesis, 552—553
Density-driven flow, 751—752
Denuded karst, 365
Dermaptera, 259
Descending, 317
Deterministic biogeography, 849—850
Detrital remanent magnetism (DMR), 586
Devils Hole, 579-580, 580f
Diagenetic environments, 750— 756
eogenesis and, 751
mesogenesis and, 751
telogenesis and, 751
Dictynidae, 790
Diffuse flow springs, 798
Dinaric karst. See also Diversity patterns, in
Dinaric karst
boundaries of, 195, 196f
caves and, 201
as classical karst, 195
climate regarding, 195—197, 229
contact karst and, 201
depressions, 199—200
described, 228—230
dolines, 199—200
flora and fauna and, 229—230
formation of, 229
human colonization of, 229
hydrological characteristics of, 197,197f
karst features of, 199—202
land use, 202
leveled surfaces, 200—201, 201f
lithology, 198-199,199f
location, 195, 228-229, 229f
poljes and, 153,199-200
relief forms, 199
roofless caves, 201—202, 202f
structural geology, 198
surface of, 201-202, 202f
tectonic map of, 196f
today, 229
traits, 195-197
uvalas, 199—200
vegetation and, 197,197f
Diplocentridae, 788
Diplopoda, 258
in Australia, 216
in Dinaric karst, 232
Millipedes, 540—541
Chordeumatida, 540t, 541
Julida, 540, 540t
map of, 540f
overview, 540
Polydesmida, 540, 540t
Pseudotremia, 541, 541f
Scoterpes, 541
Diplura, 216, 245t
Dipluridae, 789
Diptera, 260
Directional selection, 545, 545f
Directly adaptive hypothesis, 342
Discharge passages, 604
Dispersed infiltration, 387
Dissolution basins, 420, 421f
Dissolutional pavements, 427, 427f, 429f, 430
Dissolutional potholes, 420, 422f
Dissolved organic carbon (DOC), 289
Distributary springs, 800
Divergent selection, 545, 545f
Diversity patterns, in Australia
amphipods, 207
Chelicerata, 214—215
Acarina, 214
Amblypygi, 214
Araneae, 215
Opiliones, 215
Palpigradi, 214
Pseudoscorpiones, 215
Schizomida, 215
Scorpiones, 215
conclusions about, 218—219
deficiency of, 203
geographic factors
924
INDEX
Diversity patterns, in Australia ( Continued )
cave atmosphere, 205—208
energy supply, 208-209
groundwater calcretes, 207—208, 213f
overview, 204
shield regions and Cretaceous
marine transgressions, 204—205,
205f
stagnant-air zone, 205
tropical winter effect, 205
insects, 216—218
Blattodea, 216-217
Coleoptera Dystiscidae, 217—218
Collembola, 216
Diplura, 216
Orthoptera, 217
Planthoppers, 216
Zygentoma, 216
lack of study of, 204
Myriapoda, 216
Chilopoda, 216
Diplopoda, 216
overview, 203—204, 203f
stygofauna and Crustacea, 209—214
anchihaline habitats, 209, 210, 214f
Copepoda, 211
Decapoda, 213—214
Isopoda, 212—213
Ostracoda, 211—212
overview, 209
Spelaeogriphacea, 213
Syncarida, 209—211
subterranean animal examples, 206f
subterranean invertebrates distribution,
212t
vertebrates, 218
bats, 218
birds, 218
fishes, 218
snakes, 218
Diversity patterns, in Dinaric karst
Amphipoda, 233
Araneae, 232
bats, 232-233
biogeographical, 234-236
endemism, 235—236
holodinaric distribution, 235
merodinaric distribution, 235
smaller distribution areas, 235
transdinaric distribution, 235
widely spread taxa, 234
Bougainvilliidae, 234
Coleoptera, 230
Carabidae, 230—232
Cholevidae, 230
Pselaphidae, 232
Copepoda, 233
Cyprinidae, 234
Diplopoda, 232
Dreissenidae, 234
fauna composition and diversity, 230—234
aquatic, 233—234
examples, 231f
history, 230
terrestrial, 230—233
Gastropoda, 232, 233
hotspots, 236
Hercegovina, 236
Postojna-Planina Cave system, 236
Isopoda, 233-234
overview, 228—230, 229f
pollution and protection and, 237—238
Proteus anguinus, 234
Pseudoscorpiones, 232
Scutariellidae, 234
Serpulidae, 234
special assemblages, 236-237
cave hygropetric, 236
crevice systems, 237
ice caves, 236
sinking streams, 237
thermal waters, 237
Spongillidae, 234
Thermosbaenacea, 234
Diversity patterns, in Europe
Carabidae, 221, 222f
Collembola, 221-222, 224f
conservation and, 226—228
different habitats and, 226, 227t
factors influencing, 221
geographic patterns and, 224—226, 225f,
226f, 227f, 227t
hotspots, 225, 226f
Leiodidae, 221, 223f
midlatitude biodiversity ridge and,
224-225, 225f
molluscs, 221, 224f
Niphargidae, 221
other continents compared with, 220—221,
220f
overview, 219—220
troglobionts to stygobionts ratio and, 225,
227f
troglomorphy and relictness and,
221-223, 222f, 223f, 224f
Diversity patterns, in tropics
background regarding, 238, 239f
conservation issues, 249—250
bats and guano, 242—243, 249
endemism and vulnerability,
249-250
current knowledge state about, 238—240,
239f
historical context, 238
relictual vs. nonrelictual taxa, 247—249
subterranean habitats, 240—243
anchihaline and marine caves, 240
aquatic interstitial, 240
cave freshwater, 240—241
guano and guanobionts, 242
oligotrophic terrestrial, 241—242
troglophilic species, 242—243
temperate subterranean biodiversity vs.,
239f, 247
troglobiont broad-scale, 243—247
along environmental gradients, 247
geographical patterns, 243—247,
245t
Diversity patterns, in United States
cave ecology and, 252
cave/karst regions and, 252—254, 253f,
253t, 254f
discussion about, 263
frequency histograms by karst region,
256f
frequency histograms by state, 255f
habitat types and, 252
invertebrates, 254—260
Arachnida, 256—257
Crustacea, 257-258
Hexapoda, 258-259
Insecta, 259—260
Malacostraca, 257—258
Maxillopoda, 257
Mollusca, 255
Oligochaeta, 254-255
Turbellaria, 254
maps regarding, 253f, 254f
overview, 251-252, 251t
species richness comparison and, 253t
vertebrates, 260—263
Amphibia, 260, 2611
Aves, 261
Mammalia, 261—263, 262t
Osteichthyes, 260
overview, 260
Reptilia, 260—261
DMR. See Detrital remanent magnetism
DNAPLs. See Heavy nonaqueous phase
liquids
DOC. See Dissolved organic carbon
Documentation
databases, 269—271
advantages, 271
components, 269—270
data transfer and, 271
disadvantages, 271
entities and, 269
fields and, 269
foreign key and, 270
forms and, 270
future needs, 271 —272
issues, 270—271
linkages and, 270
overview, 269
queries and reports and, 270
records and, 270
tables and, 270
defined, 264
GIS and, 264-265
history, 264—269
information safety, 268—269
information storage, 266—267
in digital form, 266—267
issues, 265
access policy, 265
cave identification, 265
clerical, 265
credibility, 265
information retrieval, 265
information transfer, 265
record safety, 265
secrecy, 265
standards, 265
Krubera Cave, 443
INDEX
925
overview, 264
publication, 269
recording and, 265—266
coordination, 266
external issues, 266
helpful references, 266
information consolidation, 266
results publication, 266
storage and distribution, 266
updating, 266
retrieval, 267
standards, 267
existing, 267
planned, 267
surveying and mapping, 267—268
Voronja Cave, 443
Dolichopoda, 614
Dolichopoda laetitiae, 610—611, 610f
Dolines, 140. See also Tiankeng
collapse, 150-151,150f, 151f
cover, 150f, 152
defined, 434-435, 723
Dinaric karst, 199—200
intersection, 150f, 152
Nullarbor Caves, 570, 570f
other types of, 150—152, 150f
in Slovenia, 435f
solution
conclusions about, 150
defined, 140
drainage and, 140
drawdown, 143f, 144
evolution examples, 144f, 145—147
Eyre Peninsula, 147
filling material evacuation and, 145
frost shattering and, 145
inception, 143f, 144
as karst surface diagnostic forms,
140-150
locations, 141
midlatitude, 149
Monte Baldo, 147
Montello, 145-147, 146f, 148f
Monti Lessini, 144f, 145
morphology and size, 140—141
other evolution processes, 145—147
overview about, 140
point-recharge, 142—144, 143f, 145
Santa Ninfa Plateau, 145
slope processes and, 145
soil and, 145
specific environmental conditions
and, 147-149
structure, 141—142, 142f
subsoil, 141
tropical landscape and, 147—149,
149f
Velebit Mountains, 147
Waitomo district, 145
subsidence, 150f, 151—152
suffosion, 722
Dolonization, deep sea, 405
Double-trunked lava tunnels, 869—870
Downstream Complex, Castleguard Cave,
89-90
Dowsing, 352
Draculoides vinei, 206f
Drainage grooves, 422
Drained lobes, 871—872
Drapery, 781
Drawdown doline, 143f, 144
Dreissenidae, 234
Drip line, 283
Drip pits, 420, 421f
Dye recovery, 894, 894f, 895f
Dysderidae, 790
Dyspnoi, 796
Dystiscidae, 217—218
Dzhurinska Cave, 829t
E
Earthworms, 362
Eastern Alps, young uplift, 531
Ebb-and-flow springs, 801—802
Echolocation, 45
Ecological biogeography, 849-850
Ecological classification
accidentals, 276—277
Barr and, 276
colonization and genetic isolation and,
275
eutroglophile, 276—277
Gilbert and, 276
Joseph and, 275
occasional visitors, 275
overview about, 275
Pavan and, 276—277
Racovitza and, 275—276, 277
Schiner and, 275, 276, 277
Schiodte and, 275
selection and, 275
Sket and, 276
source and sink population and, 277
subterranean organisms, 276
subtroglophile, 276—277
Trajano and, 276
troglobites, 275—276, 277
troglophiles, 275—276, 277
trogloxenes, 275—276, 277
true vs. false cavernicoles, 276
Ecosystems
bats and, 51—54
chemoautotrophy and, 133—134
connectivity and disturbance in, 102
definition and boundaries, 99
energy flux and limitation, 100—101, lOOf
nutrients, 101-102
physical environment and habitat zones,
99-100, lOOf
Ecotones, 283
Ectothermy, 5—6
Edwards Aquifer, influent streams and,
708-709
Effective population size, 609
Efflorescent crusts, 676
Electrical resistivity imaging (ERI), 348—349,
349f
Elision, 751-752
The Elk Horn, Bobcat Cave, 79f
Ellison's Cave, 86
Embryonic cryptodolines, 141
Endokarst realm, 754- 756, 755f, 756f, 757f
Energy demand
activity and, 5
aquatic cave animals and, 4, 4t
back to network and, 6
body size conflicts and, 5
character reduction and, 6
ectothermy and neoteny and, 5—6
excitement and aggression and, 5
food scarcity, 4—6
hypoxic conditions and, 6
overview, 4
periodic starvation and, 9
terrestrial cave animals and, 4
Energy economy, 1
hypothesis, 553
Englacial caves, 354-356
Entities, 269
Entranceless caves
background regarding, 277—278
Bobcat Cave and, 280
Buckwheat Cave and, 280
Chestnut Ridge Cave System, 280
deciduous trees and, 279
entrance lifetime and, 278—279
environmental settings of, 278—279
examples of, 280
finding, 279—280
clastic ridges and, 279
digging bar and, 279
melted snow and, 279
sinks and, 279
obstacles to, 278
statistical theory regarding, 278
trash and, 279
variables that produce, 278
virgin passages and, 277
Entrances
as habitat, 283
human created, 281, 282, 282f
intrinsic, 281
lifetime, 278-279
locations of, 280
overview, 280
as paleontological/archaeological sites,
283-284
pit and shaft, 281
sinkhole, 281, 282f
sinking stream, 281
size scales of, 281, 282f
at springs, 281, 281f
statistics of, 282—283, 283f
talus, 284
truncation and collapse and, 281
types of, 281-282
valley deepening and, 281
Entrenched karst, 365, 372
Eogenesis, 751
Eogenetic karst, 437
Epigean limnic system model, 405
Epikarst. See also Shallow subterranean
habitats
as carbon dioxide reservoir, 286
concept, 284-285
926
INDEX
Epikarst ( Continued )
disagreement about, 285—286
in ecology, 284-285
in hydrogeology, 285
in karstology, 285
defined, 284, 434
erosion, 287
evapotranspiration and, 285
GPR and, 287
how it works, 286, 286f
hypotelminorheic zone and, 284
illustration of, 285f
karst evolution and morphology and,
286-287
karst landforms and, 434
as karst skin, 287
Mangin and, 284
NAPLs and, 169-170
overview about, 288
SSHs and, 685t, 686, 686f
subcutaneous karst and, 284
superficial karst and, 285
underground shallow medium and,
284-285
water
calcium and, 289
discharge and, 288
DOC and POC and, 289
temperature, 289
Epikarst communities. See also Shallow
subterranean habitats
biogeography and, 293—294
Copepodas, 291—293
ecology and, 294—295, 294f
environmental conditions, 288—289
morphological features, 293
sampling techniques, 288, 289f, 290f
taxonomy, 290—291
aquatic genera, 290—291, 291t, 292f
crustaceans, 290, 290t
terrestrial species, 291
Epikarstic aquifer, 284
Epsomite, 374, 5011, 503
stalactites, 375, 376f
Ereynetidae, 794
ERI. See Electrical resistivity imaging
Erosional caves, 865-866
Erosional tiankeng, 821f, 822, 822f
Erosion/surface water flow model, 406—407
Escape from epigean predators, 405
Escape hypothesis, 520
Eucarida, 178t, 191-193
Decapoda, 178t, 191,191f
Anomura, 178t, 192
Astacidea, 178t, 192,192t
Brachyura, 178t, 192, 193t
Caridea, 178t, 191,192t
Euhadenoecus fragilis, 610—611, 610f
Euhadenoecas puteanus, 610—611, 610f
Eukoeneniidae, 791
Eumalacostraca, 178t, 183—193
Eucarida, 178t, 191-193
Peracarida, 178t, 184—191
Syncarida, 178t, 183—184
Eupnoi, 796
European diversity patterns. See Diversity
patterns, in Europe
Eurycea, 667
blepsimolge, 668f, 669-670
E. chisholmensis, 670
E. latitans, 670
E. longicauda, 671, 672f
E. lucifuga, 671, 672f
E. nana, 670
E. naufragia, 670
E. neotenes, 669
E. pterophila, 670
E. rathbuni, 668, 668f
E. robusta, 669
E. sosorum, 669—670
E. spelaea, 670—671
E. tonkawae, 670
E. tridentifem, 670
E. troglodytes, 670
E. tynerensis, 671
E. waterlooensis, 669
other, 670—671
septentriomolge, 670
Texas cave and spring-dwelling, 667—668
typhlomolge, 668-669, 668f
Euscorpiidae, 788
Eutroglophile, 276—277
Evaporites dissolution, 758
Evapotranspiration, 285
Evolution of lineages
background about, 295
genetic differentiation models and,
297-298
molecular data overview, 301—303
subterranean habitat colonization,
295-297
hypotheses, 296
preadaptations and, 295—296
routes, 296
troglobites
single vs. multiple origin of,
298-299, 298f
troglomorphic traits, 299—301
constructive vs. regressive, 299—301
isolation variation and time, 299
Evolutionary ecology, 408—409
Evolutionary ecology, Nipliargus and
conclusions about, 562—564
distribution, 555
interactions within and between species
cannibalism, 559
chemoautotrophic bacteria
symbiosis, 560
predation and competition, 559—560,
560t
key elements comparative review,
560-562
communities: niche separation, 562,
563f
female body size, 562
morphological diversity and stasis,
560-561, 561f
key properties, 556-559
abiotic factors tolerance, 556-559
allometric growth, 556, 558f
antennae and pereopods, 556, 557f
body shape, 556, 556f, 557f
body size, 556, 557f
developmental patterns, 556, 558f
endemism/poor migratory abilities,
559
gnathopods, 556, 557f, 558f
light and, 557
maxillae I, 556, 558f
morphological variation, 556, 557f
response to hypoxia, 557-558
sexual dimorphism, 556, 558f
temperatures, 558—559
water chemistry tolerance, 559
overview about, 555—556
taxonomy challenges of, 555
Exaptations, 524
Exfiltration, 751—752
Exley, Sheck, 310, 312
Expedition caving, 308
Exploration, 103. See also Entranceless caves
basic equipment, 305—307
battery operated light sources,
305-306
carbide lamps, 305
clothing, 306
hard hats, 306, 306f
LEDs, 306
light sources, 305-306
packs, 306-307
blue holes, 160
Burnsville Cove, 75—77
Butler Cave Camp, 86
conclusions about, 309
conservation and, 309
Crystal Cave Expedition, 85—86
demands of, 85
Ellison's Cave, 86
European model for, 85
first camps used in, 85
Friars Hole System, 336
of glacier caves, 357
Jewel Cave, 411
Kazumura cave, 439
entrances, 439, 439f
local visitation, 439
survey and systematic exploration,
439
Krubera Cave, 444—445
large systems
expedition and project caving, 308
Lechuguilla Cave, 462—464
cartographic effort, 464
Chandelier Maze, 464
Coral Sea, 464
Deep Secrets, 463
Eastern Branch, 463—464
guano mining and, 462—463
before May 1986, 462
1970s, 463
1986, 463
Southwest Branch, 463
Rift, 463
Mammoth Cave System, 469—470
Mulu caves, 532
INDEX
927
need for, 277-278
Nullarbor Caves, 572
overview about, 304
PPCS, 622
Project SIMMER and, 86
reasons for, 644
safety, 309
skill levels, 304—305
large cave systems, 305
small, near-horizontal caves, 304
underwater caves, 305
vertical caves, 304
Sullivan Cave, 86
vertical caving, 307
ascending, 307, 316f, 317-318
from bottom up, 319—320
changing over, 318
descending, 307, 317
disclaimer, 314
overview regarding, 314—315
rappel, 307
rigging and, 315-316, 315f
rope and, 315
single-rope, 307
sit harness and, 316—317, 316f
traverses and tyrolians, 318—319,
319f
Vjetrenica, 863-864
Wakulla Spring underwater cave system,
878-880, 879f
water, 308-309
big river caves, 308
under water caves, 308—309
Wind Cave, and, 86
Expulsion regime, 751—752
Extraterrestrial caves, 404
Eyre Peninsula dolines, 147
F
Fl, 702
Facultative organisms, 491
Facultative visitors, 845
Fage, Louis, 219—220
Fantail, 438, 439f
Fat accumulation, 8
Faustloch, 702
Fern Bank salamander, 670
Ficco, Mike, 76
Fields, 269
Filter feeders, 20
Fishes, 20, 218, 854. See also specific fish
Fissure caves, 155
Fissure networks, 604f, 605
Fixed-beam model, 70, 71f
Flabellifera, 178t, 188, 189t
Flank margin caves, 156—159
dissolutional morphology of, 158f
formation of, 157, 158—159
freshwater lens and, 157, 157f
Hamilton's Cave, 160f
Isla de Mona and, 159, 159f
sea level and, 159
Sistema del Faro, 159f
typical, 158-159, 159f
water density and, 157
Flatworms. See Turbellaria
Flies, 362
Floor ice, 401—402
Florida. See also Wakulla Spring underwater
cave system, Florida
rivers, 708
Flowstone, 779f
Fluorescent dyes, 889, 890t
Fluorite, 5011, 506
Flutes, 420, 421f, 428
condensation, 425
defined, 679
transverse, 422
Fluviokarst, 436
Fluvio-karstic hollow, 153
Folklore. See Myth
Fontigens tartarea, 515—517
Food
adaptation to low
energy demand and, 4—6
factors concerning, 2f
food resources and, 1 —3
food scarcity types and, 1
food-finding and, 3—4
life history characters and, 6—8
overview, 1, 2f
periodic starvation and, 8—9
adaptive shift and, 11—12
behavioral adaptations and, 63, 64f
finding, 3—4
appendages and sensory equipment,
3
back to network and, 4
behavior, 3
other factors of, 4
Gammarus minus diet, 343—344
general scarcity of, 1
guano and, 363, 363f
Organ Cave and, 101
resources, 1—3
cave type influence, 3
chemoautotrophy and, 3
input, 1—3
scarcity types, 1
White Cave and, 60—61, 61f
Food sources
fecal types of, 331-334, 331t
bats and, 331t, 333
beetles and, 331t, 333
crickets and, 33It, 333
pack rats and, 3311, 332
quality, quantity, predictability, 3311,
332
raccoons and, 3311, 332
limited, 323—328
adaptations predicted from, 324—326
aquatic and terrestrial differences
from, 326
costs of, 324, 325t
exceptions proving the rule about,
326-328
glacial cycles and, 326
limestone vs. lava tube caves and,
327
observations and principles, 323—324
pollution and, 327
short time scale and, 327
specializations predicted from,
324-326, 325t
scientific method use
caveats and tradeoffs, 323
terrestrial habitats, 328—331
aquatic vs., 328—329, 328t
corpses and, 330
cricket eggs and, 330
extreme food types and, 329—331,
329t
leaf and twig litter and, 330—331
Forbesichthys agassizi, 465, 466t
Foreign key, 270
Forms, 270
Forti, P., 499
Founder events, 10—11
Fracture permeability, 384, 385f, 385t
Frasassi Cave System, Italy, 696f, 815—816,
815f
Freshwater cave, 240—241
Freshwater lens, 157, 157f
Friars Hole System
cave hydrology, 337-338, 337f
exploration history, 336
overview about, 334
paleohydrology, 338-339
cave, 338-339
cave age, 339
surface, 338
setting, 334-336
geology, 336
hydrology and hydrogeology,
334-336
location, 334, 335f
Frost shattering, 145
Frostwork, 380, 381f, 782
F-statistics, 609
Fuhl's Paradise Pit, 82
Fungus, 624
G
Galumnidae, 795
Gammarus fossarum, 653, 654t
Gammarus minus, 612
cave population origin, 345, 346f
cave streams and, 344
compound eyes and, 345—346, 347f
foods of, 343-344
general ecology of, 343-344
idealized subterranean drainage basin
and, 343, 343f
karst window and, 344—345
as model system, 347
natural selection and, 544—545
population density of, 343—344
population structure of, 345
reproduction, 343—344
springs and, 344
variation in, 344-345, 344f
Garden of Eden, 532
Garypidae, 792
Gastropoda, 20, 232, 233
Gate styles, 628-630, 629f, 630f
928
INDEX
Gating criteria, 628
Gating decision guide, 630-631
final results, 631
good reasons
to gate, 630-631
not to gate, 630
poor reasons
to gate, 630
to not gate, 630
Gelyelloida, 178t, 182
General food scarcity, 1
Genetic differentiation models, 297—298
Genetic distance, 609
Genetic isolation, 275
Genetic repertoire, 10
Genrikhova Bezdna Cave, 443, 444f
Geographic information systems (GIS),
264-265
Geoid, 348
Geophilomorpha, 539
Geophysics of location
AMT and, 349-350
cave radiolocation, 351—352, 351f
conclusions about, 352
dowsing and, 352
ERI and, 348-349, 349f
GPR and, 350, 350f
interferometry and, 350, 351 f
microgravity and, 348, 349f
overview about, 348
seismic modeling techniques and, 350
targets of, 348f
techniques, 348
TEM and, 352
thermal variation and, 350—351
tradeoffs, 348
Georgia Blind Salamander, 671—672
Ghost stories, 323
Gigalere Cave, 85
Gilbert, Janine, 101, 276
GIS. See Geographic information systems
Glaciated plateaus, 740— 742, 741f
Glacicavicola bathysciodes, 519f
Glacier caves, 105f, 106
basics about, 353
creep closure in, 354-355, 355f
cut-and-closure, 354—355, 355f
englacial, 354-356
exploration of, 357
formation processes, 354
keyhole and, 355-356, 356f
overview about, 353
recreational caving and, 642
subglacial, 356, 356f
temperature and, 353
thrust faults and, 354, 354f
Glacier milk, 724
Glaciokarst, 437
Glacio-karstic hollow, 153
The Glop Slot, Butler Cave, 78f
Gnaphosidae, 790
Gnathopods, 556, 557f, 558f
Golden Chamber, 859—860, 860f
Gonyleptidae, 796
Gostry Govdy Cave, 829t
Gouffre Berger Cave, 85
GPR. See Ground penetrating radar
Grand Canyon caves, 817—818, 818f
Gravity
sediments, 139
springs, 798— 799, 799f
Green Cathedral Cave, 536
Grike, 428-429, 429f
Grotto, Nittany, 74, 75
Ground penetrating radar (GPR), 287, 350,
350f
Guadalupe Mountains caves, 816—817, 816f
Guano
bat, 358
carnivorous, 358-359, 359f
community structure and succession
in, 360-361
conservation and, 242—243, 249
frugivorous, 358—359, 358f
hematophagous, 358-359, 359f
cave communities and science and,
357-358
conclusions about, 363—364
conservation and, 363
eutrophic caves and, 357
fauna associated with, 359—360
fauna examples, 360—361
basic taxa, 361
bugs, 362
cockroaches, 362
earthworms, 362
flies, 362
mites, 361—362, 361f
overview, 361
pseudoscorpions, 362
food web examples, 363, 363f
replacement, 626
sources, 358
studies of, 358
tropics diversity patterns and, 242
types of, 358-359
bats, 358
birds, 358
crickets, 358
Guanophages, 359
Gunung Api, 532—535, 537
Clearwater River Cave, 532—534
Clearwater River Passage, 534, 535f
Cobra-Bridge-Cloud Cave, 534
Sarawak Chamber, 534
Gunung Benarat, 535—536
Benarat Caverns System, 536
Melinau Gorge, 535-536, 535f
Terikan River Caves, 536
Gunung Buda, 536—537
Green Cathedral Cave, 536
Tardis Cave, 536f
Gunung Mulu National Park, 537
cave biology, 537
Gunung Api, 537
plant /animal species, 537
Gypsum, 500—503, 501t. See also Ukraine
giant gypsum caves
angel hair, 375
bulk crystals, 374—375
cave
conclusions about, 372—373
lengths and locations, 369t
modification of hypogene maze,
371f, 372
Ordinskaya, 372, 372f
patterns and morphology, 372, 373f
sketches, 366f
cave cotton, 375
chandeliers, 374—375
crusts, 374, 375
as evaporite mineral, 374
hypogene conduits and, 367f
karst types, 364-365
deep-seated karst, 365, 367
entrenched karst, 365, 372
intrastratal karst, 365, 367
open karst, 365, 366-367
subjacent karst, 365, 371
syngenetic karst, 365
needles, 374-375, 375f
Nullarbor Caves and, 574, 574f
occurrence of, 364-365
speleogenesis
chemistry and kinetics, 365—366
karst types and, 366—372
transverse hypogene, 367-368, 368f
speleothems and, 374—375
structure of, 777-778, 778f
Gypsum flowers
formation of, 374, 375, 375f, 782, 782f
overview about, 374
sources and depositional mechanisms,
375-377
Gyrinophilus, 672—674, 673f
G. gulolineatus, 673f, 674
G. palleucus, 673, 673f
G. porphyriticus, 672, 673f
G. subterraneus, 673f, 674
H
Hadenoecus subterraneus, 744— 746, 745f, 746f
Hahniidae, 790
Haideotriton wallacei, 671—672
Half-tubes, 422, 423, 424f
Halides, 501t, 505-506
fluorite, 501t, 506
halite, 501t, 506
overview, 501t, 505-506
Halite, 5011, 506
Nullarbor Caves and, 574, 574f
Halocline, 157
Halosbaena tulki, 214f
Hamilton's Cave, 160f
Haplozetidae, 795
Hard hats, 306, 306f
Harpacticoida, 178t, 182
Harvestment, 795- 797, 796f
Hawaii. See also Kazumura cave, Hawaii
adaptive shift and, 12—16
cixiid planthoppers, 13—14, 13f
crickets, 15
moths, 15—16
Oliarus polyphemus, 14, 14f
species-pairs occurring, 12t
INDEX
929
terrestrial Isopoda, 14—15
wolf spiders, 15
root communities in, 661t
volcanic rock caves, 866t
Head flattening, 518, 518f
Headwater Complex, Castleguard Cave,
89-90
Heavy metals
chemistry, 162—164
general, 162-163
interactions, 162,163f
iron and manganese, 163—164
pH, 162-163
redox, 162—163
trace and contaminant metals, 164
conclusions about, 165—166
defined, 161,162t
forms of, 162
potentially toxic, 162t
sources, 161—162
anthropogenic, 161—162, 163t
in karst settings, 161, 162f
natural, 161,163t
spring water concentrations, 165f
storage and transport, 164—165
schematic scenarios, 164
speleothems and cave deposits, 164
storm-enhanced, 165
suspended and bed sediments,
164-165
storage compartments, 163f
Heavy nonaqueous phase liquids
(DNAPLs), 391
Helictite Cave, 81, 81f
Helictites
beaded, 379, 380f
bushes, 382-383, 382f
formation of, 782, 782f
subaerial, 379
subaqueous, 381—383, 381f, 382f
Hemiptera, 259
Hercegovina, 236
Heteroptera, 259
Heterotrophs, 126—127, 490—491
Heterozygosity, 609
Hewitt, Bevin, 74, 75
Hexapoda, 258-259
Hexatelidae, 789
Hibernation, of bats, 48-49, 48f
Hill, C. A., 499
Hirudinea, 245t, 913-914, 913f
Historical biogeography, 849-850
Hjulstrom's Diagram, 138, 138f
Hohlaub area, 703
Holarchaeidae, 790
Holloch Cave, Switzerland, 85, 726, 727f
Homoptera, 259
Hon Chong hills, 249-250, 250f
Hondo Underground Rescue Team (HURT),
649
Hooded tickspider, 793, 793f
Hosterman Pit, 278
Howe Caverns, 740— 741, 741f
Hubbardiidae, 788- 789
Huehue Cave, 869f
Huehue Flow, 869—870, 871f
HURT. See Hondo Underground Rescue
Team
Hwanseon Cave, 696f
Hybridization, 11
Hydrogen
chemoautotrophy and, 127—128
sulfide, 758
Hydrogeology of karst aquifers. See Karst
aquifer, hydrology of
Hydrographs, 388-389, 389f
Hydromagnesite, 500
balloons, 416-417, 416f
Hydrothermal caves
chemistry, 393—394
mixing water dissolution, 393
rising thermal water dissolution, 393
subaqueous and subaerial
dissolution, 393—394, 394f
sulfides oxidation, 393
defined, 391-392
deposits, 396
character of, 396
mineralogy, 396
identification, 397
life, 396-397
morphology, 394-396
anastomotic caves, 395-396
bush-like caves, 395, 395f
caves with cupolas, 395
individual chambers, 394
isometric rooms, 394
network caves, 395
phreatic maze caves, 395—396
single-conduit caves, 394
small-scale, 396
solution porosity, 394
spongework caves, 396
regional extent, 397
Hydrothermal karst, 437
defined, 391-392
in noncarbonate rocks, 393
oxidation of sulfide ores and, 393
settings, 392-393
deep seated, 392
endokarst, 392
shallow, 392-393
Hydroxides, 501t, 505
Hydroxylapatite, 501t, 504-505
Hydryphantidae, 795
Hyena, 594
Hygrobatidae, 795
Hygropetric cave, 236
Hypochthoniidae, 795
Hypogenetic karst, 437
Hypogenic speleogenesis, 756-764
artesian, 759
basic concept, 750
chemical mechanics, 756—759, 756f, 757f
carbonic acid dissolution, 756—758
evaporites dissolution, 758
hydrogen sulfide, 758
mixed carbonate/sulfate strata
dissolution, 759
organic acids, 758
sulfuric acid, 758
thermal water cooling, 758
water mixing, 758
combined artesian/deep-rooted, 759
defined, 750
diagenetic environments, 750—756, 751
eogenesis and, 751
mesogenesis and, 751
telogenesis and, 751
distribution and settings, 759
endokarst realm, 754- 756, 755f, 756f, 757f
groundwater systems and regimes,
751-752
hydrogeologic and structural controls,
759- 761, 760f
hydrogeologic cycle, 752-753, 753f
karst settings evolution, 753— 754, 754f
mesomorphology features, 763— 764
overview, 748- 750, 749f
paleokarst and, 764
porosity patterns, 761—763
cavernous edging, 761, 762
collapse shafts, 761, 762
isolated chambers, 761, 762
network maze, 761, 762
single or rudimentary networks,
761-762
spongework maze, 761, 762
steeply-inclined passages or shafts,
761, 762
Hyporheic zone, 686-688, 687f, 688t
Hypotelminorheic zone, 284
Hypoxic conditions, 6
Hypsimetopidae, 212
I
Ice blocks, 401—402
Ice caves, 236
Ice Crawls, Castleguard Cave, 92f
Ice in caves
bamboo, 402, 403f
cold air trapping and, 400—401
cryogenic cave micropearls, 401—402, 403
crystallography and mineralogy and, 403
extraterrestrial caves and, 404
floor ice, 401—402
high latitude/altitude and, 399—400
ice blocks, 401—402
ice stalagmites, 402
isotope analysis and, 402—403
lake ice, 401—402
mechanisms causing, 399, 400f
overview about, 399
pollution and, 403
radiocarbon dating and, 402
Scarisoara Ice Cave, 403
scientific significance of, 402—404
snow and, 400
speleothems, 401f
tourism and, 403—404
ventilation cooling and, 401
Ice wedging, 72
Icicles, 575, 575f
Idealized subterranean drainage basin, 343,
343f
930
INDEX
Ideoroncidae, 792
Illinois structural basin, 734f, 736
Imprints, 871
Inception doline, 143f, 144
Incised passages, 606f
Indiana, southern, 736—737
Indirectly adaptive hypothesis, 342
Inficionado Peak, 637, 637f
Infiltration regime, 752
Inflation, 867, 868f
Influent streams, 707— 709
cave evolution and, 707, 708f
Danube River and, 707
Edwards Aquifer and, 708— 709
evolutionary sequence of, 710—712
case A, 711, 711f
case B, 711, 71 If
case C, 711, 711f
Florida, 708
function of, 709—710
Kentucky, 709, 709f, 71 Of
overview, 707
Innerbergli area, 702
Insecta, 216-218, 259-260
Blattodea, 216-217
Coleoptera, 259-260
Dystiscidae, 217—218
Collembola, 216
Dermaptera, 259
Diplura, 216
Diptera, 260
Hemiptera, 259
Heteroptera, 259
Homoptera, 259
Orthoptera, 217, 259
Planthoppers, 216
tropical diversity patterns, 245t
Zygentoma, 216
Insolation, 581, 582f
Interferometry, 350, 351 f
Internal runoff, 387—388
Interpolators, 477—479, 477f
Intersection dolines, 150f, 152
Intrastratal karst, 365, 367
Invasion
active dispersal models, 405
active colonization model, 405
adaptive-shift model, 405
climatic-relict model, 405
competitive pressure, 405
deep sea colonization, 405
epigean limnic system model, 405
escape from epigean predators, 405
marine shallow-water ecotones
model, 405
conclusions about, 409
ecological evidence, 407-408
evolutionary scenarios, 405—407
overview about, 404—405
passive dispersal models, 406—407
erosion/surface water flow model,
406-407
rafting model, 406
regression model, 406, 407f
surface water infiltration model, 406
unique vs. multiple, 407
Iron
bacteria, 133
chemistry and, 163—164
chemoautotrophy and, 129, 133
Ischyropsalidae, 796
Isla de Mona, 159,159f
Isolated chambers, 761, 762
Isometric rooms, 394
Isopoda, 178t, 187-188,189t, 258
adaptive shift and, 14—15
anchihaline habitats and, 22
Anthuridea, 178t, 188, 189t
Appalachian stream competition and,
746- 747, 746f, 747f
Asellota, 178t, 188-190,189t
in Australia, 212—213
Calabozoidea, 178t, 189t, 190
Cirolanidae, 212—213
in Dinaric karst, 233—234
Flabellifera, 178t, 188,189t
Hypsimetopidae, 212
Microcerberidea, 178t, 188,189t
Oniscidea, 178t, 189t, 190, 213
Phreatoicidea, 178t, 188,189t
Ixodidae, 794
J
Japanese folklore, 322—323
Jean-Bernard Cave, 731, 731f
Jeannel, Rene, 219—220, 404—405
Jefferson's ground sloth, 596, 597f
Jewel Cave, South Dakota
barometric airflow, 411
biology, 415—416
bats, 415-416
cave life, 415
microbiology, 416
calcite crystals of, 411, 412f
conclusion about, 417
geology, 413-415
cave fill, 414—415
Minnelusa Cap/caves relationship,
414, 414f
recent mapping, 414
history, 411
early, 411
early exploration, 411
later exploration, 411
hydrology, 415
levels, 413f
basement level, 412
chert level, 412
loft level, 412
lower level, 412
main level, 412
subchert level, 412
origin, 415
working hypothesis, 415
other Black Hills caves, 417
physical characteristics, 411—413
breakdown modification, 412—413
passage sizes, 413
three-dimensional system, 411—412,
413f
speleothems, 416—417
hydromagnesite balloons, 416—417,
416f
popcorn stalagmites, 416
Jollyville Plateau salamander, 670
Joseph, G., 275
Jubilejna Cave, 829t
Julida, 540, 540t
K
K2, 702- 703
Kakuk, Brian, 31 If
Kamenica, 859-860, 860f
Kamenitza, 140
Kane caves, Wyoming, 813, 813f, 814t
Karren, cave
biological activity influencing, 419—420,
420f
clastic material and, 420
development, 419
feature descriptions, 420—425
anastamose, 422, 423f
anastomoses, 422
avens, 423
bell holes, 423-424, 424f
bypass tubes, 423
chimneys, 423
condensation corrosion hollows, 425,
425f
condensation rills/flutes, 425
corkscrews, 423
corrosion notch, 425
cupola, 424, 425f
cusps, 425, 425f
dissolution basins, 420, 421f
dissolutional potholes, 420, 422f
drainage grooves, 422
drip pits, 420, 421f
half-tubes, 422, 423, 424f
laugdecke, 425
paragenesis, 422
pendants, 422, 423f
rills/flutes, 420, 421f, 425
scallops, 420-422, 422f
scour marks, 422
solution pockets, 423, 424f
spongework, 422
transverse flutes, 422
waterline notch, 425
flow and, 419
form controls, 419—420
overview about, 419
photokarren and, 419—420, 420f
rock properties influencing, 419, 420f
splitkarren, 428—429
Karren, surface
assemblages, 430
alpine karrenfields, 430
coastal karren, 425f, 430
pinnacle karst, 430
ruiniform karrenfield, 430
basic controls, 426—427
biological activity, 427
climate, 427
hydrodynamic, 426, 426f
INDEX
931
lithological factors, 427, 427f
dissolutional pavements, 427, 427f
features
basins, 427—428
cavernous karren, 429, 429f
channels, 428
corrosion notch, 429—430
cutter, 429
decantation runnel, 428
flute, 428
grike, 428-429, 429f
pans, 427-428, 428f
pits, 427-428
rillenkarren, 428, 428f
solution runnels, 428, 429f
spitzkarren, 429
overview about, 425—426
Pinnacles of Mulu, 426, 426f
positive forms, 426
San Salvador Island, 426f
Karst
coastal, 437
denudation, 508
ecosystems
connectivity and disturbance in, 102
definition and boundaries, 99
energy flux and limitation, 100—101,
lOOf
nutrients, 101—102
physical environment and habitat
zones, 99—100, lOOf
endokarst realm, 754- 756, 755f, 756f, 757f
eogenetic, 437
history surrounding, 432—433
hypogenetic, 437
hydrothermal, 437
sulfuric acid, 437
landforms, 433—436
caves and, 436
closed depressions, 434—435
cones and towers, 435—436, 436f
overview, 433
pavements and epikarst, 434
poljes, 435
mixed origin landscapes, 436—437
alpine karst, 436—437
fluviokarst, 436
glaciokarst, 437
mixing-zone, 437
overview about, 430—431
settings evolution, 753- 754, 754f
telogenetic, 437
window, 344—345
word origin, 431—433
karst triangle and, 432, 432f
Kras plateau and, 431 —432, 432f
location and, 431—432
world distribution, 431, 431f, 432t
Karst aquifer, hydrology of, 384—388
characteristics, 388-390
base flow/area relationships,
389-390
clastic sediments, 390
conduit systems carrying capacity,
389
hydrographs, 388-389, 389f
contaminant transport, 390—391
DNAPLs, 391
LNAPLs, 391
metals, 391
pathogens, 391
water-soluble compounds, 391
groundwater basins/flow systems,
387-388
overview, 383
piracies and spillover routes, 388
porous media aquifers and, 383—384
standard model, 387—388, 387f
allogenic recharge, 387
dispersed infiltration, 387
internal runoff, 387—388
triple porosity model, 384—385
conduit permeability, 384, 385—387,
385f, 385t
fracture permeability, 384, 385f, 385t
matrix permeability, 384, 385f, 385t
types, 385, 386f
water quality, 390
water supply, 390—391
wells and, 390
water table and, 388
Karst aquifers, modeling of
chemistry, 508-510
calcium equilibrium, 508—509, 509f
calcium flux rate, 509—510, 509f
evolution, 510—512, 512f
three-dimensional, 510—511, 512f
two-dimensional, 510
overview about, 508
single fracture, 510, 51 Of
Karst aquifers, water tracing in
definitions and objectives of, 886—887
dry set system, 892
dye amounts to use, 891
history, 887-889, 887f
passive detector types
activated carbon, 890
cotton, 891
plankton netting, 891
Quinlan Gumdrop, 891, 891f
qualitative
using passive detectors, 890-892
quantitative, 892—893
test interpretation, 893-896, 894f
dye recovery or breakthrough curve
and, 894, 894f, 895f
tracers and, 889-890
categories of, 889
cautions, 890
fluorescent dyes, 889, 890t
ideal characteristics, 889
Kartchner Caverns, 694f
Kazumura cave, Hawaii
description, 439—442
braided passages, 440—441, 441f
canyon-like passages, 439—440, 440f
extraneous tubes, 441
intermediate roofs, 441—442, 441f
multilevel/lavafall development,
441-442
plunge pools, 441—442, 442f
speleogenesis, 439-440
thermal erosion, 440—441, 440f
exploration history, 439
entrances, 439, 439f
local visitation, 439
survey and systematic exploration,
439
fantail and, 438, 439f
lava cave speleothems, 442
primary, 442
secondary, 442
overview about, 438
physical setting, 438
human impacts, 438
location, 438, 438f
surface relationship, 438
Kebara Cave, 95f, 98, 98f
Keller Well, 471f
Kentucky. See also Mammoth Cave System,
Kentucky
influent streams, 709, 709f, 71 Of
solution caves, 736— 737
Keyhole, 355-356, 356f
Komsomol'ska Cave, 829t
Kras plateau, 431—432, 432f
Kriging, 477—479
Kristal'na Cave, 829t
Krosswig, Curt, 544
Krubera Cave
Arabika Massif
deep cave evolution, 449—450
geology, 447—448, 447f
hydrogeology, 448—449
location and physiography, 443, 444f
Ortobalagan Valley in, 443, 445f
block-fault structure and, 445—447
branches, 445, 445f
combined cave profile, 446f
depth of, 443
evolution of, 449—450
exploration since 1999, 444—445
first documentation of, 443
genesis of, 729, 730f
hydrology, 447
overview about, 443
as shaft and canyon cave, 729, 730f
temperatures, 447
Kujbyshevskaja Cave, 443, 444f
L
Labeled substrate experiments, 493—494
Labyrinth caves, large, 726—729
Holloch Cave, 726, 727f
Reseau Siebenhengste-Hohgant system,
726- 729, 728f
Lake ice, 401—402
Lamarckism, 520
Lampenflora
appropriate lighting, 455, 455f
composition and ecology, 452—453
control and restriction of, 454—455, 454f
cyanobacteria, 453
definitions, 451—452
dispersion modes, 452—453
932
INDEX
Lampenflora ( Continued )
diversity, 453
factors influencing, 452
future management, 455
habitat alteration and, 453—454
killing, 455
mosses, 453
overview, 451
in Postojnska Jama, 452f
problems associated with, 454
surfaces and, 453
temperature and, 452
tourism and, 451—452
Landscape evolution
multilevel caves, 529—531
Cumberland River incision pulses,
530, 530f
eastern Alps, Austria, young uplift,
531
Mammoth Cave, 529—530, 530f
Sierra Nevada incision rates, 531
southern Rocky Mountains erosion
rates, 531
Laniatores, 796
Lankester, E. R., 404—405
Lanzaia vjetrenicae, 231f
Large labyrinth caves, 726—729
Holloch Cave, 726, 727f
Reseau Siebenhengste-Hohgant system,
726- 729, 728f
Lateral potholes, 682f, 683
Lattice data, 475
Laugdecke, 425
Lava cave, 641. See also Root communities, in
lava tubes
double-trunked tunnels, 869—870
rise, 871-872
single-trunked tunnels, 869
speleothems, 442
primary, 442
secondary, 442
superimposed-trunked tunnels, 870
tubes, 327, 658
Lavafall development, 441—442
Lechuguilla Cave, New Mexico, 816—817,
816f
biology, 459
Capitan Formation, 457—458, 457f
conservation and, 456
description, 460—462
boneyard maze, 461 —462, 462f
fissures, 461, 462f
horizontal passages, 461, 462f
vertical tubes, 461
discovery of, 456
exploration history, 462—464
cartographic effort, 464
Chandelier Maze, 464
Coral Sea, 464
Deep Secrets, 463
Eastern Branch, 463—464
guano mining and, 462—463
before May 1986, 462
1970s, 463
1986, 463
Southwest Branch, 463
Rift, 463
folds within vicinity of, 458
four units spanned by, 457—458, 457f
geologic setting, 457
as hypogene cave system, 458
map, 461f
microbial influence in, 456
mineralogy, 459—460
paleontology, 460
physical setting, 456—457, 457f
Queen Formation, 457—458, 457f
regional/local hydrogeology, 458-459
rooms and passages, 458, 459f
Seven Rivers Formation, 457—458, 457f
stratigraphy and structure, 457—458
sulfuric acid and, 456, 460—461
as world-class cave, 456
Yates Formation, 457-458, 457f, 458
LEDs. See Light-emitting diodes
Leeches, 913-914, 913f
Leeuwenhoeckiidae, 794
Legend. See Myth
Leiodidae, 221, 223f
Lepe Jame, 622, 622f
Leptodirini group, of beetles, 57, 58f
Leptodirus hochenwartii, 55, 55f, 231f, 519f
phylogenetic Mediterranean tree, 59, 60f
Leptodirus hochenwartii, 55, 55f, 231f, 519f
Leptonetidae, 790
Leptostracans, 21
Life history characters
adaptation and, 6—8
back to network and, 8
case study, 7-8, 8t
egg size and, 6—7, 7f
growth rate and, 7
longevity and, 7
Life history evolution
of Amblyopsidae, 465—466
Amblyopsis rosae, 465, 466, 466t,
467-468
Amblyopsis spelaea, 465, 466, 466t,
467-468
Chologaster cornuta, 465, 466t
Forbesichthys agassizi, 465, 466t
life history characteristics, 466—467,
466t
Typhliclithys subterraneus, 465, 466t
high reproductive effort paradox and,
467-468
Light nonaqueous phase liquids (LNAPLs),
391
Light-emitting diodes (LEDs), 88, 306
Lighting, show cave, 694—695
Limbodessus eberhardi, 206f
Limestone caves, 658. See also specific cave
Limnesiidae, 795
Limnohalacaridae, 795
Limonite, 163—164
Linkages, 270
Linyphiidae, 790
Liochelidae, 788
Liocranidae, 790
Lithobiomorpha, 539—540, 539f
Lithological contact poljes, 154, 154f
Littoral cave, 155
LNAPLs. See Light nonaqueous phase
liquids
Long inclined passages, 729— 732
Cassowary System, 731—732, 732f
Jean-Bernard Cave, 731, 731f
Looping passages, 724, 724f
Losing streams, 707—709
cave evolution and, 707, 708f
Danube River and, 707
Edwards Aquifer and, 708— 709
evolutionary sequence of, 710—712
case A, 711, 711f
case B, 711, 711f
case C, 711, 711f
Florida, 708
function of, 709—710
Kentucky, 709, 709f, 71 Of
overview, 707
Loubens, Marcel, 85
Low food adaptation
energy demand and, 4—6
factors concerning, 2f
food resources and, 1—3
food scarcity types and, 1
food-finding and, 3—4
life history characters and, 6—8
overview, 1, 2f
periodic starvation and, 8—9
Lungless salamanders, 667—674
Lycosidae, 790
M
Machaerites spelaeus, 231f
Macrochelidae, 793
Magnetostratigraphic time scale, 586, 587f
Makua Cave, 322
Malachite, 500, 50It
Malacostraca, 178t, 183-193, 257-258
Eumalacostraca, 178t, 183—193
Eucarida, 178t, 191-193
Peracarida, 178t, 184—191
Syncarida, 178t, 183—184
Phyllocarida, 183f, 183, 178t
Mammalia, 261—263, 262t
Mammals, as cave visitors
cave ecology impact of, 847
cave-dwelling species, 845—846
facultative visitors, 845
regional cavernicoles, 845
regular cavernicoles, 845—846
den sites and, 846—847
foraging habitats and, 847
knowledge gaps, 846
man and, 848—849
rare species, 848
resting and mating and, 846
wintering sites and, 847
Mammoth Cave System, Kentucky
ancient cavers
archaeological evidence, 27—28, 27f,
28f
archaeology, 26—27
chronology and, 27
INDEX
933
evidence interpretation, 28—30, 29f
prehistoric archaeology, 27—30
biology, 473—474
Cleaveland Avenue, 472f
dating, 175-176, 737
entrances, 470, 470f
exploration, 469—470
geographic and geologic setting, 471
history, 470
idealized profile through, 737, 737f
as international research magnet, 474
Keller Well, 471f
landscape evolution and, 529—530, 530f
minerals, 472—473, 473f
origin, 473
overview about, 469, 470f
passage patterns, 471—472, 471f, 472f
prehistoric explorers and miners, 469—470
recent/ongoing exploration, 469
as solution caves, 737, 737f
Manganese
bacteria, 133
chemistry and, 163—164
chemoautotrophy and, 129, 133
Mangin, A., 284
Mangkurtu mityula, 206f
Manmade show caves, 643
Mapping subterranean biodiversity
data types, 475
lattice data, 475
point patterns, 475
interpolators and, 477—479, 477f
kriging and, 477—479
modeling issues, 479—481
aggregation and scale, 479—481, 480f
excess zeroes, 481
overview about, 474—475
spatial pattern investigation, 476—477
spatial autocorrelation, 476—477
spatial process decomposition, 475—476
example, 476, 476f
large-scale trend, 475
mesoscale or small-scale variation,
475
random error or unexplained
variability, 475—476
zero-inflated spatial models and, 478—479
Marginal cave, 18
Marine caves, 240
Marine regressions
background, 482
change rate evaluation, 486
dating from stygofauna, 484—486
diverse habitat adaptability, 482
from sea to freshwater and land, 482
thalassoid limnostygobionts
biogeography of, 483, 484f, 485f
Morocco and, 483—484
as sole evidence, 483—484
Marine shallow-water ecotones model, 405
Martel, Edouard A., 103, 433
Martino's vole, 845—846, 846f
Masodon bones, 875— 876, 876f, 877f
Material compensation, 520
Mating behavior, adaptive shift and, 11
Matrix permeability, 384, 385f, 385t
Mats, microbial, 496, 498f
Maxillopoda, 178t, 180-182, 257
Copepoda, 178t, 181 — 182
Calanoida, 178t, 181,181t
Cyclopoida, 178t, 181-182
Gelyelloida, 178t, 182
Harpacticoida, 178t, 182
Misophrioida, 178t, 181, 182t
Platycopioida, 178t, 181
Mystacocarida, 178t, 180—181
Tantulocarida, 178t, 180
Maya, 322, 836
Maya caves
ancient utilization
antiquity of, 487
cave art, 488, 489f
cave burial, 487—488
cave modifications, 488, 488f
general characteristics of, 487
artificial, 489f, 490
community and, 489—490
ethnohistory and ethnography, 488—489
map, 487f
overview about, 486—487
Maze caves, 395-396
Maze development
aggressiveness boosts, 601
floodwater recharge and, 601
short flow paths and, 601
uniform recharge and, 601
Melinau Gorge, 535-536, 535f
Melinau Limestone, 532
Mesocaverns, 658
Mesogenesis, 751
Mesostigmata, 793
Mesothelae, 789
Metabolic assays, 492
Metabolic span, 7
Metastigmata, 794
Meteoric circulation regime, 752
Methane cycle, 133
Mexico. See also Sistema Huautla, Mexico
camps and, 86—87, 87f
Cueva de Villa Luz, 813-815, 813f, 814t,
815f
Micro whipscorpions, 791
Microaerophilic organisms, 491
Microbes. See also Chemoautotrophy
actinomycete colonies, 494, 497f
cave and karst diversity of, 494—497, 495f
diversity and metabolism, 490—491
aerobes and anaerobes and, 490f, 491
energy-yielding metabolic reactions,
490-491, 490f
future regarding, 497—498
history of studies on, 491 —492
early studies, 491
1960s through 1980s, 491
1980s and 1990s, 491-492
technology and, 492
methodological approaches and, 492—494
culturing and metabolic assays, 492
labeled substrate experiments,
493-494
microscopes, 494
molecular, 492—493
stable isotope ratio analysis, 493—494
microbial mats and, 496, 498f
naked eye and, 494, 496f
overview about, 490
Proteobacteria, 495—496
vermiculations, 496, 497f
Microbial mats, 496, 498f
Microcerberidea, 178t, 188,189t
Microcracks propagation model, 71
Microgravity, 348, 349f
Microstigmatidae, 789
Microzetidae, 795
Mictacea, 22,178t, 185-186, 186f
Middle East
anthropogenic deposits, 96
cave occupations, 97—99
Chalcolithic and Bronze age, 98
Middle Paleolithic, 98
Neolithic, 98
prehistoric, 97—98
Upper and Epi-Paleolithic, 98
deposits and processes, 96-97
Kebara Cave, 95f, 98, 98f
location, 95f, 96
overview about, 94—96
prehistoric sites, 95f
present environment, 96
Qafzeh Cave, 95f, 98f
Tabun Cave, 95f, 97-98, 97f
Midnight Cave, 86
Migration, 609
Milieu souterrain superficiel (MSS), 686f,
688, 688f
Millipedes, 540—541
Cambalopsidae and, 242, 244f
Chordeumatida, 540t, 541
Julida, 540, 540t
map of, 540f
overview, 540
Polydesmida, 540, 540t
Pseudotremia, 541, 541f
Scoterpe s, 541
Milyeringa veritas, 214f
Mimetidae, 790
Mineral springs, 800
Minerals. See also specific mineral
arsenates and vanadates, 506
books on, 499
carbonates, 500, 501t
aragonite, 500, 5011
calcite, 500, 5011, 503f
coralloid multiaggregates and, 500,
503f
deposition of, 500
hydromagnesite, 500
malachite, 500, 50It
halides, 501t, 505-506
fluorite, 5011, 506
halite, 5011, 506
overview, 501t, 505—506
hydrothermal caves and, 396
ice in caves and, 403
Lechuguilla Cave, 459—460
934
INDEX
Minerals ( Continued )
Mammoth Cave System, 472—473, 473f
native elements, 506
Nullarbor Caves, 574—575
calcite, 574
gypsum, 574, 574f
halite, 574, 574f
organic, 507
overview, 499—500, 501t
oxides and hydroxides, 5011, 505
phosphates, 504-505
brushite, 501t, 504—505
development of, 504, 504f
hydroxylapatite, 501t, 504-505
overview, 501t, 504
taranakite, 501t, 504—505
PPCS, 623
reaction types creating, 499
reasons to study, 507
Siebenhengste cave system, 704— 706, 706f
silicates, 501t, 505
allophane, 501t, 505
opal, 5011, 505
origins, 505
other, 505
quartz, 501t, 505
speleothems and, 777—778, 885—886
agagonite, 777
calcite, 777
gypsum, 777-778, 778f
stalactites and, 805-806, 806f
stalagmites and, 805-806, 806f
sulfate mineral crystals
overview about, 374
sources and depositional
mechanisms, 375—377
types of, 374
sulfates, 500-504, 5011
alunite, 501t, 503
barite and celestine, 5011, 503
deposition mechanisms, 503—504
epsomite, 501t, 503
gypsum, 374, 500-503, 5011, 574,
574f, 777-778, 778f
mirabilite, 501t, 503
sulfides, 507
Mines, 642
Minnelusa Cap, 414, 414f
Mirabilite, 374, 5011, 503
stalactites, 375, 376f
Misophrioida, 178t, 181, 182t
Mississippi Valley-type (MVT) ore deposits,
818-819
Mitchell, Simon, 313
Mitchell Plain, 722
Mitchell Plateau, 736
Mites, 256, 361-362, 361f, 793- 795
Mixed carbonate/sulfate strata dissolution,
759
Mixing-zone, 157
karst, 437
Mixotrophs, 126—127, 490—491
Mlynki Cave, 829t
MMS. See Superficial underground
environment
Model organisms
Astyanax mexiccinus, 37
evolution and adaptation and, 37
Modern synthetic theory of evolution, 554
Mollusca
biogeographic patterns, 514t, 515
Dinaric Mountains, 515, 515f, 516f
United States, 514t, 515
Europe, 221, 224f
overview about, 512—513
Paraegopis oberwimmeri, 512—513
Pisidium, 513—514
prosobranchs, 514—515
pulmonates, 514—515
selected species, 515-517
Congeria kusceri, 512—513, 517
Fontigens tartarea, 515—517
taxonomic patterns, 513—515, 513t
tropical diversity patterns, 245t
United States diversity patterns, 255
Zospeum, 513
Monolistra monstmosa, 231f
Monte Baldo dolines, 147
Montello dolines, 145—147, 146f, 148f
Monti Lessini dolines, 144f, 145
Morocco, 483—484
Morphine Waterfalls, Barberry Cave, 80f
Morphological adaptations
in Amblyopsidae, 522-523, 522f, 525, 525f
analysis examples, 525—527
Amblyopsidae, 522f, 525, 525f
Pseudosinella hirsuta, 526, 526f
in Collembola, 523, 523f
conclusions about, 527
in crayfish, 521, 522f
environmental factors effects, 521—522
evidence, 522—523
exaptations or preadaptations and, 524
MMS and, 524-525
occurrence of, 523-525
overview about, 517
progressive, 521
factors needed for, 521
regressive evolution hypotheses, 519—521
troglomorphy, 517—518
common characteristics of, 518—519,
520t
convergence and, 518, 518f, 519f
evolution of, 518-519
Mosses, 453
Moths, 15-16, 663
MSS. See Milieu souterrain superficiel
Muchimuk-Colibri system, 635—636, 636f
Multilevel caves
landscape interpretation using, 529—531
Cumberland River incision pulses,
530, 530f
eastern Alps, Austria, young uplift,
531
Mammoth Cave, 529—530, 530f
Sierra Nevada incision rates, 531
southern Rocky Mountains erosion
rates, 531
overview about, 528
representation of, 528f
river incision and cave record, 529
sediments and, 529, 529f
water table and, 528-529
Mulu caves, Malaysia
bird's nest soup and, 532
exploration, 532
Gunung Api, 532-535
Clearwater River Cave, 532—534
Clearwater River Passage, 534, 535f
Cobra-Bridge-Cloud Cave, 534
Sarawak Chamber, 534
Gunung Benarat, 535-536
Benarat Caverns System, 536
Melinau Gorge, 535-536, 535f
Terikan River Caves, 536
Gunung Buda, 536—537
Green Cathedral Cave, 536
Tardis Cave, 536f
Gunung Mulu National Park, 537
cave biology, 537
Gunung Api, 537
plant/animal species, 537
Melinau Limestone and, 532
overview, 531-532, 533f
Pinnacles of Mulu, 426, 426f
southern hills, 532
Deer Cave, 532, 534f
Garden of Eden, 532
Munidopsis polymorpha
aggressive behavior of, 66—67, 67f
reproductive behavior of, 63, 64f
Mushkarova Jama Cave, 829t
Mutualism, 743, 744f
MVT. See Mississippi Valley-type ore
deposits
Mygalomorphae, 789
Myriapoda
in Australia, 216
Chilopoda, 216
Diplopoda, 216
Centipedes, 539-540
Craterstigmatomorpha, 539
Geophilomorpha, 539
Lithobiomorpha, 539—540, 539f
overview, 539
Scolopendromorpha, 539
Scutigeromorpha, 539
Millipedes, 540—541
Chordeumatida, 540t, 541
Julida, 540, 540t
map of, 540f
overview, 540
Polydesmida, 540, 540t
Pseudotremia, 541, 541f
Scoterpes, 541
overview about, 538, 538f
Pauropods, 538
Symphylans, 538
tropical diversity patterns, 245t
Mysida, 21-22, 178t, 185, 186t
Mysmenidae, 790
Mystacocarida, 178t, 180—181
Mystery Cave, 742
Myth
Amarnath Cave and, 322
INDEX
935
Bible and, 322
Chinese, 322
creatures of, 321
Crete and, 321
ghost stories and, 323
Japanese, 322-323
Makua Cave and, 322
Maya, 322
Sibyl's Cave and, 321—322
Underworld in, 321
Vjetrenica and, 862
N
NAPLs. See Nonaqueous phase liquids
Natural selection. See also Species
interactions
Asellus aquations and, 545
cave regime, 546—547
in caves, 544—546
colonization and speciation role, 547—548
conclusions and prospects, 548—549
Darwin and, 543—544
Gammarus tninus and, 544—545
historical development, 543—544
modes of, 545, 545f
NMT and, 544
overview about, 543
phenotypic dines and, 545—546, 546f
population structure and, 615—617
Racovitza and, 543—544
time and, 545
Natural traps, 591
Neapliaenops tellkampfi, 744—746
Needles, gypsum, 374-375, 375f
Neff Canyon Cave, 85
Negative allometry hypothesis, 520
Nemastomatidae, 796
Nematoda, 911, 911f
Nemertea, 911, 912f
Neobisiidae, 792
Neogoveidae, 796
Neoteny, 5-6
Nesticidae, 790
Nesticus eremita, 610—611, 610f
Network caves, 395
Network maze, 602f, 761, 762
Neuromasts, 553
Neutral mutation theory (NMT), 544, 554
Neutral mutations
Asellus, 554
Astyanax fasciatus
albinism and, 550
dark melanin and, 550
deleterious risk hypothesis and,
552-553
energy economy hypothesis and, 553
eyes and, 550, 552f, 554
guanine and, 550
neuromasts and, 553
overview, 549—550
pleiotropy hypothesis and, 553
surface and cave forms regarding,
550
modern synthetic theory of evolution, 554
NMT and, 554
reduction and, 549
selection vs., 549
summary, 554—555
Ngamarlanguia luisae, 206f
Nicholson, David, 76
Nicholson, Ike, 74, 75
Niphargidae, 221
Niphargus
anchihaline habitats and, 22, 22f
conclusions about, 562—564
distribution, 555
interactions within and between species
cannibalism, 559
chemoautotrophic bacteria
symbiosis, 560
predation and competition, 559—560,
560t
key elements comparative review,
560-562
communities: niche separation, 562,
563f
female body size, 562
morphological diversity and stasis,
560-561, 561f
key properties, 556-559
abiotic factors tolerance, 556-559
allometric growth, 556, 558f
antennae and pereopods, 556, 557f
body shape, 556, 556f, 557f
body size, 556, 557f
developmental patterns, 556, 558f
endemism/poor migratory abilities,
559
gnathopods, 556, 557f, 558f
light and, 557
maxillae I, 556, 558f
morphological variation, 556, 557f
response to hypoxia, 557-558
sexual dimorphism, 556, 558f
temperatures, 558-559
water chemistry tolerance, 559
N. orcinus, 231f
N. rhenorhodanensis, 653, 654t
N. virei, 613, 653, 654t
overview about, 555-556
taxonomy challenges of, 555
Nitrate contamination, in Karst groundwater
background information, 564—565
environmental concerns and nitrogen
cycling, 565
groundwater resource protection and,
567
hydrogeologic factors, 566
nitrate loss processes, 567
springs and, 566-567
temporal variations and, 567
Nitrates, 506
Nitrites, 129
Nitrogen cycle
chemoautotrophy and, 132—133
environmental concerns and, 565
NMT. See Neutral mutation theory
Nocticolidae cockroach, 241, 242f
Nodules, 779, 780f
Nonadaptive hypothesis, 341—342
Nonaqueous phase liquids (NAPLs). See also
Heavy nonaqueous phase liquids; Light
nonaqueous phase liquids
characteristics and sources, 166—168, 167f,
167t
conclusions about, 172
density and, 166, 167f, 167t
detection, 171—172
in soils, 172
in springs and caves, 171
in wells, 171-172
distribution of, 168f
overview about, 166
solubility and, 166, 167f, 167t
storage in aquifers, 171
fractures, 171
matrix and vugs, 171
pools in conduits, 171
sediment, 171
transport into aquifers, 168—171
climate and, 169,170
epikarst and, 169—170
fractures and, 170—171
nonwetting fluid and, 169
open drains and, 170—171
sinkhole formation and, 170
sinking streams and, 170—171
soil and regolith pores and, 169
volume and rate of release and,
168-169
types of, 166,168
vapor pressure and, 166—167,167f, 167t
viscosity and, 167—168
Northern salamanders, 670
Notch, 425
Notostigmata, 793
Novice cavers, 645
Nullarbor Caves, Australia
arid environments and, 568
biology, 575
blowholes of, 570
chambers of, 570, 570f
climate and, 570—571
conclusions about, 575—576
dolines of, 570, 570f
geology, 568-571, 569f
history, 571—572
early, 571-572
exploration, 572
icicles in, 575, 575f
lakes of, 570, 571f
minerals, 574—575
calcite, 574
gypsum, 574, 574f
halite, 574, 574f
Old Homestead Cave, 574
origin and enlargement of, 572— 574, 573f
overview, 568
physical setting, 568-571, 569f
water flow and, 571
o
Occasional visitors, 275
Ochyroceratidae, 790
Offshore springs, 799-800
936
INDEX
Oklahoma salamander, 671
Old Homestead Cave, 574
Oliarus polyphemus, 14,141
Oligochaeta, 254-255, 912-913, 913f
Oligotrophic terrestrial, 241 —242
Oniscidea, 178t, 189t, 190, 213
Onychophora, 914, 914f
Onychostoma macrolepis, 107, 109t
Oonopidae, 790
Opal, 5011, 505
Open karst, 365, 366—367
Ophisternon candidum, 206f
Opiliones, 215, 256, 795- 797, 796f
Oppiidae, 795
Optymistychna Cave, 829t, 830, 831 f
Ordinskaya Cave, 372, 372f
Organ Cave, 101
Organic acids, 758
Organic minerals, 507
Organicism, 520
Orthogenesis, 520
Orthogonal vector plot, 586-587, 588f
Orthoptera, 217, 259
Ortobalagan Valley, 443, 445f
combined cave profile, 446f
major caves in, 445f
Osteichthyes, 260
Ostracoda, 178t, 182-183,182f
anchihaline habitats and, 21
in Australia, 211—212
Overflow springs, 800
Ox Bel Ha, 837
Oxides, 5011, 505
Oxygen, responses to low
anoxia survival times, 653, 654t
behavioral responses, 654
Crustacea, 653
groundwater and
DO small-scale heterogeneity in,
652-653, 652f
DO supply and consumption in,
651-652
survival evidence in, 653
metabolic responses, 655-656
during anoxia, 655—656
during postanoxic recovery, 656
overview, 651
respiratory responses, 655
Ozark Plateaus, 734f, 737- 738, 738f
Ozerna Cave, 829t
P
Pack rats, 3311, 332
Packs, cave, 306-307
Pairwise species interactions, 743, 744f
Paleoclimate records, from speleothems
dating techniques, 577—578
global keeping of, 577, 579f
overview, 577
paleoclimatology contributions, 578—585
climatic changes timing, 584
Devils Hole, 579-580, 580f
early studies: paleotemperature,
578-579
last glacial, 580-581, 580f
late Holocene, 582, 584f
other studies, 584—585
paleotempestology, 584
solar variability and climate change,
581, 583f
tracking insolation, 581, 582f
tracking paleomonsoons, 581, 582f
vegetation and soil dynamics,
583-584, 584f
stalagmites use for, 577, 578f
Paleofill, 414-415
Paleokarst, 764
Paleomagnetic records
clastic cave sediments and, 586—588, 587f
CRM and, 588
DMR and, 586
earth magnetism and, 586
interpretation, 587—588
local/global column results and, 588, 589f
magnetostratigraphic time scale and, 586,
587f
orthogonal vector plot and, 586—587, 588f
overview, 585-586
speleothems and, 588
value of, 588-589
Paleomonsoons, 581, 582f
Paleontology of caves
deposit types and examples
alluvial and colluvial transport, 591
bats, 592
bears, 593-594
cats, 595, 595f
hyena, 594
Jefferson's ground sloth, 596, 597f
large carnivorans, 592—595
large herbivores, 596
natural traps, 591
other vertebrate sites, 596—597
peccary, 596-597
raptors, 591-592
rodents, 595-596
small carnivores, 596
soft tissue preservation, 596, 596f
wolves, 594-595
dung and, 590, 590f
fossil vertebrates importance, 597—598
Lechuguilla Cave and, 460
overview about, 590—591
PPCS and, 623
Quaternary and, 590
research focus, 590-591
screenwashing and, 597
Vjetrenica and, 862
Wakulla Spring underwater cave system
and, 875-876, 876f
Paleotempestology, 584
Palpigradi, 214, 791
Panbiogeography, 850
Pans, 427-428, 428f
Parciegopis oberwimmeri, 512—513
Paragenesis, 422, 607
Paragenetic ceiling tube, 423, 424f
Paragenetic pendants, 422, 423f
Parasitidae, 793
Paroster species, 613—614
Particulate organic matter (POC), 289
Partings, 871
PASCALIS (Protocol for the Assessment and
Conservation of Aquatic Life in the
Subsurface), 220
Passage growth and development
cave degeneration, 603
cave patterns and, 599—603
anastomotic maze, 602f
branching systems origin, 601
changing conditions, 602—603
curvilinear branchwork, 602f
initial flow routes competition,
599-600, 600f
maze development, 601—602, 602f
network maze, 602f
ramiform, 602f
rectilinear branchwork, 602f
spongework maze, 602f
summary, 602f
vertical organization, 600—601, 601f
conclusion, 603
overview, 598
stages, 598-599
earliest, 599
Passages
angular passages, 604-605, 604f
below water table and, 605—606, 606f
braided, 440-441, 441f
canyon-like, 439-440, 440f, 734, 735f
conclusions about, 608
curvilinear, 604—605, 604f
definitions and concepts, 603—604
discharge, 604
fissure networks, 604f, 605
hypogenic, 604f, 606
incised, 606f
Jewel Cave, 413
Lechuguilla Cave, 458, 459f, 461, 462f
long inclined, 729— 732
Cassowary System, 731—732, 732f
Jean-Bernard Cave, 731, 731f
looping, 724, 724f
Mammoth Cave System, 471—472, 471f,
472f
morphology and evolution of, 606—608,
606f, 607f
overview, 603
patterns of, 604—606, 604f
phreatic, 72, 604, 605, 605f, 606f
stairstep development, 604, 605f
steeply-inclined, 761, 762
tapoff, 607-608, 607f
tributary, 604
types of, 604
Ukraine giant gypsum caves, 831
vadose, 604, 604f, 605, 606f
virgin, 277
Passive detectors
water tracing and
activated carbon, 890
cotton, 891
plankton netting, 891
Quinlan Gumdrop, 891, 891f
Passive dispersal models
of invasion, 406—407
erosion/surface water flow model,
406-407
INDEX
937
rafting model, 406
regression model, 406, 407f
surface water infiltration model, 406
Patchy food scarcity, 1
Pauropods, 538
Pavan, T., 276-277
Pavements, 434
Peccary, 596-597
Pendants, 422, 423f
Pennyroyal Plateau, 736
Peracarida
Amphipoda, 178t, 186, 187t
Bochusacea, 178t, 186
Cumacea, 190,191t
Isopoda, 178t, 187-188,189t
Anthuridea, 178t, 188, 189t
Asellota, 178t, 188-190, 189t
Calabozoidea, 178t, 189t, 190
Flabellifera, 178t, 188, 189t
Microcerberidea, 178t, 188, 189t
Oniscidea, 178t, 189t, 190, 213
Phreatoicidea, 178t, 188,189t
Mictacea, 178t, 185-186, 186f
Mysida, 178t, 185, 186t
Spelaeogriphacea, 178t, 184-185, 185f
Tanaidacea, 190, 190t
Thermosbaenacea, 178t, 185,185t
Peracarida, 178t, 184—191
Periodic food supply, 1
Periodic starvation, 8—9
back to network and, 9
energy demand and, 9
fat accumulation and, 8
overview, 8
Pettalidae, 796
Phalangodidae, 796— 797
Pholcidae, 790— 791
Pholeuon, 60, 61f
Phosphates, 504-505
brushite, 501t, 504—505
development of, 504, 504f
hydroxylapatite, 501t, 504—505
overview, 501t, 504
taranakite, 501t, 504—505
Photokarren, 419—420, 420f
Phreatic looping, Castleguard Cave, 90, 92f
Phreatic maze caves, 395—396
Phreatic passages, 72, 604, 605, 605f, 606f
Phreatic water, 157
Phreatoicidea, 178t, 188, 189t
Phreatoicoides gracilis, 206f
Phrynichidae, 789
Phrynidae, 789
Phyllocarida, 178t, 183, 183f
Phylogeography, 610
Pinnacle, 428—429
karst, 430
Pinnacles of Mulu, 426, 426f
Pisces, 245t
Pisidium, 513—514
Pit craters, 865
Pits, 427-428
Pivka intermittent lakes, 899—901, 901f, 902f
Pivka River, 618, 618f
Planinsko polje, 899, 900f
Plankton netting passive detector, 891
Planthoppers, 216, 663
cixiid, 13—14, 13f
Platycopioida, 178t, 181
Pleiotropy hypothesis, 553
Plethodontidae, 667—674
Eurycea, 667—668
Gyrinophilus, 672—674, 673f
Haideotriton wallacei, 671—672
Plunge pools, 441—442, 442f
POC. See Particulate organic matter
Poecilia mexicana
aggressive behavior, 67-68
reproductive behavior, 64, 65f, 67—68
Point patterns, 475
Point-recharge doline, 142—144, 143f, 145
Poljes, 140,153-155
Cerknica, 433, 434f
Cerknisko, 898-899, 898f, 899f
classification, 154
defined, 153, 435
Dinaric karst and, 153, 199—200
lakes and, 153,153f
lithological contact, 154,154f
Planinsko, 899, 900f
Pollution, 237-238, 327, 403
Polychaeta, 911-912, 912f
Polydesmida, 540, 540t
Polygenetic sink, 152—153
Pool spar, 783, 783f
Helictite Cave, 81f
Population flush, 10—11
Population structure
in aquatic troglobites
Gammarus minus, 612
historical determinants and
secondary contact, 614
hydrological routes vs. adaptive
diversity, 612—614
Niphargus virei, 613
Paroster species, 613—614
Troglocaris anophthalmus, 613
cave crickets heterozygosity and,
615-617, 616t
dispersal routes
surface vs. underground, 612
food resource heterogeneity and, 615
genetic structure estimation, 609—610
effective population size, 609
F-statistics, 609
genetic distance, 609
heterozygosity, 609
migration, 609
phylogeography, 610
genetic structure estimation and, 609—610
genetic variability and natural selection,
615-617
habitat heterogeneity and, 614—615
overview, 608—609
at regional scale, 610—612
gene flow study cases, 610—611, 610f
vicariant species, 610
troglobitic heterozygosity comparisons,
615-617, 616t
Porous media aquifers, 383—384
Postojna-Planina Cave System (PPCS)
Asellus aquations and, 32—33
as biotically diverse, 236
description, 623
entrances, 623
exploration history, 622
exstensometers used in, 620—621
Lepe Jame in, 622, 622f
physical setting, 618—622
faults, 618-619, 619f
geologic, 618-621, 619f, 620f
hydrology and topography, 618
lithological columns, 619—620, 620f
other research, 621
Pivka River, 618, 618f
special attributes, 623—624
biology, 623
mineralogy, 623
paleontology/archaeology, 623
tourist railway, 622, 622f
Postojnska Jama, Slovenia, 452f
Potassium nitrate, 676. See also Saltpetre
mining
Potholes
dissolutional, 420, 422f
lateral, 682f, 683
Poulson, Thomas, 465
PPCS. See Postojna-Planina Cave System
Preadaptation, 10, 295—296, 524
Predation, 743, 744f
Pressure ridge caves, 872
Proboscis worms, 911, 912f
Prodidomidae, 791
Progressive adaptations, 517
Project caving, 308
Project SIMMER, 86
Prokoeneniidae, 791
Prosobranchs, 514—515
Prostigmata, 794
Protection. See Conservation
Proteidae, 665—667
Proteus anguinus, 666—667, 666f
Proteobacteria, 495—496
Proterorhagiidae, 794
Proteus anguinus, 219, 230, 653, 666—667, 666f
aggressive behavior, 67-68, 67f
in Dinaric karst, 234
feeding behavior, 63, 64f
reproductive behavior, 64—65
Protocobitis, 107
Protocol for the Assessment and
Conservation of Aquatic Life in the
Subsurface. See PASCALIS
Protoschizomidae, 789
Pselaphidae, 232
Pseudochiridiidae, 792
Pseudogarypidae, 792
Pseudokarst caves, 155
Pseudoscorpiones, 791—792, 791f
in Australia, 215
in Dinaric karst, 232
guano and, 362
in United States, 256—257
Pseudosinella, 518, 518f
Pseudosinella hirsuta, 526, 526f
Pseudotremia, 541, 541f
Puka, 868-869
Pulmonates, 514-515
938
INDEX
Pygolabis humphreysi, 206f
Pyroducts, 866-870, 866t
canyon formation, 868—869, 870f
double-trunked, 869—870
formation of, 867, 868f
Huehue Cave and, 869f
Huehue Flow and, 869—870, 871f
inflation and, 867, 868f
internal development, 867—869, 870f
puka and, 868—869
single-trunked, 869
superimposed-trunked, 870
types of, 869-870
Q
Qafzeh Cave, 95f, 98f
Quartz, 501t, 505
Quartzite caves of South America
deposits and ages, 639
genesis and development controls,
638-639
Inficionado Peak, 637, 637f
lithology and occurrence, 635—637
morphology and, 637-638
anastomotic caves, 637, 638f
branchwork caves, 637, 638f
vertical network caves, 637, 638f
Muchimuk-Colibri system, 635—636, 636f
overview, 635
Roraima Sur, 635—636, 636f
Quaternary, 590
Queen Formation, 457-458, 457f
Quinlan Gumdrop, 891, 891f
Quintana Roo State, 834
R
Raccoons, 331t, 332
Racovitza, E., 219-220, 275-276, 277,
404-405, 543-544
Radio detection and ranging (RADAR), 350
Radiolocation, 351—352, 351f
Rafting model, 406
Rain pits, 140
Ramiform pattern, 602f
Rappel, 307
Raptors, 591-592
Ravelling, 720
Rebelays, 307
Rebreathers, 312, 313f
CCRs, 313
SCRs, 313
Records and, 270
Recreational caving
activities not considered
abandoned buildings, 643
overview, 642—644
shelter caves, 643
show caves, 643
cave types used for, 641—642
glacier caves, 642
lava caves, 641
mines and tunnels, 642
sea caves, 642
solution caves, 641
talus and tectonic caves, 642
contemporary trends, 647
definition, 641
demographic factors, 646—647
age, 646
frequency cycles, 647
length of trips, 647
race, 647
sex, 646—647
discouraged items for, 645
downside, 644—645
factors contributing to, 644
appeal, 644
distance to cave, 643—644
group caving suitability, 644
known existence, 643
legal access, 644
physical access, 643
levels for everyone, 646
novice cavers and, 645
other voids and, 642
people avoiding, 645—646
people continuing, 646
reasons people begin, 644
challenge/exercise, 644
exploration/adventure, 644
social aspect, 644
unique accomplishment, 644
spelunker regarding, 645
summary, 647—648
urban, 642
Rectilinear branchwork, 602f
Redox, 162-163
Regional cavernicoles, 845
Regression model, 406, 407f
Regressive adaptations, 517
Regressive evolution hypotheses, 519—521
Regular cavernicoles, 845—846
Remipedia, 21,178t, 179-180,180f, 180t
Reports, 270
Reptilia, 260—261
Rescues
background about, 648
Collins and, 648
history of, 648—649
early, 649
HURT and, 649
learning from past, 650
accident frequency, 1950s, 650
accident frequency, 1960s, 650
accident frequency, 1970s, 650
accident frequency, 1980s, 650
overview, 648
present and, 650—651
team development, 649
Reseau Siebenhengste-Hohgant system,
700-702, 726-729, 728f. See also
Siebenhengste cave system
Restoration, 632
Rhagidiidae, 794
Ribbon worms, 911, 912f
Ricinoididae, 793
Ricinulei, 793, 793f
The Rift, 463
Rigging, vertical exploration, 315—316, 315f
Rillenkarren, 428, 428f
Rills, 420, 421f, 425
Rimstone dams, 779, 780f
Robins Rift, 78
Rock shelters, 105f, 106
Rocky Mountains erosion rates, 531
Rodents, 595-596
Roofless caves, 201—202, 202f
Roosts functions, bat, 46—48
Root communities, in lava tubes
cave zones, 661-663
bad air zone, 662—663
deep zone, 662
entrance zones, 662
transition/mixing zone, 662
tree roots and, 661—662
twilight zone, 662
conservation of, 663
environmental factors, 658—661
air flow, 659
energy sources, 658-659
geomorphology influence, 661
moisture/humidity, 659
oxygen and carbon dioxide,
659-661
food web, 660f
in Hawaii and Australia, 661t
overview, 658
species morphology and adaptation, 663
crickets and cockroaches, 663
moths, 663
planthoppers, 663
Roraima Sur, 635—636, 636f
Rouch, Raymond, 99
Roundworms, 911, 91 If
Rousette fruit bats, 46
Ruiniform karrenfield, 430
Runnels, 428, 429f
S
Sabaconidae, 796
Sac Actun, 837-838
Safety, 309
Salamanders, 260, 261 1 . See also specific
salamander
blepsimolge, 668f, 669-670
conclusions about, 675
diversity patterns of, 674—675
overview, 665
Plethodontidae, 667—674
Eurycea, 667—668
Gyrinophilus, 672—674, 673f
Haideotriton wallacei, 671—672
Proteidae, 665—667
Proteus anguinus, 666—667, 666f
septentriomolge, 670
typhlomolge, 668-669, 668f
Saltpetre mining, 676—678
bacteria and, 676
definitions and importance, 676
efflorescent crusts and, 676
evidence of, 677
principle, 677—678
secondary, 678
gunpowder and, 676
history, 677
INDEX
939
other sediment extraction contrasted with,
679
potassium nitrate and, 676
rockshelters and, 676
saltpetre processing, 678—679
evidence, 678
outside, 678
steps, 678
summary, 679
white lint-like fibers and, 676, 677f
San Gabriel Springs salamander, 670
San Marcos salamander, 670
San Salvador Island karren, 426f
Santa Ninfa Plateau dolines, 145
Sarawak Chamber, 534
Scallops, 420-422, 422f
defined, 679
discharge calculation, 681
discharge data uses and pitfalls, 681—682,
681f, 682f
groundwater flow
direction and, 680, 680f
velocities and, 680-681, 680f
other phenomena, 682—683, 682f
pockets, 682-683
potholes, 682f, 683
overview about, 679
scour marks and, 422
transverse flutes and, 422
Scarisoara Ice Cave, 403
Scheloribatidae, 795
Schiner, J. R., 275, 276, 277
Schiodte, J. C., 275
Schizomida, 215, 257, 788- 789
Schmidl, Adolf, 103
Schoharie Valley, 741, 741f
Schwartz, Ben, 76
Scolopendromorpha, 539
Scorpiones, 257. See also specific Family
in Australia, 215
Families, 787-788, 788f
Scoterpes, 541
Scour marks, 422
SCR. See Semi-closed circuit rebreather
Screenwashing, 597
Scutariellidae, 234
Scutigeromorpha, 539
Sea caves, 105f, 106
as coastal caves, 155-156
human interaction with, 156
last interglacial and, 156
modern, 156
overview about, 155—156
recreational caving and, 642
ubiquity of, 156
variations of, 156, 156f
Sea level, 155
flank margin caves and, 159
Security systems, 631
Sediments, Castleguard Cave, 93
Seepage spring, 684-685
characteristics, 684, 684f
chemical and physical conditions,
684-685, 685f
species found in, 685, 685t
temperature and, 684—685, 685f, 685t
Seeps, 800-801
Seismic modeling techniques, 350
Semi-closed circuit rebreather (SCR), 313
Septentriomolge, 670
Serpulidae, 234
Sessile chaetognaths, 20
Seven Rivers Formation, 457—458, 457f
Shaft and canyon caves, 729
Krubera Cave, 729, 730f
Shallow subterranean habitats (SSHs)
epikarst, 685t, 686, 686f
fauna origin, 689
hyporheic zone, 686-688, 687f, 688t
MSS, 686f, 688, 688f
other subterranean habitats and, 689
overview, 683—684
seepage spring, 684-685
Shelta Cave, 625
Shelter caves, 643
Shields, 781, 782f
Shield's Diagram, 138f, 139
Shifflett, Tom, 76-77
Short-tailed whipscorpions, 788— 789
Show caves, 643
budget estimate of, 692t
Cango caves, 695f
development guidelines, 693—697
aboveground facilities, 696—697
access and pathways, 693—694
guide training, 697
lighting, 694—695
managers, 697
monitoring, 697
visitor frequency and numbers,
695-696
of wild cave into show cave, 693
development of, 691—693
environmental protection of, 690—691
Frasassi caves, 696f
history, 690
Hwanseon Cave, 696f
Kartchner Caverns, 694f
obtaining information about, 697
Sibyl's Cave, 321-322
Side-mount diving, 311, 312f, 313f
Siebenhengste cave system
detailed map of, 701 f
fauna discovered in, 706—707
genesis and age, 703—704, 704f, 705f
geographical and geological setting,
698-699, 698f, 699f
hydrology, 698f, 699
minerals, 704— 706, 706f
overview and exploration, 699— 703, 699f,
700f
A2, 703
Barenschacht, 699—700
FI, 702
Faustloch, 702
K2, 702- 703
Reseau Siebenhengste-Hohgant,
700-702
St. Beatus Cave, 699
Sierra Nevada incision rates, 531
Sighted salamanders, 668f, 669—670
Silicates, 50It, 505
allophane, 5011, 505
opal, 501t, 505
origins, 505
other, 505
quartz, 5011, 505
Similarity rule, 775- 776, 775f
SIMMER, Project, 86
Simmons, Ron, 76
Simmons-Mingo Cave, 86
Sinella, 518, 518f
Single fracture, 510, 510f
Single or rudimentary networks, 761—762
Single-conduit caves, 394
Single-rope techniques, 307
Single-trunked lava tunnels, 869
Sinkholes, 140
building damage by, 720— 721, 720f,
721f
causes of, 718, 719f
collapse, 150-151,150f, 151f
cover, 150f, 152
defined, 434-435, 723
Dinaric karst, 199—200
entrances, 281, 282f
evolution of, 723
examples of, 718
formation, 718, 719f
NAPLs and, 170
intersection, 150f, 152
Mitchell Plain and, 722
other types of, 150—152,150f
overview about, 718
ponding or damming and, 721
reservoirs and, 721
soil piping and, 720
solution, 718, 719f
conclusions about, 150
defined, 140
drainage and, 140
drawdown, 143f, 144
evolution examples, 144f, 145—147
Eyre Peninsula, 147
filling material evacuation and, 145
frost shattering and, 145
inception, 143f, 144
as karst surface diagnostic forms,
140-150
locations, 141
midlatitude, 149
Monte Baldo, 147
Montello, 145-147,146f, 148f
Monti Lessini, 144f, 145
morphology and size, 140—141
other evolution processes, 145—147
overview about, 140
point-recharge, 142—144, 143f, 145
Santa Ninfa Plateau, 145
slope processes and, 145
soil and, 145
specific environmental conditions
and, 147-149
structure, 141—142, 142f
subsoil, 141
940
INDEX
Sinkholes ( Continued )
tropical landscape and, 147—149,
149f
Velebit Mountains, 147
Waitomo district, 145
subsidence, 150f, 151—152
suffosion dolines and, 722
types of, 718, 719f
buried sinkhole, 719f, 722
cover-collapse sinkhole, 719f
cover-subsidence sinkhole, 719f,
721-722, 721f
subsidence sinkhole, 718—719, 719f
Sinking streams, 707—709
cave evolution and, 707, 708f
Danube River and, 707
Dinaric karst and, 237
Edwards Aquifer and, 708— 709
entrances and, 281
evolutionary sequence of, 710—712
case A, 711, 711f
case B, 711, 71 If
case C, 711, 711f
Florida, 708
function of, 709—710
Kentucky, 709, 709f, 71 Of
NAPLs and, 170-171
overview, 707
Sinocyclocheilus, 108
adipose storage, 117,118f, 119f
biogeography, 121—124
eye degeneration, 117—120,119f, 121f
humpback and horn of, 114—116, 114f,
115f
phylogenesis, 121—124, 123f
pigmentation loss, 108f, 120
reproduction, 108—112
research and conservation of, 124—125
S. hyalinus, 107, 108f, 109t
scale disappearance, 120—121
sensory apparatus, 113f, 116—117, 116f
specialized appendages, 108f, 117
speciation mechanisms, 121—124
Sironidae, 796
Sistema del Faro flank margin cave, 159f
Sistema Huautla, Mexico
Anthodite Hall, 715f
archaeology, 717—718
biology, 717
cave description, 712—716
exploration, 714f
geology, 716
hydrogeology, 713f, 714f, 717
location, 713f, 716
overview, 712
speleothems, 715f, 716f
stream network, 713f
Sistema K'oox baal, 836f
Sistema Ox Bel Ha, 837
Sistema Sac Actun, 837—838
Sit harness, 316—317, 316f
Sket, B., 276
Skiles, Wes, 312
Skocjanske, 433, 434f
Slab breakdown, 69
Slavka Cave, 829t
Slickenslides Room, Helictite Cave, 82f
Sloth, Jefferson's ground, 596, 597f
Slovenia
dolines in, 435f
Postojnska Jama, 452f
Snakes, 218
Soda straw, 807
Soil piping, 720
Solution caves, 105f, 106—107
alpinotype, 724
Chester and Crawford Uplands, 736
in high relief regions
Cassowary System, 731—732, 732f
cave examples, 726
characteristics of, 724
climate and, 724
conclusions about, 732—733
glacier milk and, 724
Holloch Cave, 726, 727f
Jean-Bernard Cave, 731, 731f
Krubera Cave, 729, 730f
large labyrinths, 726—729
long inclined passages, 729—732
looping passages, 724, 724f
overview about, 723— 724
relative chronology and, 725— 726
relative chronology example,
725- 726, 725f, 726t
Reseau Siebenhengste-Hohgant
system, 726- 729, 728f
shaft and canyon caves, 729
St. Beatus Cave, 725- 726, 725f
study of, 725
water temperature and, 724
in moderate relief regions
Appalachian Mountains, 734f,
739- 740, 740f, 741f
Appalachian Plateaus, 734f, 738— 739,
739f
canyon-like passages, 734, 735f
characteristics of, 733
dissected plateaus and, 736—737
drainage history and, 737
field examples, 735— 742
geographic distribution of, 734, 734f
glaciated plateaus, 740— 742, 741f
Illinois structural basin and, 734f, 736
Mammoth Cave and, 737, 737f
northeastern/northcentral U.S.,
740- 742, 741f
origins of, 734-735
overview about, 733
Ozark Plateaus, 734f, 737- 738, 738f
Pennyroyal and Mitchell Plateau, 736
sediment accumulation and, 737
southern Indiana and, 736—737
summary, 742—743
tubular conduits, 734— 735, 735f
western Kentucky and, 736—737
recreational caving and, 641
Solution dolines
conclusions about, 150
defined, 140
drainage and, 140
drawdown, 143f, 144
evolution examples, 144f, 145—147
Eyre Peninsula, 147
filling material evacuation and, 145
frost shattering and, 145
inception, 143f, 144
as karst surface diagnostic forms, 140—150
locations, 141
midlatitude, 149
Monte Baldo, 147
Montello, 145-147, 146f, 148f
Monti Lessini, 144f, 145
morphology and size, 140—141
other evolution processes, 145—147
overview about, 140
point-recharge, 142—144, 143f, 145
Santa Ninfa Plateau, 145
slope processes and, 145
soil and, 145
specific environmental conditions and,
147-149
structure, 141—142, 142f
subsoil, 141
tropical landscape and, 147—149, 149f
Velebit Mountains, 147
Waitomo district, 145
Solution pockets, 423, 424f
Solution runnels, 428, 429f
Solution sinkhole, 718, 719f
Solutions pans, 140
South America, quartzite caves of
deposits and ages, 639
genesis and development controls,
638-639
Inficionado Peak, 637, 637f
lithology and occurrence, 635—637
morphology and, 637-638
anastomotic caves, 637, 638f
branchwork caves, 637, 638f
vertical network caves, 637, 638f
Muchimuk-Colibri system, 635—636, 636f
overview, 635
Roraima Sur, 635—636, 636f
Sparassidae spiders, 242, 243f
Spatial pattern investigation, 476—477
Spatial process decomposition, 475—476
Species interactions
Appalachian stream competition
amphipod and isopod, 746—747,
746f, 747f
beetle predators
cricket egg prey of, 744— 746
conclusion, 748
eutrophication and competition and,
747- 748
overview about, 743— 744
pairwise, 743, 744f
Spelaeogriphacea, 178t, 184-185,185f, 213
Speleogenesis, hypogenic, 756-764
artesian, 759
basic concept, 750
chemical mechanics, 756— 759, 756f, 757f
carbonic acid dissolution, 756— 758
evaporites dissolution, 758
hydrogen sulfide, 758
INDEX
941
mixed carbonate/sulfate strata
dissolution, 759
organic acids, 758
sulfuric acid, 758
thermal water cooling, 758
water mixing, 758
combined artesian/deep-rooted, 759
deep-rooted, 759
defined, 750
diagenetic environments, 750—756
eogenesis and, 751
mesogenesis and, 751
telogenesis and, 751
distribution and settings, 759
endokarst realm, 754-756, 755f, 756f, 757f
groundwater systems and regimes,
751-752
hydrogeologic and structural controls,
759-761, 760f
hydrogeologic cycle, 752-753, 753f
karst settings evolution, 753—754, 754f
mesomorphology features, 763—764
overview, 748-750, 749f
paleokarst and, 764
porosity patterns, 761—763
cavernous edging, 761, 762
collapse shafts, 761, 762
isolated chambers, 761, 762
network maze, 761, 762
single or rudimentary networks,
761-762
spongework maze, 761, 762
steeply-inclined passages or shafts,
761, 762
Speleogenesis, telogenetic
basic dynamics, 765—767, 766f
branchwork pattern upward evolution,
768- 769, 768f
breakthrough and, 766
conduit networks evolution and, 767—769,
768f
fracture stages and, 766
overview, 765
Speleothems, 103—104. See also Gypsum
flowers; Paleoclimate records, from
speleothems; Stalactites; Stalagmites
anthodites, 379-380, 380f, 782
blue holes and, 160
Castleguard Cave, 93
cave shields, 381
classification of, 778— 783, 779t
deposition mechanisms, 778
dripping water created, 779—781
carrot-shaped stalactite, 779— 781,
781f
drapery, 781
shields, 781, 782f
stalactite-stalagmite combination,
781, 781f
stalagmites, 781, 781f
straw-shaped stalactite, 779—781,
780f
flowing water created, 779, 779t
cave beads, 779, 780f
flowstone, 779f
nodules, 779, 780f
rimstone dams, 779, 780f
frostwork, 380, 381 f
growth dynamics
uranium series dating and, 842—844,
843f
gypsum, 374-375
heavy metals and, 164
helictites
beaded, 379, 380f
bushes, 382-383, 382f
subaerial, 379
subaqueous, 381—383, 381f, 382f
ice, 401f
Jewel Cave, 416—417
hydromagnesite balloons, 416—417,
416f
popcorn stalagmites, 416
Kazumura cave, 442
primary, 442
secondary, 442
mineral precipitation and, 885—886
minerals and, 777—778
agagonite, 777
calcite, 777
gypsum, 777- 778, 778f
overview, 777
as paleoclimate archives, 785
paleomagnetic records and, 588
properties, 783— 784
calcite/aragonite problem, 783—784,
783f
color and luminescence, 784, 784f
seeping water created, 782
frostwork, 782
gypsum flower, 782, 782f
helictites, 782, 782f
shapes of, 778-783
Sistema Huautla, 715f, 716f
standing water created, 782—783
calcite rafts, 783
cave pearls, 782- 783, 783f
pool spar, 783, 783f
in volcanic caves, 784—785, 872
Speleothem deposition
chemistry
calcite precipitation / dissolution,
769- 770, 770f
overview, 769
regular stalagmite
morphology, 772- 775, 7731, 774f,
774f
similarity rule, 775—776, 775f
stalagmite growth
under variable time conditions,
776-777, 776f
thin water layers and, 770—771, 771 f
turbulent water flow and, 771—772, 772f,
773f
Speleothems, paleoclimate records from
dating techniques, 577—578
global keeping of, 577, 579f
overview, 577
paleoclimatology contributions, 578—585
climatic changes timing, 584
Devils Hole, 579-580, 580f
early studies: paleotemperature,
578-579
last glacial, 580-581, 580f
late Holocene, 582, 584f
other studies, 584—585
paleotempestology, 584
solar variability and climate change,
581, 583f
tracking insolation, 581, 582f
tracking paleomonsoons, 581, 582f
vegetation and soil dynamics,
583-584, 584f
stalagmites use for, 577, 578f
Speleothems, uranium series dating of
ancient speleothems and, 844
disequilibrium dating, 839—840, 839f
overview, 838-839
practical issues, 841 —842
cost, 842
in-cave sampling, 841, 841f
permits, 841
sample material quantity needed,
841, 842f
sample quality, 842
stalagmite focus, 842
precision and, 840—841
principles of, 839-840
reporting and assessing, 840—841
speleothem growth dynamics and,
842-844, 843f
technology, 840
uncertainties and, 840
Spelunker, 645
Sphaerochthoniidae, 795
Spiders. See also specific Family, Order or Sub
Acari, 793-795
Amblypygi, 789, 789f
of Arachnida, 256
Araneae, 789—791
biology, 786—787
conservation, 787
distribution, 786
Opiliones, 795-797, 796f
overview, 786
Palpigradi, 791
Pseudoscorpiones, 791—792, 791f
Ricinulei, 793, 793f
Schizomida, 788— 789
Scorpiones, 787-788, 788f
Sparassidae, 242, 243f
systematics, 787—797
wolf, 15
Spitzkarren, 429
Splitkarren, 428—429
Sponges, 20
Spongework, 422
caves, 396
maze, 602f
patterns, 761, 762
Spongillidae, 234
Spring salamanders, 672— 674, 673f
Springs. See also Seepage spring; Wakulla
Spring underwater cave system, Florida
Barton Springs salamander, 669—670
942
INDEX
Springs (Continued)
discharges, 801—804
ebb-and-flow springs, 801—802
hydrographs analysis, 802
magnitudes of, 801, 801t, 802f, 802t
saturation state, 804
turbidity and chemical variability,
803-804, 803f, 804f
entrances at, 281, 281f
Eurycea and, 667—668
failure, 856—857
Gammarus minus and, 344
as habitat, 804—805
heavy metals and, 165f
NAPLs and, 171
nitrate contamination and, 566—567
overview about, 797
salamanders, 669— 670, 672— 674, 673f
San Gabriel Springs salamander, 670
types of, 798-801, 798f
alluviated springs, 799, 799f
artesian springs, 799—800
conduit and diffuse flow springs, 798
gravity springs, 798- 799, 799f
mineral springs, 800
offshore springs, 799—800
overflow, underflow, and
distributary springs, 800
Seeps and wetlands, 800—801
thermal springs, 800
as water supplies, 804
West Virginia spring salamander, 673f,
674
SSHs. See Shallow subterranean habitats
St. Beatus Cave, 699, 725-726, 725f
Stable isotope ratio analysis, 493—494
Stagnant-air zone, 205
Stalactites
carrot-shaped, 779—781, 781f
column and, 805, 806f
conical, 807, 807f
defined, 805
epsomite, 375, 376f
mineralogy and formation processes,
805-806, 806f
mirabilite, 375, 376f
overview about, 805
soda straw, 807
stalagmite combination, 781, 781f
straw-shaped, 779—781, 780f
Stalagmites. See also Paleoclimate records,
from speleothems
aragonite, 809
as archives of past, 809-810
column and, 805, 806f, 809
creation of, 781, 781f
defined, 805
formation of, 807—809
growth
rate, 808-809
under variable time conditions,
776-777, 776f
ice, 402
Jewel Cave, 416
microcrystalline fabric, 809
mid-Pliocene, 809—810, 810f
mineralogy and formation processes,
805-806, 806f
morphologies, 807-808, 808f
overview about, 805
paleoclimate records from, 577, 578f
popcorn, 416
regular
morphology, 772- 775, 773f, 774f
similarity rule, 775—776, 775f
stalactite combination, 781, 781 f
uranium series dating and, 842
Stalita taenaria, 231f
"A Statistical Theory of Cave Entrance
Evolution" (Curl), 278
Steeply-inclined passages or shafts, 761, 762
Stenasellus virei, 653, 654t
Sternophoridae, 792
Stillstands, 528—529
Stiphidiidae, 791
Stone, Bill, 312
Straw-shaped stalactite, 779— 781, 780f
Stygiocaris stylifera, 214f
Stygnommatidae, 797
Stygnopsidae, 797
Stygobiotic ingolfiellid amphipods, 684f
Stygobites, 245t
Stygobromus, 853—854
Stygotrombidiidae, 795
Stylocellidae, 796
Subaerial helictites, 379
Subaqueous helictites, 381—383, 381f, 382f
Subcutaneous karst, 284
Subglacial caves, 356, 356f
Subjacent karst, 365, 371
Subsidence dolines, 150f, 151—152
Subsidence sinkhole, 718—719, 719f
Subterranean organisms, 276
Subtroglophile, 276—277
Subway, Castleguard Cave, 92f
Suffosion dolines, 722
Suffosional caves, 105f, 106
Sulaplax ensifer, 241, 241f
Sulfate mineral crystals. See also Gypsum
flowers
overview about, 374
sources and depositional mechanisms,
375-377
types of, 374
Sulfates, 500-504, 501t
alunite, 5011, 503
barite and celestine, 501t, 503
deposition mechanisms, 503—504
epsomite, 5011, 503
gypsum, 500-503, 501t
mirabilite, 5011, 503
Sulfides, 507
Sulfur compounds, reduced inorganic,
128-129
Sulfur cycle, 130-132, 132f
Sulfuric acid
hypogenic speleogenesis and, 758
karst, 437
Lechuguilla Cave and, 456
Sulfuric acid caves
chemical reactions, 812
conclusions about, 819
diagnostic features, 811—812
field examples, 812—818
Carlsbad Cavern, 816—817, 816f
Cueva de Villa Luz, 813-815, 813f,
814t, 815f
Frasassi Cave System, 815-816, 815f
Grand Canyon, 817—818, 818f
Guadalupe Mountains, 816—817,
816f
Kane caves, 813, 813f, 814t
Lechuguilla Cave, 816—817, 816f
geologic setting, 810-811, 811f
microbial interactions, 812
MVT ore deposits and, 818—819
overview, 810
uranium and, 818—819
Sullivan Cave, 86
Superficial karst, 285
Superficial underground environment
(MMS), 524-525
Superimposed-trunked lava tunnels, 870
Surface water infiltration model, 406
Swiftlets, 242, 243, 243f
Syarinidae, 792
Symphylans, 538
Syncarida, 178t, 183—184
Anaspidacea, 178t, 184,184f
in Australia, 209—211
Bathynellacea, 178t, 184,184f
Syngenetic karst, 365
Synotaxidae, 791
Synurella ambulans, 231f
T
Tables, 270
Tabun Cave, 95f, 97-98, 97f
Tafoni caves, 155
Tailless whipscorpions, 789, 789f
Talus, entrance, 284
Talus caves, 105f, 106,155
recreational caving and, 642
Tanaidacea, 190, 190t
Tantulocarida, 178t, 180
Tapoff passages, 607-608, 607f
Taranakite, 501t, 504—505
Tardis Cave, 536f
Tecto-karstic hollow, 153
Tectonic caves, 105—106,105f, 865
recreational caving and, 642
Telemidae, 791
Telogenesis, 751
Telogenetic karst, 437
Telogenetic speleogenesis
basic dynamics, 765—767, 766f
branchwork pattern upward evolution,
768- 769, 768f
breakthrough and, 766
conduit networks evolution and, 767—769,
768f
fracture stages and, 766
overview, 765
TEM. See Transient electromagnetics
Temnocephalida, 910—911, 910f, 911f
INDEX
943
Tennessee cave salamander, 673, 673f
Terikan River Caves, 536
Terminal breakdown, 69
Terrestrial cave animals, 4
Terrestrial Isopoda, 14—15
Terrestrial troglobites distribution, 851—853,
852f
beetles, 851-852, 852f
Tetrablemmidae, 791
Tetragnathidae, 791
Texas
cave and spring-dwelling Eurycea,
667-668
salamander, 669
blind, 668
Thalassoid limnostygobionts
marine regressions and
biogeography of, 483, 484f, 485f
Morocco and, 483—484
as sole evidence, 483—484
Tharp model, 71
Theraphosidae, 789
Theridiidae, 791
Theridiosomatidae, 791
Thermal erosion, 440—441, 440f
Thermal springs, 800
Thermal variation, 350—351
Thermal waters
chemistry, 393
cooling, 758
Dinaric karst and, 237
Thermobaric regime, 751—752
Thermoindicator, 402
Thermosbaenacea, 178t, 185t, 185, 257
anchihaline habitats and, 21
in Dinaric karst, 234
Thrust faults, 354, 354f
Tiankeng, 151
critical factors, 822—824
defined, 821
development stages, 822—824, 823f
distribution of, 824
features, 822-824, 824t
major world, 824t
types, 821-822, 821f, 822f
collapsed, 821, 821f, 822f
erosional, 821f, 822, 822f
water, 822
underground river and, 823, 823f
Ticks, 793-795
Tourism, 403-404, 451-452, 622, 622f
Towers, 435-436, 436f
Townsend's big-eared bat, 46f
Trajano, E., 276
Transient electromagnetics (TEM), 352
Transverse flutes, 422
Trap hypothesis, 520
Traverses and tyrolians, 318—319, 319f
Travuniidae, 797
Tree roots, 661—662
Tremarctine bears, 593
Triaenonychidae, 797
Tributary passages, 604
Tricladida, 910, 910f
Tridenchthoniidae, 792
Triple porosity model, 384—385
Troglobites, 107. See also specific troglobite
ecological classification and, 275—276, 277
habitats
lava tubes, 658
limestone caves, 658
mesocaverns, 658
tree roots, 658
single vs. multiple origin of, 298—299, 298f
tropical diversity patterns, 245t
Troglocaris anophthalmus, 613
Troglomorphy, 517—518
Troglophiles, 107, 275-276, 277
Troglosironidae, 796
Troglotayosicidae, 788
Trogloxenes, 107. See also Bats
ecological classification and, 275—277
Trombiculidae, 794
Trombidiidae, 794
Tropical landscape, sinkholes and, 147—149,
149f
Tropical winter effect, 205
Tropics diversity patterns. See Diversity
patterns, in tropics
Trout Rock caves, 739, 740f, 741f
Tubular conduits, 734— 735, 735f
Tumuli, 871-872
Tunnels, recreational caving and, 642
Turbellaria, 254, 910-911
Temnocephalida, 910—911, 910f, 911f
Tricladida, 910, 910f
tropical diversity patterns, 245t
Turbulent water flow, 771-772, 772f, 773f
Twilight zone, 283
Typhlichthys subterraneus, 465, 466t
Typhlochactidae, 788
Typhlogammarus mrazeki, 231f
Typhlomolge, 668-669, 668f
Tyrolians, 318-319, 319f
u
Ugryn Cave, 829t
Ukraine giant gypsum caves
feeding channels and, 831
hypogene transverse speleogenesis,
832-833, 832f
master passages and, 831
as mazes, 830, 830f
Optymistychna Cave, 830, 831f
patterns and morphology, 829—832, 829t
regional geology, 827-829, 828f
regional hydrogeology, 827-829, 828f
sediments and formations, 832
Underflow springs, 800
Underground shallow medium, 284—285
Underwater exploration techniques
background about, 310
carbon dioxide and, 313
equipment development, 310—311
factors contributing to, 310
fatalities and, 313—314
rebreathers, 312, 313f
CCRs, 313
SCRs, 313
side-mount diving, 311, 312f, 313f
survey tools, 312f, 314
Underworld, in myth, 321
United States diversity patterns. See
Diversity patterns, in United States
Uranium, 818—819
Uranium series dating, of speleothems
ancient speleothems and, 844
disequilibrium dating, 839—840, 839f
overview, 838-839
practical issues, 841—842
cost, 842
in-cave sampling, 841, 841f
permits, 841
sample material quantity needed,
841, 842f
sample quality, 842
stalagmite focus, 842
precision and, 840—841
principles of, 839-840
reporting and assessing, 840—841
speleothem growth dynamics and,
842-844, 843f
technology, 840
uncertainties and, 840
Urban caving, 642
Urodacidae, 788
Uropodidae, 793—794
Ursns americanus, 593
Ursus speleus, 593
U-shaped tube effect, 19
Uvalas, 152-153,199-200
V
Vadose passages, 604, 604f, 605, 606f
Vadose water, 157
Vaejovidae, 788
Valdina Farms salamander, 670
Vampire bats, 53—54
Vanadates, 506
Vandel, A., 404-405, 520
Variscite, 501t, 504-505
Velebit Mountains dolines, 147
Ventilation cooling, 401
Vermiculations, 496, 497f
Verteba Cave, 829t
Vertebrate visitors, birds and mammals
cave ecology impact of, 847
cave-dwelling species, 845—846
facultative visitors, 845
regional cavernicoles, 845
regular cavernicoles, 845—846
den sites and, 846—847
foraging habitats and, 847
knowledge gaps, 846
man and, 848—849
rare species, 848
resting and mating and, 846
wintering sites and, 847
Vertical exploration techniques, 307
ascending, 307, 316f, 317-318
from bottom up, 319—320
changing over, 318
descending, 307, 317
disclaimer, 314
overview regarding, 314—315
944
INDEX
Vertical exploration techniques ( Continued )
rappel, 307
rigging and, 315-316, 315f
rope and, 315
single-rope, 307
sit harness and, 316—317, 316f
traverses and tyrolians, 318—319, 319f
Vertical network caves, 637, 638f
Vicariant phenomena, 855
calcrete aquifers and, 857
spring failure, 856-857
stranding, 855-856
stream capture, 856
Vietnam, Hon Chong hills in, 249—250, 250f
Virgin passages, 277
Vjetrenica
cultural history, 862
artifacts, 862
myths, 862
sacred status, 862
ecological changes to, 864
exploration, 863—864
fauna, 861—862
Golden Chamber of, 859—860, 860f
habitats, 861
hydrology of, 860
kamenica of, 859—860, 860f
karst area of, 858
microclimate, 860—861
overview about, 858
paleontology, 862
protection, 864
speleomorphology of, 858-860, 859f
visitor access, 864
water bodies of, 860, 860f
Volcanic rock caves, 105f, 106
conclusions about, 872—873
limited extent
drained lobes, 871—872
imprints, 871
lava rise, 871—872
partings, 871
pressure ridge, 872
tumuli, 871—872
volcanic vents, 872
morphological indices of Hawaiian, 866t
overview about, 865
pyroducts, 866-870, 866t
canyon formation, 868—869, 870f
double-trunked, 869—870
formation of, 867, 868f
Huehue Cave and, 869f
Huehue Flow and, 869— 870, 871f
inflation and, 867, 868f
internal development, 867—869, 870f
puka and, 868—869
single-trunked, 869
superimposed-trunked, 870
types of, 869-870
secondary, 865-866
erosional caves, 865-866
pit craters, 865
tectonic caves, 865
speleothems in, 784—785, 872
Volcanic vents, 872
Voronja Cave
Arabika Massif
deep cave evolution, 449—450
geology, 447—448, 447f
hydrogeology, 448—449
location and physiography, 443, 444f
Ortobalagan Valley in, 443, 445f
block-fault structure and, 445—447
branches, 445, 445f
combined cave profile, 446f
depth of, 443
evolution of, 449—450
exploration since 1999, 444—445
first documentation of, 443
hydrology, 447
overview about, 443
temperatures, 447
w
Waitomo district dolines, 145
Wakulla Spring underwater cave system,
Florida
Ball and, 874-875
early exploration, 874
exploration and mapping, 878—880, 879f
future of, 880
geology, 875-878
geologic framework, 875, 876f, 877f
hydrology, 876- 878, 877f
paleontology, 875-876, 876f
speleogenesis, 878
history, 874
name derivation, 874
overview about, 874—875
Water caves, 308-309
Water chemistry in caves
human impact on, 886
limestone dissolution, 880—885
aqueous solution saturation state,
882-883, 883f
carbon dioxide and, 881, 881t
environmental factors influencing,
884, 884f
groundwater composition and,
881-882, 882t
strong acids role in, 884-885
mineral precipitation and speleothems
formation, 885—886
overview about, 880
Water Sinks Cave, 82
Water table, 528-529
Water tracing, in karst aquifers
definitions and objectives of, 886—887
dry set system, 892
dye amounts to use, 891
history, 887-889, 887f
passive detector types
activated carbon, 890
cotton, 891
plankton netting, 891
Quinlan Gumdrop, 891, 891f
qualitative
using passive detectors, 890-892
quantitative, 892—893
test interpretation, 893-896, 894f
dye recovery or breakthrough curve
and, 894, 894f, 895f
tracers and, 889-890
categories of, 889
cautions, 890
fluorescent dyes, 889, 890t
ideal characteristics, 889
Waterdogs, 665—667
Waterline notch, 425
West Virginia spring salamander, 673f, 674
Wetlands, 800-801
Cerknisko jezero/polje, 898-899,898f, 899f
defined, 896
distribution of major, 897, 897f
overview about, 896—898
Pivka intermittent lakes, 899—901, 901f,
902f
Planinsko polje, 899, 900f
protection, 901—902, 902t
temporary, 897
White Cave food web, 60—61, 61f
White-nose syndrome (WNS), 52—53,
625-626
cave closures and decontamination, 908
characteristics, 905—907
atypical winter behavior, 905—906
depleted fat reserves, 905
fungal infections and tissue damage,
905, 905f
immune response changes, 906
wing damage, 906—907, 906f
current and future research, 907—908
diagnostics, 904-905
European connection, 907
fungal transmission, 907
mortality and, 904
overview, 903—904
pathogen, 903f, 904
species affected by, 905
Wild caving, 643
Wilford, G. E., 532
Wind Cave, South Dakota, 86
Wind turbines, 53
Wishing Well Cave, 82
Withidae, 792
WNS. See White-nose syndrome
Wolf spiders, 15
Wolves, 594-595
Worms
Annelida, 911—914
Hirudinea, 913-914, 913f
Oligochaeta, 912-913, 913f
Polychaeta, 911-912, 912f
characterization of invertebrate types,
910-914
earthworms, 362
ecological and phylogenetic remarks, 914
general characteristics, 909
Nematoda, 911, 911f
Nemertea, 911, 912f
Onychophora, 914, 914f
studies history, 909—910
Turbellaria, 910—911
Temnocephalida, 910—911, 910f, 911f
Tricladida, 910, 910f
INDEX
945
X
Xel-Ha to Sian Ka'an Biosphere Reserve
region, 835-836, 835f
Y
Yangpterochiropera. See Bats
Yates Formation, 457—458, 457f
Yinchiroptera. See Bats
Yucatan Peninsula, underwater caves of
Campeche State and, 833-834
cenotes and, 836-837
deep well data paucity regarding, 833
geologic history, 833
hydrology, 833
pre-Mayans/Mayans and, 836
Quintana Roo State and, 834
Sistema K'oox baal, 836f
Sistema Ox Bel Ha, 837
Sistema Sac Actun, 837—838
Xel-Ha to Sian Ka'an Biosphere Reserve
region, 835-836, 835f
Yucatan State and, 834
z
Zalmoxidae, 797
Zerconidae, 794
Zero-inflated spatial models, 478—479
Zoloushka Cave, 829t
Zospeum, 513
Zospeum kusceri, 231f
Zygentoma, 216, 245t
Live Music
Archive
Librivox
Free Audio
Metropolitan Museum
Cleveland
Museum of Art
Internet
Arcade
Console Living Room
Open Library
American
Libraries
TV News
Understanding
9/11